Loading [MathJax]/jax/output/SVG/jax.js
Review Special Issues

Association between problematic TikTok use and mental health: A systematic review and meta-analysis

  • Background 

    TikTok is a significant part of social media usage, since 25.6% of the total global population has a TikTok account, and, thus, scholars should pay attention to its association with users' mental health.

    Objective 

    To synthesize and evaluate the association between problematic TikTok use and mental health.

    Methods 

    We applied the Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines in our review. The review protocol was registered with PROSPERO (CRD42024582054). We searched PubMed, Scopus, Web of Science, PsycINFO, ProQuest, and CINAHL until September 02, 2024.

    Results 

    We identified 16 studies with 15,821 individuals. All studies were cross-sectional and were conducted after 2019. Quality was moderate in 10 studies, good in three studies, and poor in three studies. Our random effects models showed a positive association between TikTok use and depression (β = 0.321, 95% confidence interval: 0.261 to 0.381, p < 0.001, I2 = 78.0%, n = 6 studies), and anxiety (β = 0.406, 95% confidence interval: 0.279 to 0.533, p < 0.001, I2 = 94.8%, n = 4 studies). Data to perform meta-analysis with the other mental health variables were limited. However, our descriptive data showed a positive association between TikTok use and body image issues, poor sleep, anger, distress intolerance, narcissism, and stress.

    Conclusion 

    Our findings suggest that problematic TikTok use has a negative association with several mental health issues. Given the high levels of TikTok use, especially among young adults, our findings are essential to further enhance our understanding of the association between TikTok use and mental health. Finally, there is a need for further studies of better quality to assess the association between problematic TikTok use and mental health in a more valid way.

    Citation: Petros Galanis, Aglaia Katsiroumpa, Zoe Katsiroumpa, Polyxeni Mangoulia, Parisis Gallos, Ioannis Moisoglou, Evmorfia Koukia. Association between problematic TikTok use and mental health: A systematic review and meta-analysis[J]. AIMS Public Health, 2025, 12(2): 491-519. doi: 10.3934/publichealth.2025027

    Related Papers:

    [1] Cui Yin Huang, Hong Liang Dai . Learning from class-imbalanced data: review of data driven methods and algorithm driven methods. Data Science in Finance and Economics, 2021, 1(1): 21-36. doi: 10.3934/DSFE.2021002
    [2] Sami Mestiri . Credit scoring using machine learning and deep Learning-Based models. Data Science in Finance and Economics, 2024, 4(2): 236-248. doi: 10.3934/DSFE.2024009
    [3] Lindani Dube, Tanja Verster . Interpretability of the random forest model under class imbalance. Data Science in Finance and Economics, 2024, 4(3): 446-468. doi: 10.3934/DSFE.2024019
    [4] Dominic Joseph . Estimating credit default probabilities using stochastic optimisation. Data Science in Finance and Economics, 2021, 1(3): 253-271. doi: 10.3934/DSFE.2021014
    [5] Michael Jacobs, Jr . Benchmarking alternative interpretable machine learning models for corporate probability of default. Data Science in Finance and Economics, 2024, 4(1): 1-52. doi: 10.3934/DSFE.2024001
    [6] Michael Jacobs Jr. . Validation of corporate probability of default models considering alternative use cases and the quantification of model risk. Data Science in Finance and Economics, 2022, 2(1): 17-53. doi: 10.3934/DSFE.2022002
    [7] Ying Li, Keyue Yan . Prediction of bank credit customers churn based on machine learning and interpretability analysis. Data Science in Finance and Economics, 2025, 5(1): 19-34. doi: 10.3934/DSFE.2025002
    [8] Jiawei He, Roman N. Makarov, Jake Tuero, Zilin Wang . Performance evaluation metric for statistical learning trading strategies. Data Science in Finance and Economics, 2024, 4(4): 570-600. doi: 10.3934/DSFE.2024024
    [9] Mazin Fahad Alahmadi, Mustafa Tahsin Yilmaz . Prediction of IPO performance from prospectus using multinomial logistic regression, a machine learning model. Data Science in Finance and Economics, 2025, 5(1): 105-135. doi: 10.3934/DSFE.2025006
    [10] Angelica Mcwera, Jules Clement Mba . Predicting stock market direction in South African banking sector using ensemble machine learning techniques. Data Science in Finance and Economics, 2023, 3(4): 401-426. doi: 10.3934/DSFE.2023023
  • Background 

    TikTok is a significant part of social media usage, since 25.6% of the total global population has a TikTok account, and, thus, scholars should pay attention to its association with users' mental health.

    Objective 

    To synthesize and evaluate the association between problematic TikTok use and mental health.

    Methods 

    We applied the Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines in our review. The review protocol was registered with PROSPERO (CRD42024582054). We searched PubMed, Scopus, Web of Science, PsycINFO, ProQuest, and CINAHL until September 02, 2024.

    Results 

    We identified 16 studies with 15,821 individuals. All studies were cross-sectional and were conducted after 2019. Quality was moderate in 10 studies, good in three studies, and poor in three studies. Our random effects models showed a positive association between TikTok use and depression (β = 0.321, 95% confidence interval: 0.261 to 0.381, p < 0.001, I2 = 78.0%, n = 6 studies), and anxiety (β = 0.406, 95% confidence interval: 0.279 to 0.533, p < 0.001, I2 = 94.8%, n = 4 studies). Data to perform meta-analysis with the other mental health variables were limited. However, our descriptive data showed a positive association between TikTok use and body image issues, poor sleep, anger, distress intolerance, narcissism, and stress.

    Conclusion 

    Our findings suggest that problematic TikTok use has a negative association with several mental health issues. Given the high levels of TikTok use, especially among young adults, our findings are essential to further enhance our understanding of the association between TikTok use and mental health. Finally, there is a need for further studies of better quality to assess the association between problematic TikTok use and mental health in a more valid way.



    Abbreviations: ADA: adaBoost; ADASYN: Adaptive synthetic sampling; ANN: Artificial neural network; AUC: Area under the curve; AUROC: Area under the ROC curve; DT: Decision tree; GB: Gradient boosting; kNN: k-Nearest neighbors; LGBM: Light gradient boosting machine; LGBM: LightGBM; LR: Logistic regression; MCC: Mathews' correlation coefficient; NB: Naïve Bayesian; RF: Random forest; ROC: Receiver operating characteristic; SGDC: Stochastic descent gradient; UNDER: Under-sampling; XGB: Extreme gradient boosting

    The accurate estimation of credit default risk plays a critical role in the financial industry, enabling banks and lenders to make informed decisions regarding loan approvals, credit limits, and pricing. Traditionally, credit risk assessment has relied on statistical models and expert judgment. However, with the advancements in machine learning techniques and the availability of large-scale credit data, there has been growing interest in utilizing these methods for credit risk prediction.

    One significant challenge in credit risk modeling is dealing with class imbalance, where the number of default instances is significantly smaller than the non-default instances. This imbalance can lead to biased models and poor predictive performance (Thabtah et al., 2020), as most machine learning algorithms are designed to maximize overall accuracy and may struggle to accurately classify the minority class. Consequently, misclassification errors related to defaults can have severe financial implications. The consequences of misclassifying credit defaults can be significant. Khemakhem and Boujelbene (2018) argued that false negatives (predicting a non-default when it is actually a default) can expose lenders to potential losses and increased credit risk. On the other hand, false positives (predicting a default when it is actually a non-default) can result in unnecessary restrictions on credit access for borrowers and potential loss of business for lenders. The impact of class imbalance on credit risk prediction has gained attention in recent research. It is crucial to understand the behaviour and limitations of machine learning algorithms (Leo et al., 2019) under imbalanced conditions to develop robust models that effectively capture the risk associated with credit defaults. This understanding can help financial institutions enhance their decision-making processes, mitigate potential losses, and ensure fair access to credit for borrowers.

    The objective of this research article is to investigate the effect of class imbalance on the estimation of default probabilities using machine learning algorithms. We aim to analyse the performance of various algorithms under different class distributions and evaluate their effectiveness in capturing default events accurately. Additionally, we will explore different techniques for addressing class imbalance, such as over-sampling, particularly ADASYN sampling and under-sampling to assess their impact on model performance. To achieve our research objectives, we will utilize a dataset of credit information, including borrower characteristics, historical payment behaviour and other relevant factors. We will compare the performance of different machine learning algorithms, such as logistic regression, random forest, gradient boosting and decision tree in estimating default probabilities. Additionally, we will evaluate the Mathews' correlation coefficient (MCC) and F1-scores of the models using metrics like the confusion matrix and area under the receiver operating characteristic curve (AUC-ROC).

    The findings of this study will contribute to the existing literature on credit risk modeling by providing insights into the effect of class imbalance on default probability estimation. This systematic investigation provides a deeper understanding of the critical impact of class imbalance on predictive analytics. By evaluating ten distinct classification models using rigorous evaluation metrics such as the area under the ROC curve, Mathews' correlation coefficient (MCC) and F1-scores, this research offers empirical insights into the strengths and weaknesses of these models in the context of imbalanced datasets. With the evidence-based assessment of various classification models and balancing techniques, this study serves as a valuable guide for practitioners and researchers dealing with imbalanced datasets. The findings offer clear directions for selecting appropriate models and applying tailored data balancing strategies, ultimately enhancing classification performance in the presence of class imbalance. It will also shed light on the strengths and limitations of different machine learning algorithms in the context of imbalanced credit data. The results will be valuable for financial institutions and regulators in developing more accurate and reliable credit risk models, enhancing credit decision-making processes, and promoting fair access to credit.

    In summary, this research article aims to bridge the gap in understanding the impact of class imbalance on credit risk prediction. By investigating the behavior of machine learning algorithms under imbalanced conditions and exploring techniques to address the imbalance, we seek to improve the accuracy and reliability of default probability estimation. The outcomes of this study will contribute to the advancement of credit risk modeling and have practical implications for the financial industry.

    One of the major challenges when building default prediction models, is the issue of imbalanced data. Class imbalance occurs whenever one majority class's training samples vastly outnumber those of the other minority class. Research has revealed that algorithms trained on an imbalanced dataset tend to suffer from a prediction biasedness and this often results in poor performance in the minority class. This paper will be exploring the results across ADASYN sampling and under-sampling. Figure 1 below outlines the proposed methodology that is adopted in this paper.

    Figure 1.  Adopted proposed approach.

    The initial and fundamental step in dealing with any sort of data is to first clean the data thoroughly and make sure that it makes sense (e.g. no nuisance entries). Given the nature of our dataset (see Section 4.), the first step was to standardise all the explanatory variables min/max scaling. The motivation for this was that whenever training an algorithm with a variable such as salary, which can range from 10000 to 665000 while a variable such as credit utilization is captured as a ratio, the classifier might assume salary is more important than credit utilization. Min/max scaling (Patro and Sahu, 2015), which was done in R-studio, scales all the numerical variables to range between 0 and 1.

