A Survey of techniques for fine-grained web traffic identification and classification

Xiaolin Gui; Yuanlong Cao; Ilsun You; Lejun Ji; Yong Luo; Zhenzhen Luo; Xiaolin Gui; Yuanlong Cao; Ilsun You; Lejun Ji; Yong Luo; Zhenzhen Luo

doi:10.3934/mbe.2022138

Mathematical Biosciences and Engineering

2022, Volume 19, Issue 3: 2996-3021. doi: 10.3934/mbe.2022138

Previous Article Next Article

Survey Special Issues

A Survey of techniques for fine-grained web traffic identification and classification

1.
School of software, Jiangxi Normal University, Nanchang 330027, China
2.
Department of Information Security Engineering, Soonchunhyang University, Asan 31538, South Korea

Academic Editor: Thippa Reddy Gadekallu

Received: 30 September 2021 Revised: 17 December 2021 Accepted: 27 December 2021 Published: 17 January 2022

After decades of rapid development, the scale and complexity of modern networks have far exceed our expectations. In many conditions, traditional traffic identification methods cannot meet the demand of modern networks. Recently, fine-grained network traffic identification has been proved to be an effective solution for managing network resources. There is a massive increase in the use of fine-grained network traffic identification in the communications industry. In this article, we propose a comprehensive overview of fine-grained network traffic identification. Then, we conduct a detailed literature review on fine-grained network traffic identification from three perspectives: wired network, mobile network, and malware traffic identification. Finally, we also draw the conclusion on the challenges of fine-grained network traffic identification and future research prospects.

Keywords:

Citation: Xiaolin Gui, Yuanlong Cao, Ilsun You, Lejun Ji, Yong Luo, Zhenzhen Luo. A Survey of techniques for fine-grained web traffic identification and classification[J]. Mathematical Biosciences and Engineering, 2022, 19(3): 2996-3021. doi: 10.3934/mbe.2022138

Related Papers:

[1]	Fugeng Zeng, Yao Huang, Peng Shi . Initial boundary value problem for a class of $p$ -Laplacian equations with logarithmic nonlinearity. Mathematical Biosciences and Engineering, 2021, 18(4): 3957-3976. doi: 10.3934/mbe.2021198
[2]	Xiaolin Fan, Tingting Xue, Yongsheng Jiang . Dirichlet problems of fractional $p$ -Laplacian equation with impulsive effects. Mathematical Biosciences and Engineering, 2023, 20(3): 5094-5116. doi: 10.3934/mbe.2023236
[3]	Luís P. Castro, Anabela S. Silva . On the solution and Ulam-Hyers-Rassias stability of a Caputo fractional boundary value problem. Mathematical Biosciences and Engineering, 2022, 19(11): 10809-10825. doi: 10.3934/mbe.2022505
[4]	Peng Shi, Min Jiang, Fugeng Zeng, Yao Huang . Initial boundary value problem for fractional $p$ -Laplacian Kirchhoff type equations with logarithmic nonlinearity. Mathematical Biosciences and Engineering, 2021, 18(3): 2832-2848. doi: 10.3934/mbe.2021144
[5]	Tong Zhang, Hiroshi Nishiura . COVID-19 cases with a contact history: A modeling study of contact history-stratified data in Japan. Mathematical Biosciences and Engineering, 2023, 20(2): 3661-3676. doi: 10.3934/mbe.2023171
[6]	Pshtiwan Othman Mohammed, Donal O'Regan, Dumitru Baleanu, Y. S. Hamed, Ehab E. Elattar . Analytical results for positivity of discrete fractional operators with approximation of the domain of solutions. Mathematical Biosciences and Engineering, 2022, 19(7): 7272-7283. doi: 10.3934/mbe.2022343
[7]	Diguo Zhai, Xinqi Bao, Xi Long, Taotao Ru, Guofu Zhou . Precise detection and localization of R-peaks from ECG signals. Mathematical Biosciences and Engineering, 2023, 20(11): 19191-19208. doi: 10.3934/mbe.2023848
[8]	Shihe Xu, Junde Wu . Qualitative analysis of a time-delayed free boundary problem for tumor growth with angiogenesis and Gibbs-Thomson relation. Mathematical Biosciences and Engineering, 2019, 16(6): 7433-7446. doi: 10.3934/mbe.2019372
[9]	Huy Tuan Nguyen, Nguyen Van Tien, Chao Yang . On an initial boundary value problem for fractional pseudo-parabolic equation with conformable derivative. Mathematical Biosciences and Engineering, 2022, 19(11): 11232-11259. doi: 10.3934/mbe.2022524
[10]	Wei Zhang, Jifeng Zhang, Jinbo Ni . Existence and uniqueness results for fractional Langevin equations on a star graph. Mathematical Biosciences and Engineering, 2022, 19(9): 9636-9657. doi: 10.3934/mbe.2022448

Abstract

1. Introduction

Climate anomaly derived from excessive carbon emission is an important issue related to the sustainable development of all aspects of human society, including politics, economy ^[1,2,3,4], environment and humanities ^{[5,6,7,8,9,10]}. The emission of greenhouse gas (GHG) is a strong contributor to the global warming ^[11]. The 2009 international energy agency (IEA) report showed that the total CO₂ emissions from China surpassed that of the United States in 2007 and occupied the first place. In 2017, 30 percentage of global CO₂ emission came from China ^[12]. In order to effectively control CO₂ emission, the Chinese government is aware of the seriousness of uncontrolled CO₂ emission, has written sustainable development into its national policy, and will continue to do so for a long time ^[13]. At the 2009 conference of UN climate change held in Copenhagen, China announced that, compared to the level of 2005, its CO₂ emission per unit of gross domestic product (GDP) would be reduced by 40–45% at the end of 2020. At the Paris Climate Conference in 2015, it promised to further reduce this indicator by 60% to 65% by 2030 ^[14].

Agriculture is the lifeblood of human social development, especially for those countries with huge populations, as well as an important source of greenhouse gases. According to the intergovernmental panel on climate change (IPCC) regulations on land utilization, the agricultural sector is responsible for 23% of CO₂ emission from this ecosystem (IPCC 2019). CO₂ emission from agriculture is the second largest source of global greenhouse gas emission, contributing about 14% of global anthropogenic greenhouse gas emission and 58% of non-anthropogenic CO₂ emission. Chinese agricultural CO₂ emission account for 17% of the total greenhouse gas carbon emissions, and the annual average growth rate of emission level is about 5% ^[15,16]. Therefore, the research on agricultural CO₂ emission can not only serve the sustainable development of the country, but also be conducive to environmental protection.

So far, various problems caused by agricultural CO₂ emission have attracted the attention of many research institutions and scholars. For example, Huang et al. ^[17] studied the temporal and spatial changes of Chinese agricultural carbon emission in the period of 1997 to 2016, and pointed that Chinese agricultural CO₂ emission generally increased, but its intensity weakened. They also found that the CO₂ emission and CO₂ emission intensity of different provinces were quite different, and the spatial correlation of CO₂ emission was obvious, which showed a high aggregation in general. Balsalobre-Lorente et al. ^[18] analyzed whether agricultural activities in BRICS countries (Brazil, Russia, India, China, and South Africa) induced carbon emission, and found that an inverted U-shaped trend was evolved between CO₂ emission and economic development, while the interaction between electricity consumption and agricultural activities had the additional deleterious effects on the environment. Wu et al. ^[19] explored the spatial characteristics and spillover effects of Chinese agricultural CO₂ emission intensity in the period of 2001 to 2018, and discovered that the provincial emission intensity of agricultural CO₂ showed a declining trend, but there was obvious heterogeneity in the rate of decline in different provinces. They also pointed that the spatial dependence of emission intensity for Chinese agricultural CO₂ had an increasing trend, and the spatial spillover effect was prominent from 2001 to 2018. He et al. ^[20] studied the regional spillover effect of agricultural carbon emission from the perspective of technology diffusion, and pointed the "swarm effect" caused direct interaction of CO₂ emission reduction in various provinces, and the spatial diffusion of agricultural technology progress was more likely to occur in areas with similar agricultural economic development. He et al. ^[21] worked at the nonlinear impact of agricultural industrial agglomeration on agricultural CO₂ emission and concluded that it had a double threshold effect. With the formation of large-scale agricultural production, agricultural CO₂ emission show an "inverted U" evolution trend. Meng et al. ^[22] explored the dynamic influence mechanism of agricultural CO₂ emission in Heilongjiang Province and found that the most important influencing factor to the increase of agricultural CO₂ emission was the agricultural economic level, followed by the agricultural production structure. In their work, the main influencing factor of restraining agricultural CO₂ emission was the efficiency of agricultural production, followed by the scale of agricultural workforce.