    Cross-validation consists of developing models explaining relationships among variables based on a subset of data, called the training data, and then testing the model on the testing data. K-fold cross-validation splits the data (K) times into training and testing data and then identifies the model that performs best in the aggregate (Cawley and Talbot, 2010). An even more refined approach reserves another 20% of the data for a separate testing stage, which is not part of model development and testing, but is used instead for out-of-sample testing of the model obtained on the basis of the training and testing datasets. In cross-validation, the training and testing data are separated and the testing data are used only when a best-fitting model has emerged (Anguita et al., 2012). In our case, we have used 5-fold cross validation on the 80% (in-sample) of the data and used the remaining 20% (out-of-sample) of the data for model testing in order to get more accurate results.

    There are two different strategies for handling missing data (Han et al., 2012). The first strategy is to simply ignore missing values and the second strategy is to consider imputation of missing values.

    Omit missing values

    The serious problem with omitting observations with missing values is that it reduces the dataset size. This is appropriate when your dataset has a small amount of missing values. There are two general approaches for ignoring missing data: listwise deletion (case deletion or complete case analysis) and pairwise deletion (available case analysis) approach. Complete case analysis approach excludes all observations with missing values for any variable of interest. This approach limits the analysis to those observations for which all values are observed which often results in biased estimate and loss of precision (Schafer and Graham, 2002). In pairwise deletion, we perform analysis with all cases in which the variables of interest are present. It does not exclude the entire unit but uses as much data as possible from every unit. The advantage of this method is that it keeps maximum available data for analysis even when some of its variables have missing values. The disadvantage of this method is that it uses different sample size for different variables (Schafer and Graham, 2002). The sample size for each individual analysis is higher than the complete case analysis.

    Impute missing values

    Missing data imputation is a procedure that replaces missing values with some plausible values (Rubin, 1976). The various imputation techniques aim to provide accurate estimation of population parameters so that power of data mining and data analysis techniques is not reduced. Optimal treatment to be given to the missing data depends on amount of missing data. Although there is no general rule on what percentage of missing data is bad, it is always better to do comparison of results before and after imputation.

    In this paper we have adopted the median imputation method for handling missingness. Median imputation is used for numerical data and our dataset was of this composition. Median imputation is a method for handling missing values by replacing missing values in a dataset with the median value of the non-missing observations of the same variable. This method assumes that the missing data are missing at random and that the median is a good representation of the central tendency of the data. The median is calculated by first ordering the nonmissing observations of a variable and then identifying the middle value or the average of the two middle values, depending on whether the number of observations is odd or even. This imputed median value is then used to replace all the missing values of that variable.

    Initially, we present a comprehensive review of the relevant literature pertaining to the research topic. Subsequently, we expound on the adopted methodology employed in this study. The models utilized are discussed and references to additional research papers are provided for supplementary understanding. Additionally, we furnish a detailed account of the analyzed dataset, coupled with an exploratory data analysis. Finally, an extensive analysis of the results of our machine learning algorithms is presented, along with recommendations for future research.

    Hyper-parameter tuning in machine learning is the process of selecting the optimal values for hyper-parameters, which are parameters set by the user that control the behaviour of the learning algorithm. The goal is to find the hyper-parameters that result in the best balance of model complexity and performance. This process can be time-consuming and computationally expensive but it is an important step in developing accurate and reliable machine learning models. If default hyper-parameters were used for the models in R, the opportunity to fine-tune the models and achieve optimal performance for the specific task or dataset may have been missed.

    In order to develop and evaluate the performance of our machine learning models, we utilized default hyper-parameters in the R programming language. While this approach may not have allowed for the fine-tuning of hyper-parameters to achieve optimal performance for our specific task and dataset, it allowed us to establish a baseline level of performance and compare the relative performance of different models. This information was valuable in guiding our model selection process and identifying areas for future improvement.

    ADASYN and under-sampling are techniques used in machine learning to address class imbalance in datasets. Under-sampling involves reducing the number of instances in the majority class to create a more balanced dataset, allowing the classifier to learn effectively from both classes. This can be achieved through methods such as random under-sampling or removing instances close to the decision boundary. On the other hand, ADASYN sampling takes a more adaptive approach by generating synthetic examples for the minority class, particularly focusing on difficult-to-learn instances. By augmenting the minority class, ADASYN aims to improve the classifier's performance and achieve better predictive accuracy. While under-sampling can lead to loss of information from the majority class, ADASYN sampling leverages the distribution of the minority class to generate synthetic samples and overcome the imbalance issue. Both techniques aim to enhance the learning process in imbalanced datasets, but they adopt different strategies to achieve a balanced representation of the classes.

    In data analytics, (Breeden, 2021) machine learning is a set of computational methods which use experience to improve performance or to make accurate predictions. Here, the word "experience" refers to past information available to the machine learning technique, classifier. In particular, data quality and data size are at the core of machine learning and, since the success of a learning algorithm depends on the data used, machine learning is strictly related to data analysis and statistics.

    Learning is a wide domain, consequently it can be ranched into subfields dealing with different types of learning. The most common partition is the one that distinguishes between supervised and unsupervised learning according to the types of training data available to the classifier (Breeden, 2021). Figure 2 depicts the word-cloud jargon of machine learning in credit risk modelling. In supervised learning, an algorithm is trained using labelled data to make predictions; this is the most common scenario when dealing with classification or regression problems. In unsupervised learning, an algorithm is fed with unlabelled data where an algorithm is tasked with learning from the data on its own and be able to make accurate predictions when given unseen data; this approach is popular in clustering and association problems. Another type of machine learning is reinforcement learning, where an intelligent agent ought to take actions in an environment in order to maximize the notion of cumulative reward. This is used largely for classification and control problems.

    Figure 2.  Machine learning in statistics.

    Decision tree (DT): Decision tree algorithm is a popular method for default prediction due to its simplicity, interpretability and its ability to handle large datasets with high dimensionality. It uses a tree-like model of decisions and their possible consequences, by recursively partitioning the feature space into smaller regions, in which the most homogeneous set of outcomes is found. However, decision trees are known to be sensitive to class imbalance since they tend to be biased towards the majority class. Breiman et al. (1984) and Fayyad and Irani (1992) give full description of the model.

    k-Nearest neighbor (kNN): The k-nearest neighbour classifier (kNN) is known to be most useful instance-based learners. kNN is a non-parametric model. kNN (Yao and Ruzzo, 2006) is a non-parametric method that makes predictions based on the majority class of the k-nearest points to a given test point. It is simple and efficient but can be sensitive to the choice of k and the distance metric used. kNN has been used in the literature for default prediction, mainly in the credit risk domain. A comprehensive description of kNN is provided by Kelleher et al. (2020), Stephens and Diesing (2014) and Wilson and Martinez (1997).

    Logistic regression model (LR): One of the most commonly used statistical models is the logistic regression model that explains the relationship of several covariates x to a binary response variable. The primary objective of the logistic regression model (Zhang et al., 2017) with multiple predictors is to construct a model to describe the relationship between a binary response variable and one or more predictor variables. Logistic regression is widely used in various fields, including medicine, finance, and social sciences, where binary classification tasks are common. It provides interpretable results because the coefficients can be examined to understand the impact of the features on the probability of the positive class. However, logistic regression assumes a linear relationship between the features and the log-odds, and may not perform well when dealing with complex nonlinear patterns in the data.

    Naïve Bayesian approach (NB): The naïve Bayes classifier is a probabilistic classification algorithm based on Bayes' Theorem that has been widely used for default prediction problems. It makes the assumption that the predictors are independent given the class label, which is called the "naive" assumption. Despite its "naive" assumption, it has been shown to be effective in several studies and it can handle class imbalance by adjusting class weights or using techniques like oversampling, under-sampling and synthetic data generation. A full description of the algorithm can be found in (De Campos et al., 2011) and (Stephens and Diesing, 2014).

    Light gradient boosting machine (LGBM): LGBM is a powerful machine learning model that has gained popularity in both regression and classification tasks. It is a gradient boosting framework that utilizes tree-based learning algorithms (Ke et al., 2017). LGBM is designed to handle large-scale datasets efficiently, making it suitable for real-world applications with high-dimensional features. One of the key strengths of LGBM lies in its ability to handle imbalanced datasets effectively. It employs a technique called gradient-based one-side sampling (GOSS) to downsample the majority class during the boosting process, which helps to improve the model's performance on minority classes. This makes LGBM particularly well-suited for classification tasks where class imbalances are common. Ke et al. (2017) and Li et al. (2022) provide more context to the LGBM model.

    Random forest (RF): As proposed by Breiman (2001), random forest is a combination of decision trees (Ho, 1995) used as an ensemble learning method for classification, regression and other tasks that operates by constructing a multitude of decision trees at training time and outputting the class that is the mode of the classification or regression of the individual trees. Deng et al. (2018) defined random forest as a type of learning method that can be used for both classification and regression problems. Furthermore, random forests are suitable for large quantities of data with substantial noise, can prevent over-fitting, and are able to distinguish important features in classification. Breiman (2001), Calderoni et al., (2015) and Booth et al., (2015) give a more detailed discussion on random forests, including a more rigorous mathematical description.

    Adaptive boosting (ADA): The Adaptive Boosting (ADA) Classifier is a popular machine learning algorithm used for classification tasks. It is a type of ensemble learning method that combines multiple weak classifiers to create a strong and accurate classifier. ADA iteratively trains a series of weak classifiers, where each subsequent classifier is designed to focus on the instances that were misclassified by the previous classifiers. This iterative process helps to improve the overall performance of the classifier. Granström and Abrahamsson (2019) provides the full details regarding the implementation of this algorithm.

    Stochastic gradient descent classifier (SGDC): The stochastic gradient descent (SGDC) classifier is a popular and efficient algorithm used in machine learning for solving classification problems. It is a variant of the gradient descent optimization algorithm (Lokeswari and Amaravathi, 2018), specifically designed for large-scale datasets. The SGDC iteratively updates the model parameters by taking small steps in the direction of the steepest gradient, aiming to minimize the loss function. Unlike traditional gradient descent (Liu et al., 2021), which calculates the gradient over the entire training dataset, SGDC performs updates on randomly selected subsets of data called mini-batches. This stochastic nature of SGDC allows for faster convergence and makes it highly suitable for working with massive datasets.

    Gradient boosting (GB): Gradient boosting is a powerful machine learning algorithm that combines multiple weak learners, typically decision trees, to create a strong predictive model. It operates by sequentially adding new models to correct the errors made by the previous models, thereby gradually improving its predictive accuracy. The algorithm works by optimizing a specific loss function through an iterative process. Each subsequent model is trained to minimize the errors or residuals of the previous models, using gradient descent optimization. Gradient boosting is known for its ability to handle complex nonlinear relationships in data and is widely used in various domains, including regression, classification, and ranking problems. It has gained popularity due to its high predictive performance and robustness. Dorogush et al., (2018) and Bentéjac et al., (2021) expand more on the model.

    Extreme gradient boosting (XGB): Another integration technique constructed by continuous iterations of weak classifier is the extreme gradient boosting. According to Ogunleye and Wang (2019), the model was proposed by Chen and Guestrin (2016) to optimize memory usage and exploit the hardware computing power, XGB decreases the execution time with an increased performance compared to many machine learning algorithms and even deep learning models. The main idea of boosting is to sequentially build sub-trees from an original tree such that each subsequent tree reduces the errors of the previous one. In this procedure, k number of regression trees are created to ensure that the prediction of the tree cluster is as close to the actual value as possible and that the generalization capability is as high as possible. More details about the procedure can be read by Dhieb et al., (2019), Ogunleye and Wang (2019) and Chen and Guestrin (2016).

    In this section we provide some information on the dataset utilised, exploratory data analysis and we also motivate the aptness for the selection of our model choice. Kaggle dataset was used in this paper, which contained 11 features and 150000 observations (Kaggle, 2023). Kaggle is a well-known platform for data science competitions, collaboration and learning. It hosts a wide variety of datasets contributed by the community, covering diverse topics and domains. These datasets are often used for data analysis, machine learning projects and research. Kaggle datasets range from structured data in CSV files to images, videos and more complex data types. Table 1, gives the dictionary to the dataset being adopted in this paper.

    Table 1.  Credit risk dataset.
    Classifier Method Type
    SeriousDlqin2yrs Person experienced 90 days past due delinquency or worse Binary
    RevolvingUtilizationOfUnsecuredLines Total balance on credit cards and personal lines of credit except real estate and no instalment debt like car loans divided by the sum of credit limits Ratio
    Age Age of borrower in years Integer
    NumberOfTime30-59DaysPastDueNotWorse Number of times borrower has been 30-59 days past due but no worse in the last 2 years (Bucket_1) Integer
    DebtRatio Monthly debt payments, alimony, living costs divided by monthly gross income Ratio
    MonthlyIncome Monthly income Numeric
    NumberOfOpenCreditLinesAndLoans Number of Open loans (insta.0=lment like car loan or mortgage) and Lines of credit (e.g. credit cards) Integer
    NumberOfTimes90DaysLate Number of times borrower has been 90 days or more past due (Bucket_3) Integer
    NumberRealEstateLoansOrLines Number of mortgage and real estate loans including home equity lines of credit Integer
    NumberOfTime60-89DaysPastDueNotWorse Number of times borrower has been 60-89 days past due but no worse in the last 2 years (Bucket_2) Integer
    NumberOfDependents Number of dependents in family excluding themselves (spouse, children etc.) Integer

     | Show Table
    DownLoad: CSV

    Roughly 2% of the data was missing, particularly within the monthly income variable as well as the number of dependents. This is shown visually in Figure 3. We thought it was worthwhile to check if there is no relationship in our explanatory variables before attempting to fit any models. This task is termed as checking for the existence of multicollinearity within the data. Figure 4 displays the results that were obtained after the test was conducted. The results show a very strong correlation in the Bucket 1 through Bucket 3 variable. Moreover, a high correlation was also seen between Bucket 1 and Number Real Estate Loans Or Lines.

    Figure 3.  Missing values.
    Figure 4.  Multicollinearity.

    The dataset had originally 7% (10026) positive cases and 93% (139974) negative cases. Since the objective of this paper was to investigate the effectiveness of various machine learning models under class imbalance, we have generated nine samples of different levels of class imbalance for each of the sampling techniques discussed in Section 2.3. As a result, we ended up with eighteen (18) samples as shown in Table 2. In under-sampling, the minority class was kept the same while the majority class was under sampled to meet the desired class imbalance. On the other hand, in ADASYN sampling the majority class was fixed at original observations while the minority class was over sampled to meet the desired samples of different class imbalance.

    Table 2.  Simulated sample sizes.
    Response Sample 1 Sample 2 Sample 3 Sample 4 Sample 5 Sample 6 Sample 7 Sample 8 Sample 9
    UNDER Yes 10026 10026 10026 10026 10026 10026 10026 10026 10026
    No 100260 66840 50130 40104 33420 28646 25065 22280 20052
    ADASYN Yes 15553 24699 34994 46658 59989 75371 93316 114524 139974
    No 139974 139974 139974 139974 139974 139974 139974 139974 139974

     | Show Table
    DownLoad: CSV

    We adopt the widely used measures of performance in the fields of credit risk to evaluate our classification algorithms. These include the area covered by the receiver operating characteristics (ROC) curve. The ROC curve tells how much a model is capable of distinguishing between classes; an excellent model will have an ROC close to 1, a poor model will have ROC close to 0.5. The ROC curve is constructed by evaluating the fraction of "true positives"(TP) and "false positives" (FP) for different threshold values. Table 3 shows the so-called confusion matrix that contains basic ingredients that we usually report on.

    Table 3.  Confusion matrix.
    Actual
    Positive Negative
    Predicted Positive True Positive (TP) False positive (FP)
    Negative False Negative (FN) True Negative (TP)

     | Show Table
    DownLoad: CSV

    We report on the following metrics,

    Note:

    These formulas are derived using 2x2 confusion matrix (Table 3) (Mitchell and Mitchell, 1997), for multi-class classification or multi-label classification the formulas will be different. ROC-AUC is a measure of the trade-off between the true positive rate (sensitivity) and the false positive rate (1-specificity) in a binary classification problem. AUC of 1 represents a perfect classifier and AUC of 0.5 represents a random classifier. Recall gain is used when the data is imbalanced and the goal is to improve recall and it is defined as: recall gain = (recall of model - recall of baseline)

    This section presents an overview of the evaluation measures obtained after training the classification models in R. Prior to fitting the models, our dataset underwent a simulation process to create nine samples with varying levels of class imbalance using both under-sampling and ADASYN sampling techniques, as depicted in Figure 5. Each sample was characterized by a specific percentage of positive target variables, such as 10% in the first sample and 90% for the remaining variables. Subsequent samples followed a similar pattern, with increasing percentages of positive target variables.

    Figure 5.  Simulation of class imbalance samples.

    The evaluation focused on five key metrics: the area under the ROC curve, Mathews' correlation coefficient (MCC), Gini coefficient, recall scores and F1-scores. The results by the Gini coefficient and the recall scores can be found in Appendix 7. These metrics were analysed for the two sampling techniques investigated in the study. The area under the ROC curve provided insights into the models' performance in distinguishing between the positive and negative classes. MCC served as a measure of the models' overall performance, taking into account true positives, true negatives, false positives, and false negatives. Last, the F1-scores provided a balanced assessment of precision and recall for the models. These evaluation measures were instrumental in assessing the performance of the classification models and drawing meaningful conclusions about their effectiveness under different levels of class imbalance.

    After generating and securely saving our simulated samples with varying degrees of class imbalance, we proceeded to the modeling phase. Each simulated dataset was divided into two parts: 80% for training the model and 20% for model testing. We employed 10 machine learning models (as discussed on Section 3.) on each of the simulated samples and reported on Mathews' correlation coefficient to compare training sets. We further used F1-scores and area under the ROC curves to validate the performance of the models on out-of-sample datasets. Table 4 below, summarises the results obtained after training and the testing the models according to Mathews' correlation coefficient.

    Table 4.  Mathews' correlation coefficient from training and testing sets.
    Model Sample 1 Sample 2 Sample 3 Sample 4 Sample 5 Sample 6 Sample 7 Sample 8 Sample 9
    Training Sets Under-Sampling LR 0.1717 0.2210 0.2259 0.2642 0.2840 0.3341 0.3062 0.4202 0.4117
    DT 0.9977 0.9987 0.9988 0.9989 0.9990 0.9993 0.9991 0.9993 0.9994
    RF 0.9975 0.9986 0.9987 0.9989 0.9990 0.9993 0.9991 0.9993 0.9993
    ADA 0.3547 0.4298 0.4565 0.4893 0.5181 0.5190 0.5261 0.5353 0.5460
    GB 0.4011 0.4743 0.4884 0.5217 0.5491 0.5552 0.5581 0.5658 0.5813
    XGB 0.5842 0.6472 0.6645 0.6941 0.7101 0.7170 0.7284 0.7316 0.7546
    LGBM 0.4700 0.5393 0.5570 0.5884 0.6049 0.6130 0.6222 0.6292 0.6388
    SGDC 0.0934 0.1457 0.1155 0.2283 0.2959 0.3614 0.3845 0.4441 0.4415
    KNN 0.5584 0.5888 0.5743 0.6001 0.6097 0.6046 0.6028 0.6049 0.6106
    NB 0.1107 0.0984 0.0955 0.0942 0.1313 0.1296 0.1045 0.1392 0.1126
    ADASYN Sampling LR 0.1823 0.2143 0.2369 0.2769 0.2741 0.3415 0.3733 0.3712 0.3843
    DT 0.9976 0.9981 0.9983 0.9984 0.9982 0.9987 0.9987 0.9988 0.9987
    RF 0.9975 0.9980 0.9983 0.9984 0.9981 0.9987 0.9987 0.9988 0.9987
    ADA 0.3741 0.4277 0.4588 0.4888 0.5035 0.5202 0.5325 0.5438 0.5421
    GB 0.4081 0.4574 0.4930 0.5178 0.5354 0.5446 0.5586 0.5658 0.5644
    XGB 0.5666 0.5987 0.6179 0.6387 0.6470 0.6541 0.6644 0.6690 0.6680
    LGBM 0.4572 0.4996 0.5313 0.5514 0.5671 0.5782 0.5868 0.5959 0.5931
    SGDC 0.0000 0.0924 0.1761 0.2688 0.2640 0.3902 0.4159 0.4330 0.4251
    KNN 0.6121 0.7047 0.7576 0.8016 0.8307 0.8525 0.8694 0.8792 0.8886
    NB 0.1012 0.1024 0.1065 0.1233 0.1299 0.1412 0.1260 0.1207 0.1257
    Testing Sets Under-Sampling LR 0.1813 0.2096 0.2388 0.2754 0.2617 0.3036 0.2967 0.4239 0.4164
    DT 0.6337 0.6225 0.6324 0.6228 0.6349 0.6374 0.6567 0.6246 0.6240
    RF 0.7333 0.7371 0.7393 0.7394 0.7387 0.7261 0.7551 0.7389 0.7472
    ADA 0.3565 0.4099 0.4489 0.4863 0.4995 0.4919 0.5256 0.5470 0.5448
    GB 0.3929 0.4356 0.4756 0.5064 0.5293 0.5177 0.5472 0.5742 0.5799
    XGB 0.4706 0.4917 0.5258 0.5714 0.5754 0.5611 0.6095 0.6054 0.6275
    LGBM 0.4159 0.4577 0.4973 0.5403 0.5443 0.5387 0.5753 0.5816 0.6061
    SGDC 0.1024 0.1344 0.1272 0.2351 0.2958 0.3302 0.3936 0.4503 0.4548
    KNN 0.3708 0.4109 0.3795 0.4279 0.4383 0.4115 0.3921 0.4213 0.4072
    NB 0.1199 0.0944 0.1008 0.1135 0.1387 0.1279 0.1153 0.1243 0.0989
    ADASYN Sampling LR 0.1718 0.2050 0.2277 0.2747 0.2806 0.3399 0.3754 0.3701 0.3887
    DT 0.7082 0.7999 0.8592 0.8806 0.9068 0.9211 0.9324 0.9363 0.9421
    RF 0.8048 0.8815 0.9178 0.9349 0.9491 0.9612 0.9627 0.9677 0.9725
    ADA 0.3641 0.4236 0.4544 0.4875 0.5113 0.5209 0.5324 0.5413 0.5444
    GB 0.3940 0.4534 0.4771 0.5157 0.5322 0.5439 0.5549 0.5559 0.5666
    XGB 0.4767 0.5319 0.5625 0.5861 0.6143 0.6237 0.6299 0.6369 0.6459
    LGBM 0.4180 0.4814 0.5025 0.5331 0.5590 0.5632 0.5766 0.5784 0.5894
    SGDC 0.0000 0.0894 0.1810 0.2768 0.2650 0.3921 0.4125 0.4273 0.4275
    KNN 0.3950 0.4919 0.5875 0.6383 0.7005 0.7374 0.7684 0.7942 0.8164
    NB 0.0895 0.1003 0.1065 0.1225 0.1296 0.1439 0.1157 0.1198 0.1209