Exploring the influencing factors of CO₂ emission in various industries has become a hot issue studied by scholars ^[23,24]. In terms of mathematical algorithm, there are mainly three factor decomposition methods: index decomposition analysis (IDA), structural decision analysis (SDA), and logarithmic mean divisia index (LMDI). SDA method is mainly depended on the input-output table generally updated every five years, which limits the application of SDA method ^[25,26,27]. IDA is simpler in comparison and only needs to use department aggregate data, so it has been used by Zha et al. ^[28] and Balezentis ^[29]. LMDI decomposition method belongs to a branch of IDA method, which achieves full decomposition and no residual. It is a common way applied to study the internal causes of change and action intensity. Alajmi ^[30] revealed the action mechanism of CO₂ emission from nine industries in Saudi Arabia from 1990 to 2016 through LMDI method. Their analysis shows that energy effect is the key role for increasing greenhouse gas emission, and the impact of activities and population also contributes to greenhouse gas emission. Moreover, the impact of activities, energy and population in the power sector is greater than that in the transport sector. Isik et al. ^[31] evaluate and reveal the generating mechanism of CO₂ emission from the transportation sector based on LMDI in Turkey from 2000 to 2017. It was found that the research subject's CO₂ emission was firstly driven by economic growth, following by the effect of population and emission intensity. Fatima et al. ^[32] discussed the main drivers of Chinese industrial CO₂ emission from 1991 to 2016. Their study showed that the first two emission sources were income and labor effects. In addition, energy structure effect had its unique contribution for CO₂ emission increasing. Energy intensity and carbon emission effect were the first two key factors for CO₂ emission reduction. Quan et al. ^[33] selected IPCC method and LMDI model to calculate and decomposed the carbon footprint to logistics industry in the period of 2000 to 2016. Their results showed that economic growth had the greatest driving effect in carbon emission, following by the positive impact of population size and energy structure. The main inhibitory factor was proved to be energy intensity. For the agricultural sector, Liu et al. ^[34] adopted LMDI model to analyze the generating mechanism of agricultural CO₂ emission in Shandong Province. Their results proved that production efficiency and industrial structure to agricultural sector, regional industrial structure and labor force played much in CO₂ emission reduction, while the development level of regional economy and the urbanization rate were the main factors for the increase of agricultural CO₂ emission. Xiong et al. ^[35] studied the influencing factors of agricultural CO₂ emission in different stages in Xinjiang and found that economy was the key point to the increase of agricultural CO₂ emission, while the efficiency factor played the opposite role. Hu et al. ^[36] discovered the characteristics and influencing mechanism about agricultural CO₂ emission in China and concluded that for the whole country, production efficiency, agricultural industrial structure, and rural population were helpful to the reduction, and industrial structure, regional economic ability, and urbanization played the opposite roles.

How to explore excellent mathematical models for solving scientific problems ^[37,38] and discover the decoupling between CO₂ emission and the economic development for various industries are issues worthy of attentions. The OECD ^[39] firstly put forward the concept of "decoupling index", and Vehmas et al. ^[40] proposed six possible theoretical frameworks for delinking and link analysis. At present, Tapio decoupling is a common method used by the researchers to find out the influence mechanism between economy and CO₂ emission. Engo ^[41] applied the Tapio decoupling method to study the decoupling relationship between the CO₂ emission and development of Cameroon's transport sector from 1990 to 2016. Their work showed that there were weak decoupling (WD), strong decoupling (SD), weak negative decoupling (WND), and strong negative decoupling (SND) during the study period. Raza et al. ^[42] also applied Tapio decoupling model to discover the decoupling state of CO₂ emission for the transport sector in Pakistan from 1984 to 2018. It was found that CO₂ emission showed expansive coupling (EC) decoupling during the period 1984–2018. The WD only occurred in the sub-periods of 1999–2003, 2004–2008, and 2009–2013. Wang et al. ^[43] also studied this topic of Chinese iron and steel industry based on Tapio decoupling model. They discovered that the carbon emission was weakly decoupled from economic growth. For the agricultural sector, Han et al. ^[44] studied the interaction between agricultural carbon emission and the development of economy in 30 provinces of China from 1997 to 2015, and found that on a regional scale, the decoupling degree between CO₂ emission and economic development was stronger for the central provinces than that of western provinces, and 24 of all provinces showed the WD state. Liu et al. ^[45] analyzed the decomposition and decoupling of regional agricultural CO₂ emission and declared the decoupling relationship between CO₂ emission and economic development presented a gradual improvement process of first EC, then WD, and finally SD. The decoupling relationship in north, northeast, east and central China was better than that in northwest, southwest and south China, and the decoupling state of a single province fluctuated greatly. Hossain et al. ^[46] explored the decoupling mechanism between agricultural development and agricultural CO₂ emission in Bangladesh and found that it showed WD in most of the time from 1990 to 2017.

In the above literature reviews, some studies have discussed the agricultural CO₂ emission characteristics. In particular, for the researches focusing on CO₂ decomposition, the LMDI method obtains relatively satisfied results. While studying the decoupling mechanism between economic development and CO₂ emission, the Tapio model obtains the temporal and spatial changes of the decoupling relationship.

However, the studies reviewed above ignore the driving factors and decoupling mechanism between the overall agricultural CO₂ emission and economic development. In addition, they rarely take Chinese agriculture sector as a research sample to specifically discuss the issue about driving factor and decoupling trend. This study aims to fill these gaps. The theoretical and practical contributions are summarized as follows: The IPCC is adopted to measure the agricultural CO₂ emission and LMDI model is utilized to identify the impact factors for agricultural CO₂ emission, and then Tapio model is developed to reflect the dependence of economic development on CO₂ emission of Chinese agriculture. The research cycle of this paper is from 1996 to 2020. The whole period is divided into five sub-periods: 1996–2000, 2001–2005, 2006–2010, 2011–2015, and 2016–2020. In addition, this study also provides policymakers with suggestions on carbon emission reduction.

Figure 1 draws the structure of this paper, and the rest is organized as follows. The methodology and data details are given in Section 2, respectively. Section 3 describes the results and discussion, and Section 4 presents the conclusions and policy recommendations.

Figure 1. The logical organization of this paper.

DownLoad: Full-Size Img PowerPoint

2. Methodology and data

2.1. Calculation of CO₂ emission

Considering the essential characteristics of agricultural production and the input of various factors, scholars have developed a more mature calculation algorithm about total CO₂ emission from agricultural production. Researchers' decisions on the selection of carbon sources for agricultural CO₂ emission measurement vary greatly ^[47]. Research on the sources of agricultural CO₂ emission focuses on actual production, taking into account the emission directly or indirectly generated from the agricultural production, including chemical fertilizer, agricultural film, agricultural diesel and other carbon sources. In addition, this study draws lessons from the previous research results ^[48,49] and measures the CO₂ emission of agricultural production from six major carbon sources, such as pesticide, chemical fertilizer, agricultural film, diesel, agricultural irrigation, and agricultural tillage.

$C = \sum {C}_{i} = \sum {T}_{i}\times {\lambda }_{i}$

(1)

In Eq (1), C indicates the total agricultural CO₂ emission, C_i represents the CO₂ emission of class ith source, T_i describes the number of class ith source, λ_i is the emission coefficient of ith CO₂ source. The relevant emission coefficient of each CO₂ source is listed in Table 1.

Table 1. Agricultural CO₂ emission sources with their emission coefficients.

CO₂ emission source	Emission coefficient
Chemical fertilizer	0.8956 Kg/Kg
Pesticides	4.9341 Kg/Kg
Rural film	5.1800 Kg/Kg
Agricultural diesel	0.5927 Kg/Kg
Agricultural sowing area	3.1260 Kg/hm²
Agricultural irrigation area	25.0000 Kg/hm²

| Show Table

DownLoad: CSV

2.2. LMDI decomposition model

This study adopts LMDI to construct the decomposition model of agricultural CO₂ emission in China. As the predecessor of LMDI model, the fundamental principle of index decomposition analysis (IDA) identity is to express a specific research object as the product form of relevant influencing factors (including agricultural production efficiency, production structure, economic development, and workforce in this paper) through the deformation of a series of mathematical identity formulas, and calculates the incremental balance of each influencing factor according to different weights. Considering this most essential mathematical characteristic, many researchers select IDA model to the study of energy consumption for analyzing the different effects of various influencing factors on the research object. The IDA identity is calculated by

${\boldsymbol{V}}^{\boldsymbol{t}} = {\sum }_{\boldsymbol{i}}{\boldsymbol{V}}_{\boldsymbol{i}}^{\boldsymbol{t}} = {\sum }_{\boldsymbol{i}}{\boldsymbol{X}}_{1\boldsymbol{i}}^{\boldsymbol{t}}{\boldsymbol{X}}_{2\boldsymbol{i}}^{\boldsymbol{t}}\cdots {\boldsymbol{X}}_{\boldsymbol{n}\boldsymbol{i}}^{\boldsymbol{t}}$

(2)

where ${\boldsymbol{V}}^{\boldsymbol{t}}$ is an energy-related aggregate or CO₂ emission in period t (0 ≤ t ≤ aT), the subscript i is the distinguishing sign of influencing factor. General assumption of that there are n factors contributing to the changes in ${\boldsymbol{V}}_{\boldsymbol{i}}^{\boldsymbol{t}}$ and each is associated with a quantifiable variable, so it exists n quantifiable variables ${\boldsymbol{X}}_{1\boldsymbol{i}}^{\boldsymbol{t}}$ , ${\boldsymbol{X}}_{2\boldsymbol{i}}^{\boldsymbol{t}}$ , $\cdots$ , ${\boldsymbol{X}}_{\boldsymbol{n}\boldsymbol{i}}^{\boldsymbol{t}}$ . Take the logarithm of ${\boldsymbol{V}}^{\boldsymbol{t}}$ and determine the derivative of period t to get the deformation of Eq (2):

$\frac{d{ln}{V}^{t}}{dt} = {\sum }_{i}{\lambda }_{i}\left(\frac{d{ln}{X}_{1i}^{t}}{dt}+\frac{d{ln}{X}_{2i}^{t}}{dt}+\cdots +\frac{d{ln}{X}_{ni}^{t}}{dt}\right)$

(3)

Take definite integrals on both sides of the Eq (3):

${ln}\frac{{V}^{t}}{{V}^{0}} = {\int }_{0}^{t}\sum \limits_{i}{\lambda }_{i}\left(\frac{d{ln}{X}_{1i}^{t}}{dt}+\frac{d{ln}{X}_{2i}^{t}}{dt}+\cdots +\frac{d{ln}{X}_{ni}^{t}}{dt}\right)dt$

$\text{ = }{\sum }_{i}{\int }_{0}^{t}\left({\lambda }_{i}\frac{d{ln}{X}_{1i}^{t}}{dt}+{\lambda }_{i}\frac{d{ln}{X}_{2i}^{t}}{dt}+\cdots +{\lambda }_{i}\frac{d{ln}{X}_{ni}^{t}}{dt}\right)dt$

(4)

In the above equation, the exact value cannot be calculated for the definite integral in brackets, and the capacity of each influencing factor is usually estimated by giving different weight functions. On this basis, Ang ^[50] proposed the LMDI, whose decomposition weight takes the average logarithmic for the two endpoints in the period.

$\left\{\begin{array}{c}L(x, y) = (y-x)/{ln}(y/x)~~if~~x\ne y\\ L(x, x) = x~~if~~x = y\end{array}\right.\left(x > 0, y > 0\right)$

(5)

Then the weight function is expressed in the following form.

$\stackrel{—}{{w}_{i}} = {\lambda }_{i} = \frac{({V}_{t}^{t}-{V}_{t}^{0}){ln}({V}_{t}^{t}/{V}_{t}^{0})}{({V}^{t}-{V}^{0}){ln}({V}^{t}/{V}^{0})}$

(6)

Taking the weight function into Eq (4).

$\frac{{V}^{t}}{{V}^{0}} = {exp}\left(\sum\limits_{i}\stackrel{—}{{w}_{i}}{ln}\frac{{X}_{1i}^{t}}{{X}_{1i}^{0}}\right){exp}\left(\sum\limits_{i}\stackrel{—}{{w}_{i}}{ln}\frac{{X}_{2i}^{t}}{{X}_{2i}^{0}}\right)\cdots {exp}\left(\sum\limits_{i}\stackrel{—}{{w}_{i}}{ln}\frac{{X}_{ni}^{t}}{{X}_{ni}^{0}}\right)\\ \text{ = }{exp}\left(\sum\limits_{i}\stackrel{—}{{w}_{i}}{ln}\frac{{X}_{1i}^{t}{X}_{2i}^{t}\cdots {X}_{ni}^{t}}{{X}_{1i}^{0}{X}_{2i}^{t}\cdots {X}_{ni}^{t}}\right)\\ \text{ = }{exp}\left(\sum\limits_{i}\frac{({V}_{t}^{t}-{V}_{t}^{0}){ln}({V}_{t}^{t}/{V}_{t}^{0})}{({V}^{t}-{V}^{0}){ln}({V}^{t}/{V}^{0})}{ln}\frac{{V}^{t}}{{V}^{0}}\right)$

$\text{ = }exp\left(\frac{{ln}({V}^{t}/{V}^{0})}{{V}^{t}-{V}^{0}}{\sum }_{i}({V}_{t}^{t}-{V}_{t}^{0})\right) = \frac{{V}^{t}}{{V}^{0}}$

(7)

From the derivation process, it is concluded that the LMDI decomposition method based on logarithmic mean is a complete decomposition method and will not produce residuals. The form of logarithm makes multiplication and addition in the formula can be converted to each other. The LMDI decomposition formula based on addition is:

$\Delta V = {V}^{{'}}-{V}^{0} = \Delta {V}_{X1}+\Delta {V}_{X2}+\cdots +\Delta {V}_{XK}+\Delta {V}_{rsd}$

(8)

where

$\Delta {V}_{XK} = \frac{({V}^{t}-{V}^{0})}{{ln}({V}^{t}/{V}^{0})}ln{D}_{XK}$

(9)

and $\Delta {V}_{rsd}$ is zero, that is zero residual in the decomposition method of LMDI.

Combining the Kaya identity ^[50] of carbon emission with the above LMDI model, Kaya equation is introduced into this study. Through deformation, the annual CO₂ emission C of agriculture is expressed as:

$\mathrm{C} = \frac{\mathrm{C}}{\mathrm{A}}\times \frac{\mathrm{A}}{\mathrm{G}}\times \frac{\mathrm{G}}{\mathrm{P}}\times \mathrm{P} = \mathrm{C}\mathrm{A}\times \mathrm{A}\mathrm{G}\times \mathrm{I}\mathrm{A}\times \mathrm{I}\mathrm{S}$

(10)

In Eq (10), C, A, G, and P represent agricultural CO₂ emission, total agricultural output value, agricultural output, forestry, animal husbandry and fishery, and the number of agricultural workforces, respectively. CA, AG, IA, and IS represent production efficiency, production structure, economic development, and workforce of the agricultural sector, respectively.

Using LMDI, the change of CO₂ emission is decomposed into above four influencing factors CA, AG, IA, and IS. Use the following Eq (11) to measure the change of CO₂ emission in period t relative to the base period zero:

${\Delta C}_{tot}^{t} = {C}^{t}-{C}^{0} = {\Delta C}_{CA}^{t}+{\Delta C}_{AG}^{t}+{\Delta C}_{IA}^{t}+{\Delta C}_{IS}^{t}$

(11)

where ${\Delta C}_{CA}^{t}$ represents the efficiency effect of agricultural production, ${\Delta C}_{AG}^{t}$ represents the effect of agricultural production structure, ${\Delta C}_{IA}^{t}$ represents the economic development effect, ${\Delta C}_{IS}^{t}$ represents the scale effect of labor force. Each one is calculated by Eqs (12)–(15) respectively.

${\Delta C}_{CA}^{t} = L\left({C}^{0}, {C}^{t}\right)\cdot ln\left({CA}^{t}/{CA}^{0}\right)$

(12)

${\Delta C}_{AG}^{t} = L\left({C}^{0}, {C}^{t}\right)\cdot ln\left({AG}^{t}/{AG}^{0}\right)$

(13)

${\Delta C}_{IA}^{t} = L\left({C}^{0}, {C}^{t}\right)\cdot ln\left({IA}^{t}/{IA}^{0}\right)$

(14)

${\Delta C}_{IS}^{t} = L\left({C}^{0}, {C}^{t}\right)\cdot ln\left({IS}^{t}/{IS}^{0}\right)$

(15)

where $L\left({C}^{0}, {C}^{t}\right)$ represents the logarithmic average weight defined as:

$L\left({C}^{0}, {C}^{t}\right) = \left({C}^{t}-{C}^{0}\right)/ln\left({C}^{t}/{C}^{0}\right)$

(16)

The contribution rate or effect value of each factor is:

$\Delta {D}_{CA} = \frac{\Delta {C}_{CA}}{\Delta C}\times 100\%$

(17)

$\Delta {D}_{AG} = \frac{\Delta {C}_{AG}}{\Delta C}\times 100\%$

(18)

$\Delta {D}_{IA} = \frac{\Delta {C}_{IA}}{\Delta C}\times 100\%$

(19)

$\Delta {D}_{IS} = \frac{\Delta {C}_{IS}}{\Delta C}\times 100\%$

(20)

2.3. Decoupling model

For evaluating the decoupling relationship between agricultural CO₂ emission and economic development, this paper adopts Tapio model:

$\beta = \frac{\%\Delta C}{\%\Delta A} = \frac{\left({C}^{t}-{C}^{0}\right)/{C}^{0}}{\left({A}^{t}-{A}^{0}\right)/{A}^{0}}$

(21)

where β represents the decoupling index of Tapio; % $\Delta$ A indicates the change rate of agricultural production value, and % $\Delta$ C is the change rate of CO₂ emission; C^t and C⁰ represent agricultural CO₂ emission for the current periods t and zero, respectively; A^t and A⁰ indicate the agricultural production value for the current periods t and zero, respectively.

Tapio decoupling index defined in Eq (21) is a ratio, which shows the percentage change for agricultural CO₂ emission if the value of agricultural production jumps by one percentage point. In order to explore a clearer decoupling trend for agricultural CO₂ emission and economic development, we divide the decoupling trend into eight states, as shown in Table 2. It should be noted that: The eight states are determined by the logical relationship " & " between five variables (%ΔC, %ΔA, β, A, and C), rather than by a single variable, such as β.

Table 2. Measures of decoupling index.

Decoupling state	%ΔC	%ΔA	β	Interpretation
SD	< 0	> 0	Β < 0	A $\uparrow$	C $\downarrow$
WD	> 0	> 0	0 < β < 0.8	A $\uparrow$	C $\nearrow$
RD	< 0	< 0	β > 1.2	A $\searrow$	C $\downarrow$
END	> 0	> 0	β > 1.2	A $\nearrow$	C $\uparrow$
SND	> 0	< 0	β < 0	A $\downarrow$	C $\uparrow$
WND	< 0	< 0	0 < β < 0.8	A $\downarrow$	C $\searrow$
EC	> 0	> 0	0.8 < β < 1.2	A $\uparrow$	C $\uparrow$
RC	< 0	< 0	0.8 < β < 1.2	A $\downarrow$	C $\downarrow$
Notes: $\uparrow$ , $\downarrow$ , $\nearrow$ and $\searrow$ represent rapid growth, rapid decrease, slow growth and slow decrease, respectively.

| Show Table

DownLoad: CSV

If the agricultural economy grows significantly and the CO₂ emission decrease rapidly, we define it as "strong decoupling" (SD) (when β < 0), which is the goal of achieving low-carbon agriculture. On the contrary, "strong negative decoupling" (SND) indicates the significantly decline of agricultural economy with the rapid emission of agricultural CO₂ (when β < 0), which is the last thing we want to see in the process of achieving agricultural economic development. The remaining six states are between SD and SND. If the agricultural economy develops significantly with a dilatory emission of agricultural CO₂ (when 0 < β < 0.8), it is defined as "weak decoupling" (WD). In case the agricultural economy slows down and CO₂ emission drops sharply (when β > 1.2), it is defined as "recurrent decoupling" (RD). "Expansive negative decoupling" (END) indicates the slow development of agricultural economy with the rapid emission of agricultural CO₂ (when β > 1.2). "Weak negative decoupling" (WND) indicates that the rapid retrogression of agricultural economy with the slow decline in agricultural CO₂ emission (when 0 < β < 0.8). "Expansive coupling" (EC) means that both rapidly increase (when 0.8 < β < 1.2), while "recurrent coupling" (RC) shows that both decrease rapidly (when 0.8 < β < 1.2).