     | Show Table
    DownLoad: CSV

    First, it is imperative to highlight the exceptional performance of the random forest (RF) and decision tree (DT) algorithms in both the training and testing phases. Even when trained on highly imbalanced datasets (Sample 1), both algorithms achieved Mathews' correlation coefficient (MCC) scores exceeding 99% and demonstrated slight improvements as the data approached balance. Remarkably, these models appeared to be relatively insensitive to class imbalance during the training stage. Visually, the dominance of RF and DT algorithms over other models is evident, as depicted in the line plots presented in Figure 6 and Figure 8. The second set of models that exhibited noteworthy performance included ADA, GB, XGB, LGBM and kNN. These models showcased significant improvements as the data became more balanced. With a 10% balanced dataset, these models achieved MCC scores ranging from 35.5–55.8%, which escalated to 54.6–61.1% for a fully balanced dataset. This trend was consistent across both sampling techniques. The line plots in Figure 6 and Figure 8 offer insights into the sensitivity of imbalanced data to these models. In contrast, LR, SGDC, and NB algorithms exhibited the poorest performance across both sampling techniques, scoring MCC values below 50% regardless of the class imbalance in the data. Further analysis on this matter will be explored in the upcoming section.

    Figure 10.  Training and testing sample by Mathews' correlation coefficient.

    The subsequent step of the analysis involved evaluating the performance of these models on separate out-of-sample datasets, which were reserved for model testing. Once again, the standout performers were the RF and DT models. When employing under-sampling, these models achieved scores ranging from as low as 63.4% and 73.3% to as high as 65.7% and 75.5%, respectively. In the case of ADASYN sampling, their scores ranged from as low as 70.8% and 80.5% to as high as 94.2% and 97.3%, respectively. Although the scores decreased compared to the training sets, the RF and DT models continued to outperform the other models. This trend is also visually represented in the line plots depicted in Figure 7 and Figure 9.

    Similarly, ADA, GB, XGB, LGBM and kNN models exhibited comparable patterns in the out-of-sample datasets as observed in the training samples, albeit with slightly lower scores. Notably, the kNN model demonstrated the most significant improvements as the data became more balanced. Despite fully balancing the datasets, the LD, LGBM, kNN, and NB models still scored MCC values below 50% when utilizing under-sampling. However, the kNN model did exhibit some enhancements in prediction quality when employing the ADASYN sampling technique, achieving an MCC score of 86.6%. On the other hand, the LD, LGBM, and NB models remained below 50% in terms of MCC scores.

    The F1-scores presented in Table 5 further reinforce the observation that the RT and DT models were the top performers during the testing stage. When employing the under-sampling technique, these models achieved scores as low as 66.7% and 74.6%, respectively, which improved to 75.5% and 83.1% as the data became more balanced. In the case of ADASYN sampling, their scores improved from 73.5% and 81.6% to 96.1% and 98.2%, respectively. Across both sampling techniques, ADA, GB, XGB, and LGBM models demonstrated gradual improvements in predictive power as the data distribution approached equilibrium. SGDC and kNN models exhibited the most significant improvements as the dataset became more balanced. At a 10% class imbalance, SGDC and kNN models achieved F1-scores of 3.4% and 40.0% respectively, which increased to 52.2% and 57.2% as the data became more balanced. This pattern held true for both sampling techniques employed in this study. On the other hand, LR and NB models performed consistently regardless of the model or sampling technique used.

    Table 5.  F1-score comparison for testing sets.
    Model Sample 1 Sample 2 Sample 3 Sample 4 Sample 5 Sample 6 Sample 7 Sample 8 Sample 9
    Under-Sampling LR 0.1078 0.1426 0.1827 0.2357 0.2348 0.3123 0.3101 0.4820 0.4882
    DT 0.6673 0.6727 0.6976 0.6995 0.7231 0.7357 0.7588 0.7449 0.7554
    RF 0.7461 0.7633 0.7789 0.7866 0.7966 0.7958 0.8244 0.8181 0.8314
    ADA 0.3604 0.4391 0.5002 0.5502 0.5782 0.5897 0.6342 0.6588 0.6687
    GB 0.3937 0.4705 0.5298 0.5698 0.6103 0.6172 0.6561 0.6849 0.7045
    XGB 0.4695 0.5244 0.5789 0.6349 0.6561 0.6599 0.7108 0.7157 0.7456
    LGBM 0.4147 0.4894 0.5541 0.6064 0.6302 0.6404 0.6833 0.6956 0.7265
    SGDC 0.0335 0.0591 0.0606 0.1611 0.2443 0.3261 0.4259 0.5126 0.5223
    KNN 0.3898 0.4505 0.4440 0.5042 0.5344 0.5268 0.5320 0.5646 0.5721
    NB 0.0680 0.0455 0.0530 0.0588 0.0933 0.0737 0.0709 0.0698 0.0573
    ADASYN Sampling LR 0.1059 0.1397 0.1713 0.2413 0.2557 0.3442 0.4043 0.4194 0.4580
    DT 0.7349 0.8259 0.8826 0.9041 0.9282 0.9412 0.9517 0.9562 0.9614
    RF 0.8162 0.8961 0.9315 0.9480 0.9610 0.9713 0.9734 0.9779 0.9817
    ADA 0.3678 0.4535 0.4987 0.5510 0.5873 0.6087 0.6346 0.6552 0.6682
    GB 0.3966 0.4870 0.5292 0.5838 0.6134 0.6353 0.6614 0.6734 0.6927
    XGB 0.4776 0.5615 0.6083 0.6474 0.6867 0.7059 0.7239 0.7388 0.7557
    LGBM 0.4181 0.5133 0.5559 0.6030 0.6397 0.6573 0.6822 0.6945 0.7130
    SGDC 0.0000 0.0259 0.0956 0.2140 0.1999 0.3953 0.4495 0.4818 0.4879
    KNN 0.4225 0.5467 0.6521 0.7108 0.7722 0.8078 0.8374 0.8607 0.8793
    NB 0.0510 0.0486 0.0632 0.0646 0.0805 0.0862 0.0668 0.0705 0.0730

     | Show Table
    DownLoad: CSV

    Finally, we present visualizations of the ROC curves, depicting the area under the curve (AUC), for the testing stage across various sample sizes. The ROC curves, as shown in Figure 12, provide further confirmation of the previous findings. While random forest (RF) and decision tree (DT) models consistently exhibited superior performance compared to other classifiers across different levels of class imbalance, it is crucial for readers to pay attention to the performance improvements of all models as the data becomes more balanced. Notably, a noteworthy observation from the ROC curves is that the AUC values for the under-sampling technique, as illustrated in Figure 13, appeared to be relatively flatter and closer to the diagonal line compared to the AUC values for the ADASYN sampling technique, as depicted in Figure 12. This observation implies that ADASYN sampling tends to produce more reliable predictions compared to the under-sampling technique.

    Figure 11.  F1-Scores comparison.
    Figure 12.  ADASYN sampling ROC curve comparison.
    Figure 13.  Under-sampling ROC curve comparison.

    The visualizations of the ROC curves provide additional evidence of the strength of RF and DT models throughout various class imbalance scenarios. Furthermore, the results highlight the importance of considering the performance improvements of all models as data balance improves. Additionally, the ROC curves suggest that ADASYN sampling may offer enhanced prediction reliability compared to under-sampling.

    In summary, the results highlight the superior performance of RF and DT models in both training and testing stages, emphasizing their robustness to class imbalance. Other models, such as ADA, GB, XGB, LGBM and kNN, showed improvements as the data became more balanced but did not surpass the performance of RF and DT models. LR, SGDC and NB models consistently performed poorly regardless of the sampling technique used. The findings also suggest that ADASYN sampling technique yielded more reliable predictions compared to under-sampling technique.

    The results obtained in this study align with some findings reported in related work on the topic of class imbalance and classification models.

    The superior performance of random forest (RF) and decision tree (DT) algorithms, especially in the presence of class imbalance, is consistent with previous research. Alija et al., (2023) and Zhou and Wang (2012) conducted a similar investigation under class imbalance and discovered that random forest tend to outperform many state-of-art classifiers such as SVM, ANN, naïve Bayes and C4.5. RF and DT models are known for their ability to handle imbalanced datasets effectively by capturing complex decision boundaries and handling both minority and majority classes well. In the paper written by Sun et al., (2018), it was also discovered that decision tree significantly outperforms other models and is effective for imbalanced enterprise credit evaluation. The high MCC scores achieved by RF and DT models in this study support their suitability for imbalanced classification tasks, as reported in previous studies. RF and DT algorithms perform well on imbalanced data due to their inherent robustness to class imbalance (Singhal et al., 2018), the use of sampling and randomness, their ability to handle overlapping regions and the benefits of ensemble methods. DT are less affected by class imbalance as they can capture patterns in both minority and majority classes (Liu et al., 2010) during the splitting process. RF, consisting of multiple decision trees trained on bootstrap samples, introduces randomness and diversity, enabling the algorithm to learn from both classes. DTs can form partitions that help separate the minority class instances, improving classification performance. Finally, the ensemble nature of RF leverages the collective wisdom of multiple trees, further enhancing its ability to handle imbalanced data (Liu et al., 2010).

    The results also align with previous studies that have highlighted the challenges faced by logistic regression, stochastic gradient descent classifier, and naïve Bayes algorithms in imbalanced classification tasks. These models often struggle to handle class imbalance, resulting in lower MCC scores and poorer performance compared to other algorithms. Logistic regression, stochastic gradient descent classifier, and naïve Bayes algorithms face challenges in imbalanced classification tasks due to the skewed class distribution (Aljedaani et al., 2022), loss function optimization, assumption of feature independence, and sensitivity to data representation. According to Das et al., (2018), the skewed class distribution can lead to biased models and difficulties in capturing patterns for the minority class. The loss functions used by LR and SGDC may prioritize the majority class, resulting in biased decision boundaries and poor performance on the minority class. NB's assumption of feature independence can disregard rare but discriminative features for the minority class. Additionally, these algorithms may struggle to find sufficient evidence to accurately model the minority class (Das et al.,, 2018) due to its under-representation. To overcome these challenges, techniques like resampling, adjusting class weights, using different loss functions, or employing specialized algorithms designed for imbalanced data can be applied. The consistent poor performance of LR, SGDC and NB models in this study reinforces the need to carefully select appropriate classifiers when dealing with imbalanced datasets.