2.4. Data sources

In order to ensure the reliability of data, this study applies time series statistics, with a span of 1996–2020. The data of six major carbon sources of agricultural production are selected, among which the data of pesticide, chemical fertilizer, agricultural film, sowing area, diesel, GDP and irrigation area, agriculture, animal husbandry, fishery and forestry GDP, and total rural population are mainly from China Statistical and Rural Statistical Yearbook. For GDP of agriculture, agriculture, animal husbandry, fishery and forestry GDP, 1996 is the starting year, and comparable data are obtained.

3. Results and discussion

Our empirical design is divided into three parts: Measurement of CO₂ emission for agricultural sector, influencing factor decomposition, and analysis of decoupling trend. They are strictly corresponding to the three parts of the methodology section, and also in line with the design idea of empirical analysis, that is, the order of measurement, decomposition and decoupling. Each part of this section is closely related and has a connecting design rationality. In the first empirical analysis, we refer to literatures ^[48,49] and select six carbon sources for the measurement of agricultural CO₂ emission. For the empirical analysis about decomposition, we comprehensively discuss the experimental results from two aspects of different five-year plans and different factors. In the subsection 3.3 of decoupling analysis, we present the discussion and analysis for the decoupling states of different five-year plans.

3.1. CO₂ emission measurement for Chinese agricultural sector

The curve related to the CO₂ emission in Figure 2 shows the phased trend. In addition, the contribution of each carbon source is quite different. It can be seen from Figure 2 that during the period from 1996 to 2020, the overall CO₂ emissions from Chinese agricultural sector raised by 47.84%, from 53.49 million tons to 79.08 million tons. The CO₂ emission curve of the agricultural sector is divided into two stages: a stable growth stage (1996–2015) and a rapid decline stage (2016–2020). During the first stage (1996–2015), the agricultural CO₂ emission raised by 70.89%, and the total emissions increased from 53.49 million tons to 91.41 million tons. During the second stage (2016–2020), the reduction rate of agricultural CO₂ emission was -14.32%. For the perspective of carbon source structure, from 1996 to 2020, the CO₂ emission from chemical fertilizer were the largest, accounting for 60.27% as a whole, followed by rural diesel, film and pesticide, accounting for 14.07, 13.02 and 9.95%, respectively, while the CO₂ emission from irrigation were small, accounting for 2.00%, and the CO₂ emission from sowing were very small, only 0.68%.

Figure 2. Agricultural CO₂ emission and the contributions of different carbon sources.

DownLoad: Full-Size Img PowerPoint

3.2. Analysis of influencing factors

Table 3, Figures 3 and 4 present the statistical results which show that: From the perspective of contribution and contribution rate, agricultural workforce and agricultural production efficiency play two major roles in curbing CO₂ emission, while agricultural economic development is the most key factor to promote the growth of CO₂ emission. According to the five-year plans, the study estimates the impact of each factor during the period 1996–2020. Table 3 and Figure 3 present the results of factor decomposition from different aspects, where ∆C_CA and ∆C_AG represent the efficiency effect and the structural effect about agricultural production, respectively. ∆C_IA represents the economic development effect, ∆C_IS represents the scale effect of labor force, ∆C represents is the sum of the above four factors. From 1996 to 2020, 28.6469 million tons is the change of total CO₂ emissions from the agricultural sector, of which the contributions from productive efficiency, productive structure, economic development, and labor are -20.4657 million tons, 5.5485 million tons, 76.8315 million tons and -33.2674 million tons, respectively. As shown in Figure 4, the respective contribution rates are -71.44, 19.37, 268.20 and -116.13%, respectively.

Table 3. Decomposition distribution for agricultural CO₂ emission (10, 000 t).

	∆C_CA	∆C_AG	∆C_IA	∆C_IS	∆C
1996–2000	1386.60	128.18	-166.50	-339.16	1009.12
2001–2005	200.62	-393.81	1892.28	-536.31	1162.79
2006–2010	-2090.88	688.55	3377.20	-817.03	1157.84
2011–2015	-676.29	130.51	2244.24	-930.20	768.25
2016–2020	-2180.60	170.63	2212.19	-1435.53	-1233.31
1996–2020	-2046.57	554.85	7683.15	-3326.74	2864.69

| Show Table

DownLoad: CSV

Figure 3. The change of effect value for the four factors.

DownLoad: Full-Size Img PowerPoint

Figure 4. Contributions of all factors to the changes of CO2 emission.

DownLoad: Full-Size Img PowerPoint

From the perspective of different five-year plans, the factor of agricultural economy plays a weak negative role to the CO₂ emission during the period of ninth five-year plan. As shown in Figures 3 and 4, during 1996–2000, the contribution rate is -16.50%, reducing agricultural CO₂ emission by 1.665 million tons. This is mainly because, as shown in Figure 5, during this period, agricultural GDP shows a tendency of increasing first and decreasing later, from 2043.3 billion in 1996 to 1856.4 billion in 2000.

Figure 5. Changes in agricultural GDP and agricultural workforce.

DownLoad: Full-Size Img PowerPoint

Figures 3–5 show that, after the tenth five-year plan, the contribution of economic development for the agricultural CO₂ emission starts getting stronger. From 2001 to 2005, the CO₂ emission increases by 18.92 million tons to 2006–2010, the CO₂ emission increases by 33.77 million tons, and the contribution rate reaches 291.68% from 162.74%. This is mainly because despite the short-term decline in agricultural CO₂ emission in 2009 affected by the 2008 financial crisis, the agricultural GDP still reaches 3555.4 billion in 2010, and the growth rate reaches 38.97% in the period of eleventh five-year plan, while the agricultural workforce decreases rapidly, from 745 million to 656 million. During 2011–2015 and 2016–2020, economic growth slows down, while the agricultural workforce continues to decline, and the increase of CO₂ emission caused by economic development is still about 22 million tons.

Moreover, Figure 5 concludes that: Although the agricultural workforce has decreased, their quality and agricultural technical skills have been continuously improved. Energy saving measure, emission reduction technology and strong risk resistance have promoted the growth of regional low-carbon agriculture.

Agricultural workforce plays the second key role for curbing CO₂ emission. The factor of agricultural workforce had been inhibiting the rise of CO₂ emission during the ninth to the thirteenth five-year plan. The CO₂ emission was declined by 3.3916 million tons in the period of the ninth five-year plan, while the it was reduced by 14.3553 million tons from 2016 to 2020. This phenomenon shows that the force is gradually strengthened. As shown in Figure 5, in the past 25 years, the population of agricultural workforces had changed from 850.85 million to 509.79 million, and the proportion of reduction is about 40.08%. The main reason for this phenomenon is that under the background of industrialization and urbanization, a large number of farmers have transferred from agricultural industry to non-agricultural industry, and agricultural production has gradually concentrated in the hands of large grain growers. With the implementation of policies such as the land transfer system, the scale of agricultural production has expanded year by year.

Figure 6 concluded that the efficiency of agricultural production promotes firstly and inhibits later agricultural CO₂ emission. The change time of inhibition occurs at the period of eleventh five-year plan, that is, before 2006–2010, the factor of agricultural production efficiency promotes the increase of CO₂ emission, while during the eleventh to the thirteenth ones, it inhibits CO₂ emission. This is mainly because at the period of ninth five-year plan, the government did not make much efforts to control the agricultural environment, the agricultural economic development could not keep up with the rising rate of agricultural CO₂ emission, and the efficiency of agricultural production raised from 4260 tons/one hundred million to 5450 tons/one hundred million. In the next period, the efficiency of agricultural production reached the peak of 5950 tons/one hundred million in 2003 and decreased to 5620 tons/one hundred million in 2005, which explained why the factor of agricultural production efficiency presented a small contribution in promoting CO₂ emission during the period of tenth five-year plan. From 2006 to 2020, the efficiency of agricultural production did not decline much, resulting in little effect on CO₂ emission reduction. The rapid rising of agricultural production efficiency during the periods of eleventh and thirteenth five-year plan strengthened the reduction effect of CO₂ emission year by year, which also showed that the environmental policy increased agricultural productivity and reduced the environmental pressure on China's agricultural sector.

Figure 6. Changes in agricultural production efficiency and agricultural production structure.

DownLoad: Full-Size Img PowerPoint

Except the period from 2001 to 2005, the structural factor of agricultural production was the secondary role to the growth of agricultural CO₂ emission. The impact from the structure of agricultural production for CO₂ emission is classified into two stages. From 1996 to 2005, The first one showed that it promoted first and then inhibited. The second stage, 2006–2020, showed a promoting effect, but the effect was gradually reduced. In the first stage, although the structural factor of agricultural production promoted the CO₂ emission, their role was small, and the contribution rate was only 12.7%. For the period of tenth five-year plan shown in Figure 6, the proportion of agricultural output value decreased from 0.597 to 0.563, indicating that the structural factor of agricultural production has achieved 3.9381 million tons of CO₂ emission reduction. In the second stage, the proportion of agricultural output value showed a fluctuating upward trend. The structural factors of agricultural production increased CO₂ emission by 6.8855 million tons during this period, which was the second-largest role in promoting the rising of CO₂ emission. Later, in the twelfth and the thirteenth five-year plans, although it still promoted the emission of CO₂, the contribution rate was only 16.99 and 13.83%.

3.3. Analysis on decoupling

The results in Table 4 conclude that the development of agricultural economy and the CO₂ emission had four states of decoupling for five five-year plan periods (SND, END, WD, WD, SD and WD) from 1996 to 2020. As shown in Table 4, during 1996–2020, the decoupling degree of β was 0.489, showing a WD state. This result shows that during the past five periods of five-year plan, with the rapid economic development, the agricultural CO₂ emission is increasing slowly. This can be explained that during this period, the advanced agricultural high-tech, the implementation of circular agriculture, and the optimization of industrial structure have kept the output value of the agricultural sector growing steadily, while the pressure on the ecological environment caused by agricultural CO₂ emission has increased. In order to conform to the development concept of "resource-saving and environment-friendly society", governments at all levels have implemented some successful environmental protection policies. The concept of energy conservation and emission reduction has taken root in the hearts of the people, and the development of low-carbon agriculture has become the consensus of the whole society. The rising rate of CO₂ emission from agricultural production begins to be lower than that of total CO₂ emissions from agricultural production factors. The WD state also occurs during the eleventh five-year plan, as well as its next one.