    ADA, gradient boosting, extreme gradient boosting, light gradient boosting machine, and k-nearest neighbors models can be sensitive to imbalanced data due to their underlying mechanisms and characteristics:

    1. Data weighting and boosting: Models like ADA, GB, XGB and LGBM utilize boosting techniques, where multiple weak classifiers are combined to form a strong classifier. In the presence of imbalanced data, these models tend to assign higher weights to misclassified instances from the minority class during the training process. This weighting scheme can result in an overemphasis on the minority class, potentially leading to misclassifications and biased decision boundaries (Okey et al., 2022).

    2. Loss function optimization: Boosting algorithms aim to minimize a loss function by iteratively fitting models to the training data. In imbalanced datasets, the loss function (Fernando and Tsokos, 2021) used may not adequately capture the cost of misclassifying the minority class. As a result, the models might prioritize minimizing the overall loss (Laradji et al., 2015), which is dominated by the majority class, leading to a bias towards the majority class and reduced performance on the minority class.

    3. Nearest neighbor-based approach: kNN algorithm makes predictions based on the class labels of its nearest neighbors. In the presence of imbalanced data, the sparsity (Padmaja et al., 2007) of the minority class can lead to situations where the nearest neighbors of a minority instance predominantly belong to the majority class. This can result in misclassifications and a tendency to favor the majority class during classification.

    The gradual improvements observed in the performance of these models as the data became more balanced are consistent with the notion that as the class distribution becomes more even, classifiers tend to achieve better results. This observation supports the idea that balancing techniques, such as ADASYN sampling, can alleviate the negative impact of class imbalance on the performance of classifiers.

    Overall, the findings of this study align with existing research on the performance of classification models in the presence of class imbalance. The superiority of RF and DT models, the challenges faced by LR, SGDC and NB models, and the performance improvements with data balancing techniques are consistent with previous findings. These results contribute to the growing body of knowledge on class imbalance and provide further evidence of the effectiveness of certain algorithms in imbalanced classification tasks.

    In conclusion, this study investigated the performance of various classification models in the presence of class imbalance. The results shed light on the impact of class distribution on the effectiveness of different algorithms and the importance of data balancing techniques. The findings highlight the outstanding performance of random forest and decision tree algorithms, which consistently outperformed other models in both training and testing stages. These models demonstrated robustness to class imbalance and achieved high Mathews' correlation coefficient scores even when trained on highly imbalanced datasets. The visual representations and area under the ROC curves further supported their superiority over other classifiers.

    On the other hand, logistic regression, stochastic gradient descent classifier, and naïve Bayes models exhibited poor performance regardless of the class imbalance in the data. These models struggled to handle imbalanced datasets and scored lower MCC values compared to other algorithms. The study also highlighted the performance improvements of models such as ADA, GB, XGB, LGBM, and kNN as the data became more balanced. These models showed increased predictive power and achieved higher MCC scores as the class distribution became more even. The results further emphasized the effectiveness of ADASYN sampling techniques in producing more reliable predictions compared to under-sampling techniques. The findings of this study align with prior research on imbalanced classification tasks, providing further evidence of the superiority of RF and DT models and the challenges faced by LR, SGDC, and NB models. The results contribute to the existing body of knowledge on class imbalance and highlight the importance of selecting appropriate algorithms and employing data balancing techniques for improved classification performance.

    Overall, this study emphasizes the need for careful consideration of the choice of classification models and the implementation of data balancing techniques when dealing with imbalanced datasets. The results can inform practitioners and researchers in selecting the most suitable models for imbalanced classification tasks and guide the development of more effective approaches to address class imbalance challenges. This thorough investigation provides a deeper comparison of class imbalance's pivotal influence on predictive analytics. Through assessing ten diverse classification models using robust evaluation metrics, such as ROC curve area, Mathews' correlation coefficient (MCC) and F1-scores, this study furnishes empirical insights into these models' strengths and weaknesses within imbalanced datasets. Guided by data-driven model assessments and balancing approaches, this research serves as a valuable roadmap for practitioners and researchers grappling with imbalanced datasets. The results provide explicit guidelines for model selection and tailored data balancing techniques, ultimately enhancing classification performance in the face of class imbalance.

    However, this study has certain limitations that should be acknowledged. First, the analysis was conducted using a specific dataset with its own characteristics, and the results may not generalize to other datasets or domains. Therefore, it is crucial to validate these findings on different datasets to ensure their applicability in diverse contexts. In our case, we have simulated eighteen (18) different samples. Second, the study focused solely on the performance of classification models and did not delve into the underlying reasons for the observed differences in performance. Future research could explore the specific factors contributing to the effectiveness or ineffectiveness of different models in handling class imbalance. This could involve examining feature importance, model interpretability, or identifying specific patterns in the data that affect model performance.

    This work is based on the research supported wholly/in part by the National Research Foundation of South Africa (Grant Number 126885). This work is based on research supported in part by the Department of Science and Innovation (DSI) of South Africa. The grant holder acknowledges that opinions, findings, and conclusions or recommendations expressed in any publication generated by DSI-supported research are those of the authors and that the DSI accepts no liability whatsoever in this regard.

    The authors declare they have not used Artificial Intelligence (AI) tools in the creation of this article, except for the machine learning models used for the analysis.

    I would like to express my deepest gratitude to my supervisor, Prof. Tanja Verster, for their unwavering guidance, invaluable guidance, and exceptional mentorship throughout the course of this research. I would also like to extend my gratitude to the NWU BMI Centre for availing the resources for me and the wonderful staff we have.

    The authors declare no conflict of interest.

    This section provides more results that were obtained from training and testing of the machine learning models as summarised by the Gini score (Table 6) and the recall measure (Table 7). The results also conformed with the findings obtained using MCC, F1-score and the AUROC measures in Section 5.Once again, decision tree and random forest models outperformed the rest of the models across various samples of varying class imbalance regardless of the sampling technique used.

    Table 6.  Training results by the gini score.
    Model Sample 1 Sample 2 Sample 3 Sample 4 Sample 5 Sample 6 Sample 7 Sample 8 Sample 9
    Under-sampling LR 0.0558 0.0744 0.0968 0.1288 0.1254 0.1744 0.1704 0.3076 0.3088
    DT 0.6574 0.6506 0.6652 0.6552 0.6602 0.6448 0.6756 0.6414 0.6502
    RF 0.6432 0.6698 0.7030 0.7088 0.7154 0.7136 0.7504 0.7276 0.7426
    ADA 0.2408 0.3064 0.3574 0.4054 0.4254 0.4312 0.4764 0.4986 0.5022
    GB 0.2682 0.34 0.3904 0.4276 0.465 0.4658 0.5056 0.534 0.5506
    XGB 0.3344 0.3936 0.4452 0.5074 0.5248 0.5232 0.9414 0.5784 0.6126
    LGBM 0.2856 0.357 0.4194 0.4734 0.4934 0.4976 0.5436 0.5498 0.5834
    SGDC 0.0266 0.0646 0.0824 0.1386 0.1094 0.1796 0.1768 0.2898 0.3068
    KNN 0.2746 0.323 0.3044 0.3594 0.3792 0.3592 0.352 0.3804 0.3748
    NB 0.033 0.0208 0.0238 0.027 0.0428 0.034 0.0312 0.032 0.0244
    ADASYN sampling LR 0.0544 0.0726 0.0898 0.1318 0.1394 0.1998 0.2438 0.2514 0.2804
    DT 0.7322 0.8368 0.8952 0.9164 0.9362 0.9476 0.9534 0.955 0.9578
    RF 0.7348 0.8616 0.9156 0.9424 0.9592 0.9708 0.9726 0.9768 0.9800
    ADA 0.2472 0.3202 0.3546 0.4036 0.4344 0.4518 0.4744 0.4914 0.5006
    GB 0.2716 0.3554 0.3906 0.4438 0.4682 0.4862 0.5104 0.516 0.5338
    XGB 0.3432 0.4298 0.4754 0.5178 0.5592 0.5778 0.5946 0.6026 0.6242
    LGBM 0.2892 0.381 0.4212 0.4686 0.5016 0.5168 0.5392 0.5454 0.5626
    SGDC 0 0.0126 0.0486 0.1168 0.1076 0.2414 0.2824 0.3076 0.311
    KNN 0.3122 0.4454 0.5638 0.64 0.7172 0.7624 0.7968 0.8236 0.843
    NB 0.0236 0.0226 0.0282 0.0302 0.0368 0.0406 0.0298 0.0316 0.0328

     | Show Table
    DownLoad: CSV
    Table 7.  Training and testing results by recall.
    Model Sample 1 sample 2 Sample 3 Sample 4 Sample 5 Sample 6 Sample 7 Sample 8 Sample 9
    Training Under-sampling LR 0.0613 0.0899 0.1304 0.1838 0.2603 0.3653 0.4479 0.5493 0.6505
    DT 0.9957 0.9987 0.9986 0.9995 0.9995 0.9999 0.9998 0.9999 0.9999
    RF 0.996 0.9989 0.9988 0.9995 0.9995 0.9999 0.9998 0.9999 0.9999
    ADA 0.2868 0.375 0.4438 0.5088 0.5526 0.6044 0.6424 0.6988 0.7434
    GB 0.3034 0.4183 0.4909 0.5481 0.5944 0.6394 0.6873 0.753 0.7894
    XGB 0.3999 0.507 0.5876 0.6403 0.6897 0.7322 0.7717 0.8142 0.8494
    LGBM 0.3264 0.4557 0.5279 0.5774 0.6328 0.6759 0.7209 0.7672 0.807
    SGDC 0.0152 0.0283 0.0897 0.1911 0.2922 0.372 0.4042 0.4393 0.5186
    KNN 0.4448 0.7002 0.8518 0.9336 0.9749 0.9918 0.9974 0.9994 0.9998
    NB 0.0271 0.0315 0.0369 0.0329 0.0343 0.0357 0.039 0.0453 0.0442
    ADASYN sampling LR 0.0629 0.0853 0.1348 0.1904 0.2653 0.3608 0.4547 0.5487 0.6561
    DT 0.6707 0.8811 0.962 0.9832 0.9928 0.9982 0.9994 0.9997 0.9999
    RF 0.6531 0.8801 0.9579 0.9827 0.9934 0.9986 0.9994 0.9997 0.9999
    ADA 0.2875 0.3687 0.4433 0.51 0.5583 0.606 0.6475 0.6948 0.7485
    GB 0.2985 0.4083 0.4919 0.5444 0.6002 0.6434 0.6937 0.7492 0.7916
    XGB 0.3302 0.4595 0.5568 0.6171 0.6761 0.727 0.7669 0.8031 0.8471
    LGBM 0.3036 0.4405 0.5194 0.571 0.6315 0.6758 0.7232 0.762 0.8102
    SGDC 0.0154 0.0288 0.0927 0.1949 0.2994 0.3703 0.415 0.4372 0.5213
    KNN 0.2811 0.5047 0.6958 0.8272 0.9172 0.9674 0.9876 0.9973 0.9993
    NB 0.026 0.0297 0.0346 0.0329 0.0354 0.0377 0.038 0.045 0.0441
    Testing Under-sampling LR 0.0525 0.0699 0.0762 0.0841 0.0978 0.0949 0.0976 0.0996 0.101
    DT 0.9972 0.9975 0.9969 0.9974 0.9979 0.9975 0.9976 0.9982 0.998
    RF 0.9969 0.997 0.9971 0.9974 0.9979 0.9971 0.9974 0.9984 0.9978
    ADA 0.2419 0.313 0.3514 0.372 0.3808 0.3857 0.3951 0.4001 0.4004
    GB 0.2576 0.3428 0.3884 0.3998 0.4218 0.4247 0.4336 0.4461 0.4465
    XGB 0.3569 0.4476 0.4768 0.5018 0.527 0.5282 0.5345 0.5475 0.5546
    LGBM 0.2789 0.3677 0.4191 0.4469 0.4628 0.4716 0.4784 0.493 0.4911
    SGDC 0.0149 0.0149 0.0014 0.023 0.0467 0.0507 0.0147 0.0491 0.0511
    KNN 0.2569 0.317 0.3387 0.3542 0.363 0.3588 0.3682 0.3782 0.3666
    NB 0.0278 0.0276 0.033 0.0336 0.0336 0.0356 0.0392 0.0339 0.0327
    ADASYN sampling LR 0.0474 0.072 0.0943 0.089 0.0979 0.1003 0.1022 0.1037 0.1026
    DT 0.3115 0.3724 0.3939 0.4327 0.4347 0.4062 0.4187 0.4269 0.4422
    RF 0.2448 0.3229 0.3687 0.4122 0.4101 0.4072 0.4152 0.4107 0.4328
    ADA 0.2409 0.3106 0.3495 0.3847 0.38 0.3892 0.3885 0.3865 0.3981
    GB 0.2423 0.3248 0.3756 0.4232 0.4091 0.417 0.4253 0.42 0.4373
    XGB 0.2483 0.3263 0.3766 0.4072 0.4061 0.4082 0.4152 0.4181 0.4417
    LGBM 0.2473 0.3283 0.3727 0.4292 0.4292 0.4286 0.4424 0.4368 0.4517
    SGDC 0.0138 0.0137 0.002 0.0285 0.0492 0.0458 0.0111 0.0469 0.0501
    KNN 0.151 0.2126 0.2379 0.2656 0.2485 0.2655 0.2386 0.2572 0.2395
    NB 0.0242 0.0265 0.041 0.035 0.0382 0.0326 0.0302 0.0296 0.0263

     | Show Table
    DownLoad: CSV


    Authors' contributions



    Conceptualization: Petros Galanis, Aglaia Katsiroumpa, Zoe Katsiroumpa; methodology: Petros Galanis, Aglaia Katsiroumpa, Zoe Katsiroumpa, Polyxeni Mangoulia, Evmorfia Koukia; software: Petros Galanis, Aglaia Katsiroumpa, Zoe Katsiroumpa, Parisis Gallos; validation: Zoe Katsiroumpa, Polyxeni Mangoulia, Parisis Gallos, Ioannis Moisoglou; formal analysis: Petros Galanis, Aglaia Katsiroumpa, Zoe Katsiroumpa, Parisis Gallos; resources: Petros Galanis, Zoe Katsiroumpa, Polyxeni Mangoulia, Parisis Gallos, Ioannis Moisoglou; data curation: Petros Galanis, Zoe Katsiroumpa, Polyxeni Mangoulia, Parisis Gallos, Ioannis Moisoglou; writing–original draft preparation: Petros Galanis, Aglaia Katsiroumpa, Zoe Katsiroumpa, Polyxeni Mangoulia, Parisis Gallos, Ioannis Moisoglou, Evmorfia Koukia; writing–review and editing: Petros Galanis, Aglaia Katsiroumpa, Zoe Katsiroumpa, Polyxeni Mangoulia, Parisis Gallos, Ioannis Moisoglou, Evmorfia Koukia; supervision: Petros Galanis.

    Conflict of interest



    Petros Galanis is an editorial board member for AIMS Public Health, and he is also a guest editor of AIMS Public Health Special Issue. Polyxeni Mangoulia and Ioannis Moisoglou are guest editors of AIMS Public Health Special Issues. They were not involved in the editorial review or the decision to publish this article. All authors declare that there are no competing interests.