Table 4. Decoupling distribution of Chinese agricultural CO₂ emission.

Period	%ΔC	%ΔA	β	Decoupling state
1996–2000	0.200	-0.066	-3.034	SND
2001–2005	0.192	0.157	1.227	END
2006–2010	0.160	0.518	0.310	WD
2011–2015	0.092	0.179	0.511	WD
2016–2020	-0.135	0.118	-1.146	SD
1996–2020	0.568	1.162	0.489	WD

| Show Table

DownLoad: CSV

For the period of 1996–2000, CO₂ emission from the agricultural sector showed an SND state with a decoupling degree of -3.034. This indicated that during this period, while agricultural economic development declined rapidly, CO₂ emission increased rapidly, and the decoupling state for the economic development was the worst. Investigating its reason, first, during this period, the utilization of agricultural chemical fertilizer continued to increase, and the CO₂ emission generated by the increase of chemical fertilizer increased the overall agricultural CO₂ emission by 49.05%. Second, in the late 1990s, phased changes took place in agriculture and rural economy. Floods and droughts gradually increased, the per capita cultivated land area decreased, the constraints on agricultural economic development increased, and the actual economic output value decreased slightly.

The period of tenth five-year plan presented the END state with a decoupling degree of 1.227, which showed that CO₂ emission continued to increase rapidly, and the agricultural economic development had a slow growth rate. Compared with the ninth five-year plan period, the government has made forward-looking adjustments to China's economic structure in this period, and the basic position of the agricultural economy has been valued and strengthened. At the same time, although greater efforts have been made to develop agricultural high-tech and industrialization, the cycle is long, resulting in that the CO₂ emission have not been alleviated.

The agricultural CO₂ emission presented a SD state during 2016–2020. While economic growth, CO₂ emission decreased. This is mainly because, first, in order to achieve remarkable results in agricultural scientific and technological innovation, the agricultural industry with biotechnology and information technology as the core has developed rapidly. Second, Chinese government has vigorously promoted the reduction on the consumption of chemical fertilizer, as well as the reduction and harm control of pesticides. Just in the period of thirteenth five-year plan, the consumption of chemical fertilizer decreased by 12.82%, due to the zero-growth action plan for fertilizer consumption. Under the guidance of this policy, China has made great efforts to develop green agriculture and promoted clean technology, achieves negative growth of pesticide consumption, and reduces the use of chemical fertilizer.

4. Conclusions and policy implication

4.1. Conclusions

Some works ^[51,52,53] point out that the environmental responsibility issues arising from China's sustainable development need urgent attention, and one main source of CO₂ emission is from the agricultural production in China. In order to realize the low-carbon agriculture, our work aims to adopt the LMDI method for studying the driving factors and the Tapio model for exploring the decoupling trend of China's agricultural CO₂ emission from 1996–2020, which could help government find accurate CO₂ emission reduction policies. The following conclusions are presented by analyzing and discussing the experimental results.

1) The trend of agricultural CO₂ emission is divided into two stages. From 1996 to 2015, the agricultural CO₂ emission increased rapidly, of which the emission from chemical fertilizers accounted for 60.47%. From 2016 to 2020, the use of chemical fertilizer was controlled, and the CO₂ emission was alleviated and gradually decreased.

2) The agricultural workforce and production structure are the two main factors to inhibit CO₂ emission, while the agricultural economic development is a key role to promote the growth of CO₂ emission, as well as the production efficiency.

3) Between 1996 and 2020, there were four decoupling states. From the ninth to the thirteenth five-year plan, the decoupling state successively appeared SND, END, WD and SD, and the decoupling state gradually improved, which concluded that if the agricultural economy developed rapidly, the agricultural CO₂ emission had been gradually alleviated.

Based on LMDI model, the key factors to the agricultural CO₂ emission are identified. The analysis results of Tapio model not only describe the decoupling degree, but also intuitively show the causes of the decoupling state.

We frankly admit that our research still has the following shortcomings. Firstly, we have taken into account the existing circumstances of Chinese agricultural production and the experience of previous studies when selecting carbon sources, but this may not be a template for other nations or regions. Secondly, China is a large country with rich resources, and agricultural production varies greatly among different regions and provinces, while this paper has not conducted further research about the driving factors and decoupling trend analysis. Our follow-up work will focus on the above two aspects and conduct more in-depth research.

4.2. Policy implication

Till now, the CO₂ emission reduction measures in the Chinese agricultural sector still need the further perfection and improvement, so some suggestions about the policy are put forward.

1) Change the mode of agricultural economic growth and establish a sustainable agricultural ecosystem. This paper reveals that the development of agricultural economy is the main factor leading to the increase emission of agricultural CO₂ in China. According to the requirements of agricultural supply-side structural reform, the policy making departments should pay more attentions to the quality and efficiency of emission reduction policies, build a low-carbon agricultural factor market. Society should abandon the extensive growth of agricultural production, and more financial and material resources should be invested in sustainable and intensive development.

2) Intensify efforts to develop advanced agricultural technology and improve agricultural production efficiency. Recent years have witnessed the efficiency of agricultural production has gradually become the important contribution to the reduction of agricultural CO₂ emission. Therefore, relevant departments need to devote more energy to developing agricultural scientific research, especially vigorously develop precision agriculture and ecological recycling technology, strengthen the construction of agricultural socialized service system, demonstrate and promote new technologies such as balanced fertilization, encourage the scientific utilization of chemical fertilizers and pesticides, and promote the sowing of harmless agricultural products.

3) Continue to pay attention to the education about the professional level of agricultural workforce. Based on the current situation about the gradual reduction of agricultural workforce, improve the regional cultivated land transfer system, pilot the farmland contract system, implement the technical promotion, policy support and guidance of low-carbon agriculture for large agricultural households, and strive for the multiple objectives including the green agriculture, environmental protection, and harmonious development of rural economy while increasing income.

Acknowledgments

All the authors are extremely grateful to the anonymous reviewers for their wonderful comments and insightful suggestions.

Conflict of interest

All authors declare no conflicts of interest in this paper.

References

[1]	F. Hernández-Campos, K. Jeffay, F. D. Smith, Tracking the evolution of web traffic: 1995–2003, in 11th IEEE/ACM International Symposium on Modeling, Analysis and Simulation of Computer Telecommunications Systems, (2003), 16–25. https://doi.org/10.1109/MASCOT.2003.1240638
[2]	H. Schulze, K. Mochalski, Internet study 2008/2009, Ipoque Rep., 37 (2009), 351–362.
[3]	T. Zimmermann, J. Rüth, B. Wolters, O. Hohlfeld, How HTTP/2 pushes the web: An empirical study of HTTP/2 server push, in 2017 IFIP Networking Conference (IFIP Networking) and Workshops, (2017), 1–9. https://doi.org/10.23919/IFIPNetworking.2017.8264830
[4]	O. Hohlfeld, J. Rüth, K. Wolsing, T. Zimmermann, Characterizing a meta-CDN, in International Conference on Passive and Active Network Measurement, (2018), 114–128. https://doi.org/10.1007/978-3-319-76481-8_9
[5]	F. Lichtblau, F. Streibelt, T. Krüger, P. Richter, A. Feldmann, Detection, classification, and analysis of inter-domain traffic with spoofed source IP addresses, in Proceedings of the 2017 Internet Measurement Conference, (2017), 86–99. https://doi.org/10.1145/3131365.3131367
[6]	A. Al-Najjar, S. Teed, J. Indulska, M. Portmann, Flow-based load balancing of web traffic using OpenFlow, in 2017 27th International Telecommunication Networks and Applications Conference (ITNAC), (2017), 1–6. https://doi.org/10.1109/ATNAC.2017.8215411
[7]	Cisco, Cisco Visual Networking Index: Global Mobile Data Traffic Forecast Update, 2016–2021, 2017. Available from: https://www.ramonmillan.com.
[8]	W. Li, A. W. Moore, M. Canini, Classifying HTTP traffic in the new age, ACM SIGCOMM, 8 (2008), 17–22.
[9]	J. Liu, C. Fang, N. Ansari, Request dependency graph: A model for web usage mining in large-scale web of things, IEEE Internet Things J., 3 (2016), 598–608. https://doi.org/10.1109/JIOT.2015.2452964 doi: 10.1109/JIOT.2015.2452964
[10]	L. Vassio, I. Drago, M. Mellia, Detecting user actions from HTTP traces: toward an automatic approach, in 2016 International Wireless Communications and Mobile Computing Conference (IWCMC), (2016), 50–55. https://doi.org/10.1109/IWCMC.2016.7577032
[11]	G. Scavo, Z. B. Houidi, S. Traverso, R. Teixeira, M. Mellia, WeBrowse: mining HTTP logs online for network-based content recommendation, preprint, arXiv: 1602.06678.
[12]	P. Fiadino, A. Bar, P. Casas, HTTPTag: a flexible on-line HTTP classification system for operational 3G networks, in 2013 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS), (2013), 71–72. https://doi.org/10.1109/INFCOMW.2013.6970744
[13]	X. Gui, J. Liu, Q. Lv, C. Dong, Z. Lei, Probabilistic top-k query: model and application on web traffic analysis, China Commun., 13 (2016), 123–137. https://doi.org/10.1109/CC.2016.7513208 doi: 10.1109/CC.2016.7513208
[14]	J. Sun, L. She, H. Chen, W. Zhong, C. Chang, Z. Chen, et al., Automatically identifying apps in mobile traffic, Concurrency Comput. Pract. Exper., 28 (2016), 3927–3941. https://doi.org/10.1002/cpe.3703 doi: 10.1002/cpe.3703
[15]	G. Aceto, D. Ciuonzo, A. Montieri, A. Pescapé, Mobile encrypted traffic classification using deep learning: experimental evaluation, lessons learned, and challenges, IEEE Trans. Network Serv. Manage., 16 (2019), 445–458. https://doi.org/10.1109/TNSM.2019.2899085 doi: 10.1109/TNSM.2019.2899085
[16]	P. Białczak, W. Mazurczyk, Characterizing anomalies in malware-generated HTTP traffic, Secur. Commun. Networks, 2020 (2020). https://doi.org/10.1155/2020/8848863 doi: 10.1155/2020/8848863
[17]	J. Li, H. Zhang, Z. Wei, The weighted word2vec paragraph vectors for anomaly detection over HTTP traffic, IEEE Access, 8 (2020), 141787–141798. https://doi.org/10.1109/ACCESS.2020.3013849 doi: 10.1109/ACCESS.2020.3013849
[18]	G. D'Angelo, F. Palmieri, Network traffic classification using deep convolutional recurrent autoencoder neural networks for spatial–temporal features extraction, J. Network Comput. Appl., 173 (2021), 102890. https://doi.org/10.1016/j.jnca.2020.102890 doi: 10.1016/j.jnca.2020.102890
[19]	S. Dong, Y. Xia, T. Peng, Traffic identification model based on generative adversarial deep convolutional network, Ann. Telecommun., (2021), 1–15. https://doi.org/10.1007/s12243-021-00876-6 doi: 10.1007/s12243-021-00876-6
[20]	T. T. Nguyen, G. Armitage, A survey of techniques for internet traffic classification using machine learning, IEEE Commun. Surv. Tutorials, 10 (2008), 56–76. https://doi.org/10.1109/SURV.2008.080406 doi: 10.1109/SURV.2008.080406
[21]	A. Callado, C. Kamienski, G. Szabó, B. P. Gero, J. Kelner, S. Fernandes, et al., A survey on internet traffic identification, IEEE Commun. Surv. Tutorials, 11 (2009), 37–52. https://doi.org/10.1109/SURV.2009.090304 doi: 10.1109/SURV.2009.090304
[22]	A. Dainotti, A. Pescape, K. C. Claffy, Issues and future directions in traffic classification, IEEE Network, 26 (2012), 35–40. https://doi.org/10.1109/MNET.2012.6135854 doi: 10.1109/MNET.2012.6135854
[23]	M. Finsterbusch, C. Richter, E. Rocha, J. Muller, K. Hanssgen, A survey of payload-based traffic classification approaches, IEEE Commun. Surv. Tutorials, 16 (2013), 1135–1156. https://doi.org/10.1109/SURV.2013.100613.00161 doi: 10.1109/SURV.2013.100613.00161
[24]	D. Naboulsi, M. Fiore, S. Ribot, R. Stanica, Large-scale mobile traffic analysis: a survey, IEEE Commun. Surv. Tutorials, 18 (2015), 124–161. https://doi.org/10.1109/COMST.2015.2491361 doi: 10.1109/COMST.2015.2491361
[25]	P. Velan, M. Cermak, P. Celeda, M. Drasar, A survey of methods for encrypted traffic classification and analysis, Int. J. Network Manage., 25 (2015), 355–374. https://doi.org/10.1002/nem.1901 doi: 10.1002/nem.1901
[26]	D. Acarali, M. Rajarajan, N. Komninos, I. Herwono, Survey of approaches and features for the identification of HTTP-based botnet traffic, J. Network Comput. Appl., 76 (2016), 1–15. https://doi.org/10.1016/j.jnca.2016.10.007 doi: 10.1016/j.jnca.2016.10.007
[27]	W. Pan, G. Cheng, X. Guo, S. Huang, Review and perspective on encrypted traffic identification research, J. Commun., 37 (2016), 154–167. https://doi.org/10.11959/j.issn.1000-436x.2016187 doi: 10.11959/j.issn.1000-436x.2016187
[28]	F. Pacheco, E. Exposito, M. Gineste, C. Baudoin, J. Aguilar, Towards the deployment of machine learning solutions in network traffic classification: A systematic survey, IEEE Commun. Surv. Tutorials, 21 (2018), 1988–2014. https://doi.org/10.1109/COMST.2018.2883147 doi: 10.1109/COMST.2018.2883147
[29]	S. Rezaei, X. Liu, Deep learning for encrypted traffic classification: an overview, IEEE Commun. Mag., 57 (2019), 76–81. https://doi.org/10.1109/MCOM.2019.1800819 doi: 10.1109/MCOM.2019.1800819
[30]	A. D'Alconzo, I. Drago, A. Morichetta, M. Mellia, P. Casas, A survey on big data for network traffic monitoring and analysis, IEEE Trans. Network Serv. Manage., 16 (2019), 800–813. https://doi.org/10.1109/TNSM.2019.2933358 doi: 10.1109/TNSM.2019.2933358
[31]	W. M. Shbair, T. Cholez, J. François, I. Chrisment, A survey of HTTPS traffic and services identification approaches, preprint, arXiv: 2008.08339.
[32]	G. Aceto, D. Ciuonzo, A. Montieri, A. Pescape, Toward effective mobile encrypted traffic classification through deep learning, Neurocomputing, 409 (2020), 306–315. https://doi.org/10.1016/j.neucom.2020.05.036 doi: 10.1016/j.neucom.2020.05.036
[33]	A. Shahraki, M. Abbasi, A. Taherkordi, A. D. Jurcut, . Active learning for network traffic classification: a technical study, preprint, arXiv: 2106.06933.
[34]	S. Dong, R. Li, Traffic identification method based on multiple probabilistic neural network model, Neural Comput. Appl., 31 (2019), 473–487. https://doi.org/10.1007/s00521-017-3081-x doi: 10.1007/s00521-017-3081-x
[35]	H. Tang, Z. Li, Design and implementation of a DPI-Based P2P traffic control system, Inf. Secur. Commun. Privacy, 6 (2007).
[36]	M. Soysal, E. G. Schmidt, Machine learning algorithms for accurate flow-based network traffic classification: evaluation and comparison, Perform. Eval., 67 (2010), 451–467. https://doi.org/10.1016/j.peva.2010.01.001 doi: 10.1016/j.peva.2010.01.001
[37]	S. Dong, Multi class SVM algorithm with active learning for network traffic classification, Expert Syst. Appl., 176 (2021), 114885. https://doi.org/10.1016/j.eswa.2021.114885 doi: 10.1016/j.eswa.2021.114885
[38]	F. Haddadi, A. N. Zincir-Heywood, Benchmarking the effect of flow exporters and protocol filters on botnet traffic classification, IEEE Syst. J., 10 (2016), 1390–1401. https://doi.org/10.1109/JSYST.2014.2364743 doi: 10.1109/JSYST.2014.2364743
[39]	T. Bakhshi, B. Ghita, On internet traffic classification: A two-phased machine learning approach, J. Comput. Networks Commun., 2016 (2016). https://doi.org/10.1155/2016/2048302 doi: 10.1155/2016/2048302
[40]	S. Dong, X. Zhang, D. Zhou, Auto adaptive identification algorithm based on network traffic flow, Int. J. Comput. Commun. Control, 9 (2014), 672–685. http://dx.doi.org/10.1145/1080091.1080119 doi: 10.1145/1080091.1080119
[41]	Y. Dong, J. Zhao, J. Jin, Novel feature selection and classification of Internet video traffic based on a hierarchical scheme, Comput. Networks, 119 (2017), 102–111. https://doi.org/10.1016/j.comnet.2017.03.019 doi: 10.1016/j.comnet.2017.03.019
[42]	S. Dong, W. Liu, D. Zhou, Y. Qi, NSVM: A new SVM algorithm based on traffic flow metric, J. Internet Technol., 16 (2015), 1005–1014.
[43]	R. Dubin, A. Dvir, O. Pele, O. Hadar, I know what you saw last minute—encrypted http adaptive video streaming title classification, IEEE Trans. Inf. Forensics Secur., 12 (2017), 3039–3049. https://doi.org/10.1109/TIFS.2017.2730819 doi: 10.1109/TIFS.2017.2730819
[44]	H. D. Trinh, A. F. Gambin, L. Giupponi, M. Rossi, P. Dini, Mobile traffic classification through physical control channel fingerprinting: a deep learning approach, IEEE Trans. Network Serv. Manage., 2020 (2020). https://doi.org/10.1109/TNSM.2020.3028197 doi: 10.1109/TNSM.2020.3028197
[45]	M. Xie, J. Fu, Y. Wang, G. Peng, Monitoring and blocking methods of HTTP traffic injection in mobile web browser, J. Wuhan Univ., 63 (2017), 385–396.
[46]	G. Rizothanasis, N. Carlsson, A. Mahanti, Identifying user actions from HTTP (S) traffic, in 2016 IEEE 41st Conference on Local Computer Networks (LCN), (2016), 555–558. https://doi.org/10.1109/LCN.2016.91
[47]	J. Manzoor, I. Drago, R. Sadre, How HTTP/2 is changing web traffic and how to detect it, in 2017 Network Traffic Measurement and Analysis Conference (TMA), (2017), 1–9. https://doi.org/10.23919/TMA.2017.8002899
[48]	J. Muehlstein, Y. Zion, M. Bahumi, I. Kirshenboim, R. Dubin, A. Dvir, et al., Analyzing HTTPS encrypted traffic to identify user's operating system, browser and application, in 2017 14th IEEE Annual Consumer Communications & Networking Conference (CCNC), (2017), 1–6. https://doi.org/10.1109/CCNC.2017.8013420
[49]	T. Petsas, A. Papadogiannakis, M. Polychronakis, E. P. Markatos, T. Karagiannis, Measurement, modeling, and analysis of the mobile app ecosystem, ACM Trans. Model. Perform. Eval. Comput. Syst., 2 (2017), 7. https://doi.org/10.1145/2993419 doi: 10.1145/2993419
[50]	M. Rapoport, P. Suter, E. Wittern, O. Lhotak, J. Dolby, Who you gonna call? Analyzing web requests in Android applications, in 2017 IEEE/ACM 14th International Conference on Mining Software Repositories (MSR), (2017), 80–90. https://doi.org/10.1109/MSR.2017.11
[51]	Z. B. Houidi, G. Scavo, S. Ghamri-Doudane, A. Finamore, S. Traverso, M. Mellia, Gold mining in a river of internet content traffic, in International Workshop on Traffic Monitoring and Analysis, Springer, (2014), 91–103. https://doi.org/10.1007/978-3-642-54999-1_8
[52]	UNIBS, 2011. Available from: http://netweb.ing.unibs.it/ ntw/tools/traces/.
[53]	CIC, 2021. Available from: https://www.unb.ca/cic/datasets/.
[54]	UMass, 2021. Available from: http://skuld.cs.umass.edu/traces/network/README-webident2.
[55]	CAIDA, 2021. Available from: https://catalog.caida.org/search?query=types=dataset.
[56]	WIDE, 2021. Available from: http://mawi.wide.ad.jp/mawi/.
[57]	G. Maier, A. Feldmann, V. Paxson, A. Mark, On dominant characteristics of residential broadband internet traffic, in Proceedings of the 9th ACM SIGCOMM Conference on Internet Measurement Conference, (2009), 90–102. https://doi.org/10.1145/1644893.1644904
[58]	L. Bernaille, R. Teixeira, I. Akodkenou, A. Soule, K. Salamatian, Traffic classification on the fly, ACM SIGCOMM Comput. Commun. Rev., 36 (2006), 23–26. https://doi.org/10.1145/1129582.1129589 doi: 10.1145/1129582.1129589
[59]	L. Bernaille, R. Teixeira, K. Salamatian, Early application identification, in Proceedings of the 2006 ACM CoNEXT Conference, (2006), 1–12. https://doi.org/10.1145/1368436.1368445
[60]	L. Bernaille, R. Teixeira, Early recognition of encrypted applications, in International Conference on Passive and Active Network Measurement, (2007), 165–175. https://doi.org/10.1007/978-3-540-71617-4_17
[61]	A. Este, F. Gringoli, L. Salgarelli, On the stability of the information carried by traffic flow features at the packet level, ACM SIGCOMM Comput. Commun. Rev., 39 (2009), 13–18. https://doi.org/10.1145/1568613.1568616 doi: 10.1145/1568613.1568616
[62]	N. Huang, G. Jai, H. Chao, Early identifying application traffic with application characteristics, in 2008 IEEE International Conference on Communications, (2008), 5788–5792. https://doi.org/10.1109/ICC.2008.1083
[63]	N. Huang, G. Jai, H. Chao, Y. Tzang, H. Chang, Application traffic classification at the early stage by characterizing application rounds, Inf. Sci., 232 (2013), 130–142. https://doi.org/10.1016/j.ins.2012.12.039 doi: 10.1016/j.ins.2012.12.039
[64]	T. T. Nguyen, G. Armitage, P. Branch, S. Zander, Timely and continuous machine-learning-based classification for interactive IP traffic, IEEE/ACM Trans. Networking, 20 (2012), 1880–1894. https://doi.org/10.1109/TNET.2012.2187305 doi: 10.1109/TNET.2012.2187305
[65]	G. He, M. Yang, J. Luo, L. Zhang, Online identification of tor anonymous communication traffic, J. Commun., 24 (2013), 540–556.
[66]	L. Chen, J. Gong, Fast application-level traffic classification using NetFlow records, J. Commun., 33 (2012), 145–152. https://doi.org/1000-436X(2012)01-0145-08
[67]	S. Dong, W. Ding, Traffic classification model based on fusion of multiple classifiers with flow preference, J. Commun., 34 (2013), 143–152. https://doi.org/10.3969/j.issn.1000-436x.2013.10.017 doi: 10.3969/j.issn.1000-436x.2013.10.017
[68]	V. Paxson, Empirically derived analytic models of wide-area TCP connections, IEEE/ACM Trans. Networking, 2 (1994), 316–336. https://doi.org/10.1109/90.330413 doi: 10.1109/90.330413
[69]	C. Dewes, A. Wichmann, A. Feldmann, An analysis of Internet chat systems, in Proceedings of the 3rd ACM SIGCOMM Conference on Internet Measurement, (2003), 51–64. https://doi.org/10.1145/948205.948214
[70]	T. Lang, G. Armitage, P. Branch, H. Choo, A synthetic traffic model for half-life, in Aust. Telecommun. Networks Appl. Conference, 2003 (2003), 1–5.
[71]	T. Lang, P. Branch, G. Armitage, A synthetic traffic model for Quake3, in Proceedings of the 2004 ACM SIGCHI International Conference on Advances in Computer Entertainment Technology, (2004), 233–238. https://doi.org/10.1145/1067343.1067373
[72]	S. Cheung, R. Crawford, M. Dilger, J. Frank, J. Hoagland, K. Levitt, et al. The design of GrIDS: A graph-based intrusion detection system, in Technical Report CSE-99-2, UC Davis Computer Science Department, (1999).
[73]	M. Iliofotou, P. Pappu, M. Faloutsos, M. Mitzenmacher, S. Singh, G. Varghese, Network monitoring using traffic dispersion graphs (tdgs), in Proceedings of the 7th ACM SIGCOMM Conference on Internet Measurement, (2007), 315–320. https://doi.org/10.1145/1298306.1298349
[74]	M. Iliofotou, H. Kim, M. Faloutsos, M. Mitzenmacher, P. Pappu, G. Varghese, Graption: a graph-based P2P traffic classification framework for the internet backbone, Comput. Networks, 55 (2011), 1909–1920. https://doi.org/10.1016/j.comnet.2011.01.020 doi: 10.1016/j.comnet.2011.01.020
[75]	Cisco Visual Networking Index: Global Mobile Data Traffic Forecast Update, 2017–2022 White Paper, 2019. Available from: https://branden.biz/wp-content/uploads/2019/05/white-paper-c11-738429.pdf.
[76]	A. Gember, A. Anand, A. Akella, A comparative study of handheld and non-handheld traffic in campus wi-fi networks, in International Conference on Passive and Active Network Measurement, (2011), 173–183. https://doi.org/10.1007/978-3-642-19260-9_18
[77]	J. Liu, T. Li, G. Chen, Y. Hua, Z. Lei, Mining and modelling the dynamic patterns of service providers in cellular data network based on big data analysis, China Commun., 10 (2013), 25–26. https://doi.org/10.1109/CC.2013.6723876 doi: 10.1109/CC.2013.6723876
[78]	S. Dong, D. Zhou, W. Ding, Traffic classification model based on integration of multiple classifiers, J. Comput. Inf. Syst., 8 (2012), 10429–10437.
[79]	X. Gui, J. Liu, C. Li, Q. Lv, Z. Lei, Fine-grained analysis of cellular smartphone usage characteristics based on massive network traffic, J. China Univ. Posts Telecommun., 23 (2016), 70–75. https://doi.org/10.1016/S1005-8885(16)60035-3 doi: 10.1016/S1005-8885(16)60035-3
[80]	Y. Li, J. Yang, N. Ansari, Cellular smartphone traffic and user behavior analysis, in 2014 IEEE International Conference on Communications (ICC), (2014), 1326–1331. https://doi.org/10.1109/ICC.2014.6883505
[81]	M. Z. Shafiq, L. Ji, A. X. Liu, J. Wang, Characterizing and modeling internet traffic dynamics of cellular devices, in Proceedings of the ACM SIGMETRICS Joint International Conference on Measurement and Modeling of Computer Systems, (2011), 305–316. https://doi.org/10.1145/2007116.2007148
[82]	Q. Xu, J. Erman, A. Gerber, Z. Mao, J. Pang, S. Venkataraman, Identifying diverse usage behaviors of smartphone apps, in Proceedings of the 2011 ACM SIGCOMM conference on Internet measurement conference, (2011), 329–344. https://doi.org/10.1145/2068816.2068847
[83]	F. T. Vincent, R. Spolaor, M. Conti, I. Martinovic, Appscanner: automatic fingerprinting of smartphone apps from encrypted network traffic, in 2016 IEEE European Symposium on Security and Privacy (EuroS & P), (2016), 439–454. https://doi.org/10.1109/EuroSP.2016.40
[84]	M. Shen, J. Zhang, L. Zhu, K. Xu, X. Du, Y. Liu, Encrypted traffic classification of decentralized applications on ethereum using feature fusion, in 2019 IEEE/ACM 27th International Symposium on Quality of Service (IWQoS), (2019), 1–10. https://doi.org/10.1145/3326285.3329053
[85]	G. Aceto, D. Ciuonzo, A. Montieri, A. Pescape, MIMETIC: mobile encrypted traffic classification using multimodal deep learning, Comput. Networks, 165 (2019), 106944. https://doi.org/10.1016/j.comnet.2019.106944 doi: 10.1016/j.comnet.2019.106944
[86]	G. Aceto, D. Ciuonzo, A. Montieri, A. Pescape, Multi-classification approaches for classifying mobile app traffic, J. Network Comput. Appl., 103 (2018), 131–145. https://doi.org/10.1016/j.jnca.2017.11.007 doi: 10.1016/j.jnca.2017.11.007
[87]	G. Xie, Q. Li, Y. Jiang, D. Tao, G. Shen, R. Li, et al., SAM: self-attention based deep learning method for online traffic classification, in Proceedings of the Workshop on Network Meets AI & ML, (2020), 14–20. https://doi.org/10.1145/3405671.3405811
[88]	C. Liu, L. He, G. Xiong, Z. Cao, Z. Li, Fs-net: a flow sequence network for encrypted traffic classification, in IEEE INFOCOM 2019-IEEE Conference on Computer Communications, (2019), 1171–1179. https://doi.org/10.1109/INFOCOM.2019.8737507
[89]	M. Lopez-Martin, B. Carro, A. Sanchez-Esguevillas, J. Lloret, Network traffic classifier with convolutional and recurrent neural networks for internet of things, IEEE Access, 5 (2017), 18042–18050. https://doi.org/10.1109/ACCESS.2017.2747560 doi: 10.1109/ACCESS.2017.2747560
[90]	T. Shapira, Y. Shavitt, Flowpic: encrypted internet traffic classification is as easy as image recognition, in IEEE INFOCOM 2019-IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS), (2019), 680–687. https://doi.org/10.1109/INFCOMW.2019.8845315
[91]	F-SECURE, Threat Description Bluetooth-Worm: SymbOS/Cabir, 2021. Available from: https://www.f-secure.com/v-descs/cabir.shtml.
[92]	F-SECURE, Mobile Threat Report Q4 2011, 2021. Available from: https://www.f-secure.com/documents/996508/1030743/mobile_threat_report_q4_2011.pdf.
[93]	F-SECURE, Threat Description Bluetooth-Worm: SymbOS/Cabir, 2021. Available from: https://www.f-secure.com/v-descs/worm_iphoneos_ikee.shtml.
[94]	GDATASECURITYLAB, FakePlayer, 2021. Available from: https://www.gdata.at/securitylabs/mobile/mobile-malware/.
[95]	F-SECURE, Mobile Threat Report 2012–2014, 2021. Available from: https://www.f-secure.com/en/web/labs_global/whitepapers.
[96]	APPTHORITY, Mobile Threat Report, 2021. Available from: http://info.appthority.com/hubfs/website-LEARN-content/Appthority-Mobile-Threat-Report-Q12015.pdf.
[97]	S. C. Peng, A survey on malware containment models in smartphones, Appl. Mech. Mater., 263 (2013), 3005–3011. https://doi.org/10.4028/www.scientific.net/AMM.263-266.3005 doi: 10.4028/www.scientific.net/AMM.263-266.3005
[98]	S. PENG, S. Yu, A. Yang, Smartphone malware and its propagation modeling: a survey, Commun. Surv. Tutorials, 16 (2014), 925–941. https://doi.org/10.1109/SURV.2013.070813.00214 doi: 10.1109/SURV.2013.070813.00214
[99]	G. Suarez-Tangil, J. E. Tapiador, P. Peris-Lopez, A. Ribagorda, Evolution, detection and analysis of malware for smart devices, Commun. Surv. Tutorials, 16 (2014), 961–987. https://doi.org/10.1109/SURV.2013.101613.00077 doi: 10.1109/SURV.2013.101613.00077
[100]	Y. Zhou, X. Jiang, Dissecting android malware: characterization and evolution, in 2012 IEEE Symposium on Security and Privacy, (2012), 95–109. https://doi.org10.1109/SP.2012.16
[101]	Y. Liu, L. Zhang, J. Liang, S. Qu, Z. Ni, Detecting trojan horses based on system behavior using machine learning method, in 2010 International Conference on Machine Learning and Cybernetics, (2010), 855–860. https://doi.org/10.1109/ICMLC.2010.5580591
[102]	V. K. Gudipati, A. Vetwal, V. Kumar, A. Adeniyi, A. Abuzneid, Detection of trojan horses by the analysis of system behavior and data packets, in 2015 Long Island Systems, Applications and Technology, (2015), 1–4. https://doi.org/10.1109/LISAT.2015.7160176
[103]	J. Nazario, T. Holz, As the net churns: fast-flux botnet observations, in 2008 3rd International Conference on Malicious and Unwanted Software (MALWARE), (2008), 24–31. https://doi.org/10.1109/MALWARE.2008.4690854
[104]	W. Yu, X. Wang, P. Calyam, D. Xuan, W. Zhao, Modeling and detection of camouflaging worm, IEEE Trans. Dependable Secure Comput., 8 (2011), 377–390. https://doi.org/10.1109/TDSC.2010.13 doi: 10.1109/TDSC.2010.13
[105]	NSFOCUS Information Technology Co. Ltd., XcodeGhost automatically Checking, 2015. Available from: https://cloud.nsfocus.com/#/krosa/views/initcdr/secalertindex.
[106]	PANGU JAILBREAK, Statistical Report for XcodeGhost Virus, 2015. Available from: http://x.pangu.io/.
[107]	A. Shabtai, U. Kanonov, Y. Elovici, C. Glezer, Y. Weiss, "Andromaly": a behavioral malware detection framework for android devices, J. Intell. Inf. Syst., 38 (2012), 161–190. https://doi.org/10.1007/s10844-010-0148-x doi: 10.1007/s10844-010-0148-x
[108]	Y. Cao, R. Ji, L. Ji, X. Shao, G. Lei, H. Wang, MPTCP-meLearning: a multi-expert learning-based MPTCP extension to enhance multipathing robustness against network attacks, IEICE Trans. Inf. Syst., E104-D (2021). https://doi.org/10.1587/transinf.2021NGP0009 doi: 10.1587/transinf.2021NGP0009
[109]	F. Song, L. Li, I. You, H. Zhang, Enabling heterogeneous deterministic networks with smart collaborative theory, IEEE Network, 35 (2021), 64–71. https://doi.org/10.1109/MNET.011.2000613 doi: 10.1109/MNET.011.2000613
[110]	F. Song, Z. Ai, H. Zhang, I. You, S. Li, Smart collaborative balancing for dependable network components in cyber-physical systems, IEEE Trans. Ind. Inf., 17 (2021), 6916–6924. https://doi.org/10.1109/TII.2020.3029766 doi: 10.1109/TII.2020.3029766
[111]	C. J. Wright, Towards Real Time Characterization of Grain Growth from the Melt, Columbia University, 2020.
[112]	G. Aceto, D. Ciuonzo, A. Montieri, A. Pescape, DISTILLER: encrypted traffic classification via multimodal multitask deep learning, J. Network Comput. Appl., 183 (2021), 102985. https://doi.org/10.1016/j.jnca.2021.102985 doi: 10.1016/j.jnca.2021.102985
[113]	Z. Bu, B. Zhou, P. Cheng, K. Zhang, Z. Ling, Encrypted network traffic classification using deep and parallel network-in-network models, IEEE Access, 8 (2020), 132950–132959. https://doi.org/10.1109/ACCESS.2020.3010637 doi: 10.1109/ACCESS.2020.3010637
[114]	G. Bovenzi, L. Yang, A. Finamore, A first look at class Incremental Learning in Deep Learning Mobile Traffic Classification, preprint, arXiv: 2107.04464.
[115]	F. Song, M. Zhu, Y. Zhou, I. You, H. Zhang, Smart collaborative tracking for ubiquitous power IoT in edge-cloud interplay domain, IEEE Internet Things J., 7 (2020), 6046–6055. https://doi.org/10.1109/JIOT.2019.2958097 doi: 10.1109/JIOT.2019.2958097
[116]	F. Song, Z. Ai, Y. Zhou, I. You, R. Choo, H. Zhang, Smart collaborative automation for receive buffer control in multipath industrial networks, IEEE Trans. Ind. Inf., 16 (2020), 1385–1394. https://doi.org/10.1109/TII.2019.2950109 doi: 10.1109/TII.2019.2950109

This article has been cited by:

1.	Yasir A. Madani, Mohammed Nour A. Rabih, Faez A. Alqarni, Zeeshan Ali, Khaled A. Aldwoah, Manel Hleili, Existence, Uniqueness, and Stability of a Nonlinear Tripled Fractional Order Differential System, 2024, 8, 2504-3110, 416, 10.3390/fractalfract8070416
2.	Jaromír Kukal, Michal Beneš, Simple difference schemes for multidimensional fractional Laplacian and fractional gradient, 2025, 1311-0454, 10.1007/s13540-025-00386-5

Reader Comments

Your name:*

Email:*
© 2022 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

Mathematical Biosciences and Engineering

4.4

Metrics

Article views(4317) PDF downloads(205) Cited by(1)

Preview PDF

Download XML

Export Citation

Article outline

Show full outline

Figures and Tables

Figures(1) / Tables(4)

Mathematical Biosciences and Engineering

A Survey of techniques for fine-grained web traffic identification and classification

Related Papers:

Abstract

1. Introduction

2. Methodology and data

2.1. Calculation of CO₂ emission

2.2. LMDI decomposition model

2.3. Decoupling model

2.4. Data sources

3. Results and discussion

3.1. CO₂ emission measurement for Chinese agricultural sector

3.2. Analysis of influencing factors

3.3. Analysis on decoupling

4. Conclusions and policy implication

4.1. Conclusions

4.2. Policy implication

Acknowledgments

Conflict of interest

References

This article has been cited by:

Reader Comments

通讯作者: 陈斌, bchen63@163.com

Metrics

Figures and Tables

Other Articles By Authors

Catalog

Mathematical Biosciences and Engineering

A Survey of techniques for fine-grained web traffic identification and classification

Related Papers:

Abstract

1. Introduction

2. Methodology and data

2.1. Calculation of CO2 emission

2.2. LMDI decomposition model

2.3. Decoupling model

2.4. Data sources

3. Results and discussion

3.1. CO2 emission measurement for Chinese agricultural sector

3.2. Analysis of influencing factors

3.3. Analysis on decoupling

4. Conclusions and policy implication

4.1. Conclusions

4.2. Policy implication

Acknowledgments

Conflict of interest

References

This article has been cited by:

Reader Comments

通讯作者: 陈斌, bchen63@163.com

Metrics

Figures and Tables

Other Articles By Authors

Related pages

Tools

Export File

Citation

Format

Content

Catalog

2.1. Calculation of CO₂ emission

3.1. CO₂ emission measurement for Chinese agricultural sector