    [1] StatistaSocial Media & User-generated content (2024). [cited 2025 February 01]. Available from: https://www.statista.com/statistics/278414/number-of-worldwide-social-network-users/
    [2] BacklinkoTikTok statistics you need to know (2024). [cited 2025 February 01]. Available from: https://backlinko.com/tiktok-users
    [3] Merchant RM, Lurie N (2020) Social media and emergency preparedness in response to novel coronavirus. JAMA 323: 2011-2012. https://doi.org/10.1001/jama.2020.4469
    [4] Helm PJ, Jimenez T, Galgali MS, et al. (2022) Divergent effects of social media use on meaning in life via loneliness and existential isolation during the coronavirus pandemic. J Soc Pers Relat 39: 1768-1793. https://doi.org/10.1177/02654075211066922
    [5] Khan MN, Ashraf MA, Seinen D, et al. (2021) Social media for knowledge acquisition and dissemination: The impact of the COVID-19 pandemic on collaborative learning driven social media adoption. Front Psychol 12: 648253. https://doi.org/10.3389/fpsyg.2021.648253
    [6] Rosen AO, Holmes AL, Balluerka N, et al. (2022) Is social media a new type of social support? Social media use in spain during the COVID-19 pandemic: A mixed methods study. Int J Environ Res Public Health 19: 3952. https://doi.org/10.3390/ijerph19073952
    [7] Yang JZ, Liu Z, Wong JC (2022) Information seeking and information sharing during the COVID-19 pandemic. Commun Q 70: 1-21. https://doi.org/10.1080/01463373.2021.1995772
    [8] Szeto S, Au AKY, Cheng SKL (2024) Support from social media during the COVID-19 Pandemic: A systematic review. Behav Sci (Basel) 14: 759. https://doi.org/10.3390/bs14090759
    [9] Hmidan A, Seguin D, Duerden EG (2023) Media screen time use and mental health in school aged children during the pandemic. BMC Psychol 11: 202. https://doi.org/10.1186/s40359-023-01240-0
    [10] Seguin D, Kuenzel E, Morton JB, et al. (2021) School's out: Parenting stress and screen time use in school-age children during the COVID-19 pandemic. J Affect Disord Rep 6: 100217. https://doi.org/10.1016/j.jadr.2021.100217
    [11] De Rosis S, Lopreite M, Puliga M, et al. (2021) The early weeks of the Italian Covid-19 outbreak: Sentiment insights from a Twitter analysis. Health Policy 125: 987-994. https://doi.org/10.1016/j.healthpol.2021.06.006
    [12] Lopreite M, Panzarasa P, Puliga M, et al. (2021) Early warnings of COVID-19 outbreaks across Europe from social media. Sci Rep 11: 2147. https://doi.org/10.1038/s41598-021-81333-1
    [13] Tsao SF, Chen H, Tisseverasinghe T, et al. (2021) What social media told us in the time of COVID-19: A scoping review. Lancet Digit Health 3: e175-e194. https://doi.org/10.1016/S2589-7500(20)30315-0
    [14] Haenlein M, Anadol E, Farnsworth T, et al. (2020) Navigating the new era of influencer marketing: How to be successful on Instagram, TikTok, & Co. Calif Manage Rev 63: 5-25. https://doi.org/10.1177/0008125620958166
    [15] Pedrouzo S, Krynski L (2023) Hyperconnected: Children and adolescents on social media. The TikTok phenomenon. Arch Argent Pediatr 121: e202202674. https://doi.org/10.5546/aap.2022-02674.eng
    [16] Grandinetti J (2023) Examining embedded apparatuses of AI in Facebook and TikTok. AI & Soc 38: 1273-1286. https://doi.org/10.1007/s00146-021-01270-5
    [17] Kang H, Lou C (2022) AI agency vs. human agency: understanding human–AI interactions on TikTok and their implications for user engagement. J Comput-Mediat Comm 27: zmac014. https://doi.org/10.1093/jcmc/zmac014
    [18] Montag C, Yang H, Elhai JD (2021) On the psychology of TikTok use: A first glimpse from empirical findings. Front Public Health 9: 641673. https://doi.org/10.3389/fpubh.2021.641673
    [19] TikTokMental and behavioral health (2024). [cited 2025 February 01]. Available from: https://www.tiktok.com/community-guidelines/en/mental-behavioral-health?lang=en
    [20] Lookingbill V (2022) Examining nonsuicidal self-injury content creation on TikTok through qualitative content analysis. Libr Inform Sci Res 44: 101199. https://doi.org/10.1016/j.lisr.2022.101199
    [21] Center for Countering Digital HateTikTok pushes harmful content promoting eating disorders and self-harm into users' feeds (2022). [cited 2025 February 01]. Available from: https://counterhate.com/research/deadly-by-design/
    [22] Karizat N, Delmonaco D, Eslami M, et al. (2021) Algorithmic folk theories and identity: How TikTok users co-produce knowledge of identity and engage in algorithmic resistance. Proc ACM Hum-Comput Interact 5: 1-44. https://doi.org/10.1145/3476046
    [23] Abbouyi S, Bouazza S, El Kinany S, et al. (2024) Depression and anxiety and its association with problematic social media use in the MENA region: A systematic review. Egypt J Neurol Psychiatry Neurosurg 60: 15. https://doi.org/10.1186/s41983-024-00793-0
    [24] Ahmed O, Walsh E, Dawel A, et al. (2024) Social media use, mental health and sleep: A systematic review with meta-analyses. J Affect Disord 367: 701-712. https://doi.org/10.1016/j.jad.2024.08.193
    [25] Alonzo R, Hussain J, Stranges S, et al. (2021) Interplay between social media use, sleep quality, and mental health in youth: A systematic review. Sleep Med Rev 56: 101414. https://doi.org/10.1016/j.smrv.2020.101414
    [26] Baker DA, Algorta GP (2016) The relationship between online social networking and depression: A systematic review of quantitative studies. Cyberpsychol Behav Soc Netw 19: 638-648. https://doi.org/10.1089/cyber.2016.0206
    [27] Casale S, Banchi V (2020) Narcissism and problematic social media use: A systematic literature review. Addict Behav Rep 11: 100252. https://doi.org/10.1016/j.abrep.2020.100252
    [28] Marino C, Gini G, Vieno A, et al. (2018) The associations between problematic Facebook use, psychological distress and well-being among adolescents and young adults: A systematic review and meta-analysis. J Affect Disord 226: 274-281. https://doi.org/10.1016/j.jad.2017.10.007
    [29] Shannon H, Bush K, Villeneuve PJ, et al. (2022) Problematic social media use in adolescents and young adults: Systematic review and meta-analysis. JMIR Ment Health 9: e33450. https://doi.org/10.2196/33450
    [30] Wu W, Huang L, Yang F (2024) Social anxiety and problematic social media use: A systematic review and meta-analysis. Addict Behav 153: 107995. https://doi.org/10.1016/j.addbeh.2024.107995
    [31] Amin S, Iftikhar A, Meer A (2022) Intervening effects of academic performance between TikTok obsession and psychological wellbeing challenges in university students. Online Media Soc 3: 244-255. https://doi.org/10.71016/oms/qy5har60
    [32] Hendrikse C, Limniou M (2024) The use of Instagram and TikTok in relation to problematic use and well-Being. J Technol Behav Sci 9: 846-857. https://doi.org/10.1007/s41347-024-00399-6
    [33] Rogowska AM, Cincio A (2024) Procrastination mediates the relationship between problematic TikTok use and depression among young adults. J Clin Med 13: 1247. https://doi.org/10.3390/jcm13051247
    [34] Williams M, Lewin KM, Meshi D (2024) Problematic use of five different social networking sites is associated with depressive symptoms and loneliness. Curr Psychol 43: 20891-20898. https://doi.org/10.1007/s12144-024-05925-6
    [35] Conte G, Iorio GD, Esposito D, et al. (2024) Scrolling through adolescence: A systematic review of the impact of TikTok on adolescent mental health. Eur Child Adolesc Psychiatry 16. https://doi.org/10.1007/s00787-024-02581-w
    [36] Walker E, Hernandez AV, Kattan MW (2008) Meta-analysis: Its strengths and limitations. Cleve Clin J Med 75: 431-439. https://doi.org/10.3949/ccjm.75.6.431
    [37] Egger M, Smith GD, Phillips AN (1997) Meta-analysis: Principles and procedures. BMJ 315: 1533-1537. https://doi.org/10.1136/bmj.315.7121.1533
    [38] Finckh A, Tramèr MR (2008) Primer: strengths and weaknesses of meta-analysis. Nat Rev Rheumatol 4: 146-152. https://doi.org/10.1038/ncprheum0732
    [39] Ioannidis JPA, Lau J (1999) Pooling research results: Benefits and limitations of meta-Analysis. Jt Comm J Qual Improv 25: 462-469. https://doi.org/10.1016/S1070-3241(16)30460-6
    [40] Csikszentmihalyi M (2002) Flow: the classic work on how to achieve happiness. London, Rider: Random House.
    [41] Montag C, Lachmann B, Herrlich M, et al. (2019) Addictive features of social media/messenger platforms and freemium games against the background of psychological and economic theories. Int J Environ Res Public Health 16: 2612. https://doi.org/10.3390/ijerph16142612
    [42] Andreassen CS, Billieux J, Griffiths MD, et al. (2016) The relationship between addictive use of social media and video games and symptoms of psychiatric disorders: A large-scale cross-sectional study. Psychol Addict Behav 30: 252-262. https://doi.org/10.1037/adb0000160
    [43] Andreassen CS, Torsheim T, Brunborg GS, et al. (2012) Development of a Facebook addiction scale. Psychol Rep 110: 501-517. https://doi.org/10.2466/02.09.18.PR0.110.2.501-517
    [44] Griffiths M (2005) A ‘components' model of addiction within a biopsychosocial framework. J Subst Abuse 10: 191-197. https://doi.org/10.1080/14659890500114359
    [45] Sun Y, Zhang Y (2021) A review of theories and models applied in studies of social media addiction and implications for future research. Addict Behav 114: 106699. https://doi.org/10.1016/j.addbeh.2020.106699
    [46] Varona MN, Muela A, Machimbarrena JM (2022) Problematic use or addiction? A scoping review on conceptual and operational definitions of negative social networking sites use in adolescents. Addict Behav 134: 107400. https://doi.org/10.1016/j.addbeh.2022.107400
    [47] Alonso J, Liu Z, Evans-Lacko S, et al. (2018) Treatment gap for anxiety disorders is global: Results of the world mental health surveys in 21 countries. Depress Anxiety 35: 195-208. https://doi.org/10.1002/da.22711
    [48] Kazdin AE (2000) Encyclopedia of psychology. Washington, DC: American Psychological Association 145-152.
    [49] Carstens E, Moberg GP (2000) Recognizing pain and distress in laboratory animals. ILAR J 41: 62-71. https://doi.org/10.1093/ilar.41.2.62
    [50] (2013) American Psychiatric AssociationDiagnostic and statistical manual of mental disorders: DSM-5. Washington: American Psychiatric Association 329-354. https://doi.org/10.1176/appi.books.9780890425596
    [51] Della Sala S (2021) Encyclopedia of behavioral neuroscience. San Diego: Elsevier Science & Technology 552-557.
    [52] Pavlova M, Latreille V (2019) Sleep disorders. Am J Med 132: 292-299. https://doi.org/10.1016/j.amjmed.2018.09.021
    [53] Peplau LA, Perlman D (1982) Loneliness: A sourcebook of current theory, research and therapy. New York: Wiley 1-18.
    [54] Ellis L, Hoskin A, Ratnasingam M (2018) Handbook of social status correlates. US: Elsevier 1-14.
    [55] Miller DN (2011) Life satisfaction. Encyclopedia of Child Behavior and Development . Boston, MA: Springer US 887-889. https://doi.org/10.1007/978-0-387-79061-9_1659
    [56] Huppert FA (2009) Psychological well-being: Evidence regarding its causes and consequences. Appl Psych Health Well 1: 137-164. https://doi.org/10.1111/j.1758-0854.2009.01008.x
    [57] Moher D, Shamseer L, Clarke M, et al. (2015) Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev 4: 1. https://doi.org/10.1186/2046-4053-4-1
    [58] Park CU, Kim HJ (2015) Measurement of inter-rater reliability in systematic review. Hanyang Med Rev 35: 44. https://doi.org/10.7599/hmr.2015.35.1.44
    [59] McHugh ML (2012) Interrater reliability: The kappa statistic. Biochem Med (Zagreb) 22: 276-282. https://doi.org/10.11613/BM.2012.031
    [60] Santos WM dos, Secoli SR, Püschel VA de A (2018) The Joanna Briggs institute approach for systematic reviews. Rev Lat Am Enfermagem 26: e3074. https://doi.org/10.1590/1518-8345.2885.3074
    [61] Nieminen P (2022) Application of standardized regression coefficient in meta-analysis. BioMedInformatics 2: 434-458. https://doi.org/10.3390/biomedinformatics2030028
    [62] Vittinghoff E (2005) Regression methods in biostatistics: Linear, logistic, survival, and repeated measures models. New York: Springer 65-82.
    [63] Higgins J (2003) Measuring inconsistency in meta-analyses. BMJ 327: 557-560. https://doi.org/10.1136/bmj.327.7414.557
    [64] Wallace BC, Schmid CH, Lau J, et al. (2009) Meta-analyst: Software for meta-analysis of binary, continuous and diagnostic data. BMC Med Res Methodol 9: 80. https://doi.org/10.1186/1471-2288-9-80
    [65] Al-Garni AM, Alamri HS, Asiri WMA, et al. (2024) Social media use and sleep quality among secondary school students in Aseer region: A cross-sectional study. Int J Gen Med 17: 3093-3106. https://doi.org/10.2147/IJGM.S464457
    [66] Asad K, Ali F, Awais M (2022) Personality traits, narcissism and TikTok addiction: A parallel mediation approach. Int J Media Inf Lit 7: 293-304. https://doi.org/10.13187/ijmil.2022.2.293
    [67] Sarman A, Tuncay S (2023) The relationship of Facebook, Instagram, Twitter, TikTok and WhatsApp/Telegram with loneliness and anger of adolescents living in Turkey: A structural equality model. J Pediatr Nurs 72: 16-25. https://doi.org/10.1016/j.pedn.2023.03.017
    [68] Sha P, Dong X (2021) Research on adolescents regarding the indirect effect of depression, anxiety, and stress between TikTok use disorder and memory loss. Int J Environ Res Public Health 18: 8820. https://doi.org/10.3390/ijerph18168820
    [69] Yang Y, Adnan H, Sarmiti N (2023) The relationship between anxiety and TikTok addiction among university students in China: Mediated by escapism and use intensity. Int J Media Inf Lit 8: 458-464. https://doi.org/10.13187/ijmil.2023.2.458
    [70] Yao N, Chen J, Huang S, et al. (2023) Depression and social anxiety in relation to problematic TikTok use severity: The mediating role of boredom proneness and distress intolerance. Comput Hum Behav 145: 107751. https://doi.org/10.1016/j.chb.2023.107751
    [71] López-Gil JF, Chen S, Jiménez-López E, et al. (2023) Are the use and addiction to social networks associated with disordered eating among adolescents? Findings from the EHDLA study. Int J Ment Health Addict 22: 3775-3789. https://doi.org/10.1007/s11469-023-01081-3
    [72] Masciantonio A, Bourguignon D, Bouchat P, et al. (2021) Don't put all social network sites in one basket: Facebook, Instagram, Twitter, TikTok, and their relations with well-being during the COVID-19 pandemic. PLoS One 16: e0248384. https://doi.org/10.1371/journal.pone.0248384
    [73] Landa-Blanco M, García YR, Landa-Blanco AL, et al. (2024) Social media addiction relationship with academic engagement in university students: The mediator role of self-esteem, depression, and anxiety. Heliyon 10: e24384. https://doi.org/10.1016/j.heliyon.2024.e24384
    [74] Sagrera CE, Magner J, Temple J, et al. (2022) Social media use and body image issues among adolescents in a vulnerable Louisiana community. Front Psychiatry 13: 1001336. https://doi.org/10.3389/fpsyt.2022.1001336
    [75] Blackburn MR, Hogg RC (2024) #ForYou? the impact of pro-ana TikTok content on body image dissatisfaction and internalisation of societal beauty standards. PLoS One 19: e0307597. https://doi.org/10.1371/journal.pone.0307597
    [76] Nasidi QY, Norde AB, Dahiru JM, et al. (2024) Tiktok usage, social comparison, and self-esteem among the youth: Moderating role of gender. Galactica Media 6: 121-137. https://doi.org/10.46539/gmd.v6i2.467
    [77] Hartanto A, Quek FYX, Tng GYQ, et al. (2021) Does social media use increase depressive symptoms? A reverse causation perspective. Front Psychiatry 12: 641934. https://doi.org/10.3389/fpsyt.2021.641934
    [78] Van Zalk N (2016) Social anxiety moderates the links between excessive chatting and compulsive internet use. Cyberpsychology 10: 3. https://doi.org/10.5817/CP2016-3-3
    [79] Adams SK, Kisler TS (2013) Sleep quality as a mediator between technology-related sleep quality, depression, and anxiety. Cyberpsychol Behav Soc Netw 16: 25-30. https://doi.org/10.1089/cyber.2012.0157
    [80] Liu S, Wing YK, Hao Y, et al. (2019) The associations of long-time mobile phone use with sleep disturbances and mental distress in technical college students: A prospective cohort study. Sleep 42. https://doi.org/10.1093/sleep/zsy213
    [81] Caplan SE (2002) Problematic internet use and psychosocial well-being: Development of a theory-based cognitive–behavioral measurement instrument. Comput Hum Behav 18: 553-575. https://doi.org/10.1016/S0747-5632(02)00004-3
    [82] Sanders CE, Field TM, Diego M, et al. (2000) The relationship of Internet use to depression and social isolation among adolescents. Adolescence 35: 237-242.
    [83] Davis RA (2001) A cognitive-behavioral model of pathological internet use. Comput Hum Behav 17: 187-195. https://doi.org/10.1016/S0747-5632(00)00041-8
    [84] Morahan-Martin J, Schumacher P (2003) Loneliness and social uses of the internet. Comput Hum Behav 19: 659-671. https://doi.org/10.1016/S0747-5632(03)00040-2
    [85] Kross E, Verduyn P, Demiralp E, et al. (2013) Facebook use predicts declines in subjective well-being in young adults. PLoS One 8: e69841. https://doi.org/10.1371/journal.pone.0069841
    [86] Chou HT, Edge N (2012) “They are happier and having better lives than I am”: The impact of using Facebook on perceptions of others' lives. Cyberpsychol Behav Soc Netw 15: 117-121. https://doi.org/10.1089/cyber.2011.0324
    [87] Sagioglou C, Greitemeyer T (2014) Facebook's emotional consequences: Why Facebook causes a decrease in mood and why people still use it. Comput Hum Behav 35: 359-363. https://doi.org/10.1016/j.chb.2014.03.003
    [88] Primack BA, Shensa A, Sidani JE, et al. (2017) Social media use and perceived social isolation among young adults in the U.S. Am J Prev Med 53: 1-8. https://doi.org/10.1016/j.amepre.2017.01.010
    [89] Tandoc EC, Ferrucci P, Duffy M (2015) Facebook use, envy, and depression among college students: Is facebooking depressing?. Comput Hum Behav 43: 139-146. https://doi.org/10.1016/j.chb.2014.10.053
    [90] Smith RH, Kim SH (2007) Comprehending envy. Psychol Bull 133: 46-64. https://doi.org/10.1037/0033-2909.133.1.46
    [91] Block JJ (2008) Issues for DSM-V: Internet addiction. Am J Psychiatry 165: 306-307. https://doi.org/10.1176/appi.ajp.2007.07101556
    [92] Morrison CM, Gore H (2010) The relationship between excessive internet use and depression: A questionnaire-based study of 1,319 young people and adults. Psychopathology 43: 121-126. https://doi.org/10.1159/000277001
    [93] Meier EP, Gray J (2014) Facebook photo activity associated with body image disturbance in adolescent girls. Cyberpsychol Behav Soc Netw 17: 199-206. https://doi.org/10.1089/cyber.2013.0305
    [94] Wang R, Yang F, Haigh MM (2017) Let me take a selfie: Exploring the psychological effects of posting and viewing selfies and groupies on social media. Telemat Inform 34: 274-283. https://doi.org/10.1016/j.tele.2016.07.004
    [95] O'Keeffe GS, Clarke-Pearson K (2011) The impact of social media on children, adolescents, and families. Pediatrics 127: 800-804. https://doi.org/10.1542/peds.2011-0054
    [96] Qin Y, Musetti A, Omar B (2023) Flow experience is a key factor in the likelihood of adolescents' problematic TikTok use: The moderating role of active parental mediation. Int J Environ Res Public Health 20: 2089. https://doi.org/10.3390/ijerph20032089
    [97] Qin Y, Omar B, Musetti A (2022) The addiction behavior of short-form video app TikTok: The information quality and system quality perspective. Front Psychol 13: 932805. https://doi.org/10.3389/fpsyg.2022.932805
    [98] Elhai JD, Yang H, Fang J, et al. (2020) Depression and anxiety symptoms are related to problematic smartphone use severity in Chinese young adults: Fear of missing out as a mediator. Addict Behav 101: 105962. https://doi.org/10.1016/j.addbeh.2019.04.020
    [99] Pan W, Mu Z, Zhao Z, et al. (2023) Female users' TikTok use and body image: Active versus passive use and social comparison processes. Cyberpsychol Behav Soc Netw 26: 3-10. https://doi.org/10.1089/cyber.2022.0169
    [100] Geng Y, Gu J, Wang J, et al. (2021) Smartphone addiction and depression, anxiety: The role of bedtime procrastination and self-control. J Affect Disord 293: 415-421. https://doi.org/10.1016/j.jad.2021.06.062
    [101] Feng Y, Meng D, Guo J, et al. (2022) Bedtime procrastination in the relationship between self-control and depressive symptoms in medical students: From the perspective of sex differences. Sleep Med 95: 84-90. https://doi.org/10.1016/j.sleep.2022.04.022
    [102] Ha JH, Kim SY, Bae SC, et al. (2007) Depression and Internet addiction in adolescents. Psychopathology 40: 424-430. https://doi.org/10.1159/000107426
    [103] LeBourgeois MK, Hale L, Chang AM, et al. (2017) Digital media and sleep in childhood and adolescence. Pediatrics 140: S92-S96. https://doi.org/10.1542/peds.2016-1758J
    [104] Zubair U, Khan MK, Albashari M (2023) Link between excessive social media use and psychiatric disorders. Ann Med Surg (Lond) 85: 875-878. https://doi.org/10.1097/MS9.0000000000000112
    [105] El Asam A, Samara M, Terry P (2019) Problematic internet use and mental health among British children and adolescents. Addict Behav 90: 428-436. https://doi.org/10.1016/j.addbeh.2018.09.007
    [106] Kuss D, Griffiths M, Karila L, et al. (2014) Internet addiction: A systematic review of epidemiological research for the last decade. Curr Pharm Des 20: 4026-4052. https://doi.org/10.2174/13816128113199990617
    [107] Fan R (2023) The impact of TikTok short videos on anxiety level of juveniles in Shenzhen China. Proceedings of the 2022 International Conference on Science Education and Art Appreciation (SEAA 2022) . Paris: Atlantis Press SARL 535-542. https://doi.org/10.2991/978-2-494069-05-3_66
    [108] Mironica A, Popescu CA, George D, et al. (2024) Social media influence on body image and cosmetic surgery considerations: A systematic review. Cureus 16: e65626. https://doi.org/10.7759/cureus.65626
    [109] Vincente-Benito I, Ramírez-Durán MDV (2023) Influence of social media use on body image and well-being among adolescents and young adults: A systematic review. J Psychosoc Nurs Ment Health Serv 61: 11-18. https://doi.org/10.3928/02793695-20230524-02
    [110] Galanis P, Katsiroumpa A, Moisoglou I, et al. (2024) The TikTok addiction scale: Development and validation. AIMS Public Health 11: 1172-1197. https://doi.org/10.3934/publichealth.2024061
    [111] Kwon M, Kim DJ, Cho H, et al. (2013) The smartphone addiction scale: Development and validation of a short version for adolescents. PLoS One 8: e83558. https://doi.org/10.1371/journal.pone.0083558
    [112] Chen IH, Strong C, Lin YC, et al. (2020) Time invariance of three ultra-brief internet-related instruments: Smartphone Application-Based Addiction Scale (SABAS), Bergen Social Media Addiction Scale (BSMAS), and the nine-item Internet Gaming Disorder Scale- Short Form (IGDS-SF9) (Study Part B). Addict Behav 101: 105960. https://doi.org/10.1016/j.addbeh.2019.04.018
    [113] Luo T, Qin L, Cheng L, et al. (2021) Determination the cut-off point for the Bergen social media addiction (BSMAS): Diagnostic contribution of the six criteria of the components model of addiction for social media disorder. J Behav Addict 10: 281-290. https://doi.org/10.1556/2006.2021.00025
    [114] Zarate D, Hobson BA, March E, et al. (2023) Psychometric properties of the Bergen social media addiction scale: An analysis using item response theory. Addict Behav Rep 17: 100473. https://doi.org/10.1016/j.abrep.2022.100473
    [115] Servidio R, Griffiths MD, Di Nuovo S, et al. (2023) Further exploration of the psychometric properties of the revised version of the Italian smartphone addiction scale–short version (SAS-SV). Curr Psychol 42: 27245-27258. https://doi.org/10.1007/s12144-022-03852-y
    [116] Galanis P, Katsiroumpa A, Moisoglou I, et al. Determining an optimal cut-off point for TikTok addiction using the TikTok addiction scale. [Preprint] (2025). https://doi.org/10.21203/rs.3.rs-4782800/v1
  • publichealth-12-02-027-s001.pdf
  • This article has been cited by:

    1. Lindani Dube, Tanja Verster, Interpretability of the random forest model under class imbalance, 2024, 4, 2769-2140, 446, 10.3934/DSFE.2024019
    2. Albert Verasius Dian Sano, Faqir M. Bhatti, Eka Miranda, Mediana Aryuni, Alfi Yusrotis Zakiyyah, Charles Bernando, The Impact of Augmentation and SMOTE Implementation on the Classification Models Performance: A Case Study on Student Academic Performance Dataset, 2024, 245, 18770509, 282, 10.1016/j.procs.2024.10.253
    3. Shotaro Mizuno, Tsubura Noda, Kaoru Mogushi, Takeshi Hase, Yoritsugu Iida, Katsuyuki Takeuchi, Yasuyoshi Ishiwata, Shinichi Uchida, Masashi Nagata, Prediction of Vancomycin-Associated Nephrotoxicity Based on the Area under the Concentration–Time Curve of Vancomycin: A Machine Learning Analysis, 2024, 47, 0918-6158, 1946, 10.1248/bpb.b24-00506
    4. Mohammed Alenezi, Fatih Anayi, Michael Packianather, Mokhtar Shouran, Enhancing Transformer Protection: A Machine Learning Framework for Early Fault Detection, 2024, 16, 2071-1050, 10759, 10.3390/su162310759
    5. Ying Li, Keyue Yan, Prediction of bank credit customers churn based on machine learning and interpretability analysis, 2025, 5, 2769-2140, 19, 10.3934/DSFE.2025002
    6. Selma Remidi, Abdelhak Boutaleb, Salah Eddine Tachi, Yacine Hasnaoui, Robert Szczepanek, Abderraouf Seffari, Ensemble machine learning model for exploration and targeting of Pb-Zn deposits in Algeria, 2025, 18, 1865-0473, 10.1007/s12145-025-01718-y
    7. Jintong Han, Adrian Chong, Hybrid-CGAN: A Hybrid Approach Combining Simulation and Generative Adversarial Networks for Fault Detection and Diagnosis in Buildings, 2025, 03787788, 115531, 10.1016/j.enbuild.2025.115531
    8. Molham Sakkal, Abdallah Abou Hajal, Machine learning predictions of tumor progression: How reliable are we?, 2025, 191, 00104825, 110156, 10.1016/j.compbiomed.2025.110156
    9. Miguel García-Torres, Francisco Saucedo, Federico Divina, Santiago Gómez-Guerrero, RFMSU: A multivariate symmetrical uncertainty-based random forest, 2026, 169, 00313203, 111939, 10.1016/j.patcog.2025.111939
    10. Neha Kumari, Suvendu Kanungo, Jit Mukherjee, 2025, Chapter 5, 978-3-031-94120-7, 57, 10.1007/978-3-031-94121-4_5
    11. Soukaina Hakkal, Ayoub Ait Lahcen, Leveraging graph neural network for learner performance prediction, 2025, 293, 09574174, 128724, 10.1016/j.eswa.2025.128724
  • Reader Comments
  • © 2025 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(1577) PDF downloads(138) Cited by(0)

Figures and Tables

Figures(3)  /  Tables(3)

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog