Meta-analysis of voice disorders databases and applied machine learning techniques

Sidra Abid Syed; Munaf Rashid; Samreen Hussain; Sidra Abid Syed; Munaf Rashid; Samreen Hussain

doi:10.3934/mbe.2020404

Mathematical Biosciences and Engineering

2020, Volume 17, Issue 6: 7958-7979. doi: 10.3934/mbe.2020404

Previous Article Next Article

Research article Special Issues

Meta-analysis of voice disorders databases and applied machine learning techniques

1.
Biomedical Engineering Department, Ziauddin University Faculty of Engineering Science Technology and Management, Karachi, Pakistan
2.
Software Engineering Department, Ziauddin University Faculty of Engineering Science Technology and Management, Karachi, Pakistan
3.
Vice Chancellor, Begum Nusrat Bhutto Women University, Sukkur, Pakistan

Received: 20 August 2020 Accepted: 25 October 2020 Published: 11 November 2020

Background and ObjectiveVoice disorders are pathological conditions that directly affect voice production. Computer based diagnosis may play a major role in the early detection and in tracking and even development of efficient pathological speech diagnosis, based on a computerized acoustic evaluation. The health of the Voice is assessed by several acoustic parameters. The exactness of these parameters is often linked to algorithms used to estimate them for speech noise identification. That is why main effort of the scientists is to study acoustic parameters and to apply classification methods that achieve a high precision in discrimination. The primary aim of this paper is for a meta-analysis on voice disorder databases i.e. SVD, MEEI and AVPD and machine learning techniques applied on it.

Materials and MethodsThis field of study was systematically reviewed in compliance with PRISMA guidelines. A search was performed with a set of formulated keywords on three databases i.e. Science Direct, PubMed, and IEEE Xplore. A proper screening and analysis of articles were performed after which several articles were also excluded.

ResultsForty-five studies that fulfills the eligibility criteria were included in this meta-analysis. After applying eligibility criteria on the peer reviewed and research article and studies that were published in authentic journals and conferences proceedings till June 2020 were chosen for further full-text screening. In general, only those articles that used voice recordings from SVD, MEEI and AVPD databases as a dataset is included in this meta-analysis.

ConclusionWe discussed the strengths and weaknesses of SVD, MEEI and AVPD. After detailed analysis of the studies including the techniques used and outcome measurements, it was also concluded that Support Vector Machine (SVM) is the most common used algorithm for the detection of voice disorders. Other than was also noticed that researchers focus on supervised techniques for the clinical diagnosis of voice disorder rather than using unsupervised techniques. It was also concluded that more work needs to be on voice pathology detection using AVPD database.

Keywords:

Citation: Sidra Abid Syed, Munaf Rashid, Samreen Hussain. Meta-analysis of voice disorders databases and applied machine learning techniques[J]. Mathematical Biosciences and Engineering, 2020, 17(6): 7958-7979. doi: 10.3934/mbe.2020404

Related Papers:

[1]	Yoichi Enatsu, Yukihiko Nakata . Stability and bifurcation analysis of epidemic models with saturated incidence rates: An application to a nonmonotone incidence rate. Mathematical Biosciences and Engineering, 2014, 11(4): 785-805. doi: 10.3934/mbe.2014.11.785
[2]	Toshikazu Kuniya, Hisashi Inaba . Hopf bifurcation in a chronological age-structured SIR epidemic model with age-dependent infectivity. Mathematical Biosciences and Engineering, 2023, 20(7): 13036-13060. doi: 10.3934/mbe.2023581
[3]	Peter Rashkov, Ezio Venturino, Maira Aguiar, Nico Stollenwerk, Bob W. Kooi . On the role of vector modeling in a minimalistic epidemic model. Mathematical Biosciences and Engineering, 2019, 16(5): 4314-4338. doi: 10.3934/mbe.2019215
[4]	Yongbing Nie, Xiangdong Sun, Hongping Hu, Qiang Hou . Bifurcation analysis of a sheep brucellosis model with testing and saturated culling rate. Mathematical Biosciences and Engineering, 2023, 20(1): 1519-1537. doi: 10.3934/mbe.2023069
[5]	Kunlun Huang, Xintian Jia, Cuiping Li . Analysis of modified Holling-Tanner model with strong Allee effect. Mathematical Biosciences and Engineering, 2023, 20(8): 15524-15543. doi: 10.3934/mbe.2023693
[6]	Juan Ye, Yi Wang, Zhan Jin, Chuanjun Dai, Min Zhao . Dynamics of a predator-prey model with strong Allee effect and nonconstant mortality rate. Mathematical Biosciences and Engineering, 2022, 19(4): 3402-3426. doi: 10.3934/mbe.2022157
[7]	Xiaoli Wang, Junping Shi, Guohong Zhang . Bifurcation analysis of a wild and sterile mosquito model. Mathematical Biosciences and Engineering, 2019, 16(5): 3215-3234. doi: 10.3934/mbe.2019160
[8]	Mengyun Xing, Mengxin He, Zhong Li . Dynamics of a modified Leslie-Gower predator-prey model with double Allee effects. Mathematical Biosciences and Engineering, 2024, 21(1): 792-831. doi: 10.3934/mbe.2024034
[9]	Zhenliang Zhu, Yuming Chen, Zhong Li, Fengde Chen . Dynamic behaviors of a Leslie-Gower model with strong Allee effect and fear effect in prey. Mathematical Biosciences and Engineering, 2023, 20(6): 10977-10999. doi: 10.3934/mbe.2023486
[10]	Fang Wang, Juping Zhang, Maoxing Liu . Dynamical analysis of a network-based SIR model with saturated incidence rate and nonlinear recovery rate: an edge-compartmental approach. Mathematical Biosciences and Engineering, 2024, 21(4): 5430-5445. doi: 10.3934/mbe.2024239

Abstract

1. Introduction

In this paper, we consider the infectious disease transmission models. Denoted by $S(t)$ , $I(t)$ , and $R(t)$ , the numbers of susceptible, infective, and recovered or removed individuals at time $t$ , respectively. The classical susceptible-infective-recovered-susceptible (SIRS) model has the form

$\begin{equation} \left\{ \begin{aligned} &\dot{S} = b-d S-g(I) S+ \delta R, \\ &\dot{I} = g(I) S-(d+\mu) I,\\ &\dot{R} = \mu I-(d+\delta) R, \end{aligned} \right. \end{equation}$

(1.1)

where $b > 0$ and $d > 0$ represent the recruitment rate and the natural death rate of population respectively, $\mu > 0$ expresses the natural recovery rate of the infective individuals, and $\delta > 0$ denotes the rate at which recovered individuals lose immunity and return to the susceptible class. $g(I)S$ is the incidence rate, and $g(I)$ measures the infection force of a disease.

In most epidemic models, the incidence rate adopts the form of mass action with bilinear interactions, namely $g(I)S = k IS$ , where the constant $k$ is the probability of transmission per contact and $g(I) = k I$ is unbounded when $I\geq0$ . The model (1.1) with $g(I) = k I$ usually has at most one endemic equilibrium and has no period orbit. The disease will die out if the basic reproduction number is less than one and will persist otherwise (see Hethcote ^[4]). Recently, different types of nonlinear incidence rates have been applied in the study of epidemic diseases. For example, Liu et al. ^[5] proposed a general nonlinear incidence rate defined by

$\begin{equation} g(I)S = \frac{k I^p S}{1+\alpha I^q}, \end{equation}$

(1.2)

where $k I^p$ measures the infection force of the disease, $1/(1+\alpha I^q)$ describes the inhibition effect from the behavioral change of the susceptible individuals when the number of infectious individuals increases (see Capasso and Serio ^[6]). $\alpha\geq 0$ measures the psychological or inhibitory effect, $p, \, k, \, q$ are real constants with $p > 0, \, \, k > 0, \, \, q\geq 0$ . Note that $g(I)S = kSI$ in (1.2) if $\alpha = 0$ and $p = 1$ .

For general $p$ and $q$ , Hu et al. in ^[1] studied the SIRS model (1.1) with (1.2). For simplicity, they considered the reduced system

$\begin{equation} \left\{ \begin{aligned} &\dot{I} = \frac{k I^p}{1+\alpha I^q} \left(\frac{b}{d}-I-R\right)-(d+\mu) I, \\ &\dot{R} = \mu I-(d+\delta) R. \end{aligned} \right. \end{equation}$

(1.3)

By calculation, $E(I^*, R^*)$ is a positive equilibrium of system (1.3) if and only if $I^* > 0$ satisfies

$\begin{equation} \frac{1}{K}I^p-I^{p-1}+\frac{\alpha}{\sigma}I^{q}+ \frac{1}{\sigma} = 0, \end{equation}$

(1.4)

where

$\begin{equation*} K = \frac{b(d+\delta)}{d(d+\delta+\mu)},\,\, \sigma = \frac{k b}{d(\mu+d)}. \end{equation*}$

It was shown in ^[1] that the model (1.3) can have very rich and complex dynamical behaviors. More precisely, system (1.3) can have multiple endemic equilibria and different types of bifurcations, including Hopf and Bogdanov-Takens bifurcations. Note that the expressions of the positive equilibria cannot be explicitly expressed for general $p$ and $q$ . They only gave the calculation formula of the first Lyapunov constant of the unique positive equilibrium (if it is linearly a center), which can be found in many books. In addition, the first Lyapunov constant is a very complex function of $I^*$ . In other words, the exact codimension of Hopf bifurcation remains unknown. Furthermore, the conditions for determining the codimension of Bogdanov-Takens bifurcation are complex functions of $I^*$ . Thus, they can not determine the maximum codimension of Bogdanov-Takens bifurcation.

In the paper ^[3], the nonlinear function $g(I)$ is classified into the three types: unbounded incidence function with $p > q$ , saturated incidence function with $p = q$ , and nonmonotone incidence function $p < q$ . The results on the dynamics of system (1.1) can be found in ^{[2,3,6,7,8,9,10,11]} and reference therein.

For saturated incidence function with $p = q$ , as early as $1991$ , Hethcote and van den Driessche ^[7] showed that Hopf bifurcation can occur in system (1.1) with saturated incidence rate (1.2) (when $q = p$ ). For similar reasons as in ^[1], they did not determine the exact codimension of Hopf bifurcation. For the special case of $q = p = 1$ , Gomes et al. ^[12] showed the existence of backward bifurcations, oscillations, and Bogdanov-Takens points in SIR and SIS models. For the special case of $q = p = 2$ , system (1.1) with (1.2) was analysed by Ruan and Wang in ^[2]. It was shown that it can undergo Bogdanov-Takens bifurcation of codimension two and Hopf bifurcation of codimension one. Later, Tang et al. in ^[3] calculated the higher order Lyapunov constants of the weak focus and proved that the maximal order of the weak focus is two. They also obtained the coexistence of a limit cycle and a homoclinic loop. Recently, Lu et al. ^[13] considered a more general model and proved that the codimension of Bogdanov-Takens bifurcation is at most two. These results indicate that the case $q = p > 1$ may be very complicated and deserves further investigation.

Although a lot of work on SIRS models have been obtained, there are still a lot of open problem for it. In order to understand the dynamical behavior of system (1.1) and also motivated by the works in ^[1], ^[2], ^[3], ^[7] and ^[13], we focused on the model (1.1) with a general saturated incidence rate $k I^p/(1+\alpha I^p)$ , i.e., the case $q = p$ in (1.2). Based on the results of ^[1], the model with $p\leq 1$ has the simple dynamics. Hence, we study the SIRS epidemic model (1.1) with $q = p > 1$ in this paper, i.e., the system defined as

$\begin{equation} \left\{ \begin{aligned} &\dot{S} = b-d S-\dfrac{k S I^p}{1+ \alpha I^p}+ \delta R, \\ &\dot{I} = \dfrac{k S I^p}{1+ \alpha I^p}-(d+\mu) I, \\ &\dot{R} = \mu I-(d+\delta) R, \end{aligned} \right. \end{equation}$

(1.5)

where $S(0)\geq 0$ , $I(0)\geq 0$ , $R(0)\geq 0$ and $k, \alpha, b, d, \mu, \delta > 0, p > 1$ .

For general $p > 1$ , to study the model (1.5), one of the difficulties is that the coordinate of positive equilibria cannot be explicitly expressed for $p$ . The second difficulty is that the model of this type is often not a polynomial differential system or can not transformed into a polynomial differential system. Finally, another difficulty is the lack of the methods to determine the order of a weak focus. In this paper, we will provide some available methods and techniques to overcome these difficulties, and perform qualitative and bifurcation analysis of system (1.5). We find that system (1.5) have the complicated dynamical behaviors and bifurcation phenomena like the special case of $p = 2$ . It is shown that the codimension of Bogdanov-Takens bifurcation is at most two, which coincides with the corresponding results for $p = 2$ in ^[13]. In addition, This also improves the results for $p = 2$ in ^[2] and ^[3]. Moreover, our results on Hopf bifurcation coincides with the ones for $p = 2$ in ^[3] and ^[13], and also can be seen as a complement of the results for $p = 2$ obtained in ^[1] and ^[2]. Furthermore, there exist some critical values $\alpha = \alpha_0$ (i.e., $b = b_0$ or $\delta = \delta_0$ ) and $\delta = \delta_2$ such that: (i) if $\alpha > \alpha_0$ (i.e., $b < b_0$ , or $\delta < \delta_0$ ), the disease will die out; (ii) if $\alpha = \alpha_0$ (i.e., $b = b_0$ , or $\delta = \delta_0$ ) and $\delta\leq \delta_2$ , the disease will die out for almost all positive initial populations; (iii) if $\alpha = \alpha_0$ (i.e., $b = b_0$ , or $\delta = \delta_0$ ) and $\delta > \delta_2$ , the disease will persist in the form of a positive coexistent steady state for some positive initial populations; and (iv) if $\alpha < \alpha_0$ , the disease will persist in the form of multiple positive periodic coexistent oscillations and coexistent steady states for some positive initial populations.

Though the SIRS epidemic models with nonlinear incidence rates have been extensively studied, to the best of our knowledge, this is the first time that the exact codimension of Hopf bifurcation and Bogdanov-Takens bifurcation have been determined in epidemic model for general parameter. It is worth mentioning that the difficulties of study of system (1.5), as mentioned above, are not only a common problem of other infectious disease models, but also a common problem of chemical molecular reaction models, physical systems, etc. The methods and techniques developed in this paper can be applied to study most of the complex dynamical systems.

The rest of this paper is organized as follows. In Section $2$ , we give some preliminary results. Section $3$ is devoted to study saddle-node bifurcation, Bogdanov-Takens bifurcation and Hopf bifurcation, and illustrate these results by simulation.

2. Preliminary results

In this section, we are going to provide some preliminary results, which are the basis for proving the main results.

2.1. Model reduction

To simplify the system (1.5), we first give the following lemma.

Lemma 2.1. System $(1.5)$ has the invariant manifold defined by $S(t)+I(t)+R(t) = b/d$ , which is attracting in the first quadrant.

Proof. Let $N(t) = S(t)+I(t)+R(t)$ . It follows from (1.5) that

$\begin{equation*} \begin{aligned} \dot{N} = b -d N. \end{aligned} \end{equation*}$

Obviously, $N(t) = b/d$ is a stable equilibrium of the above equation, which implies the conclusion.

It follows from Lemma 2.1 that the plane $S(t)+I(t)+R(t) = b/d$ is a limit set of system (1.5) and all important dynamical behaviors of (1.5) are on this plane. Thus, in the rest of this paper we consider the following limit system in the first quadrant

$\begin{equation} \left\{ \begin{aligned} &\dot{I} = \dfrac{k I^p}{1+ \alpha I^p} \left(\dfrac{b}{d}- I- R\right)-(d+\mu) I, \\ &\dot{R} = \mu I-(d+\delta) R. \end{aligned} \right. \end{equation}$

(2.1)

Denote

$\begin{equation*} \begin{aligned} \mathcal{D} = \left\{(I,R)\mid I\geq 0, R\geq0, I+R\leq \frac{b}{d}\right\}. \end{aligned} \end{equation*}$

Obviously $\mathcal{D}$ is a positive invariant set of system (2.1).

Take the scalings

$\begin{equation*} \begin{aligned} I = \left(\dfrac{d+\delta}{k}\right)^{\frac{1}{p}} x,\,\, R = \left(\dfrac{d+\delta}{k}\right)^{\frac{1}{p}} y,\,\, t = \dfrac{1}{d+\delta} \tau, \end{aligned} \end{equation*}$

to get

$\begin{equation} \left\{ \begin{aligned} &\dot{x} = \dfrac{x^p}{1+ \beta x^p} (A- x- y)- m x, \\ &\dot{y} = n x- y, \end{aligned} \right. \end{equation}$

(2.2)

where we rewrite $\tau$ into $t$ , and

$\begin{equation} \begin{aligned} \beta = \alpha \dfrac{d+\delta}{k},\,\, A = \dfrac{b}{d} \left(\dfrac{k}{d+\delta}\right)^{\frac{1}{p}},\,\, m = \dfrac{d+\mu}{d+\delta},\,\, n = \dfrac{\mu}{d+\delta}, \end{aligned} \end{equation}$

(2.3)

which give

$\begin{equation} \begin{aligned} \beta > 0, \quad A > 0, \quad p > 1,\quad m > n > 0. \end{aligned} \end{equation}$

(2.4)

The positively invariant set $\mathcal{D}$ becomes

$\begin{equation*} \begin{aligned} \mathcal{D} = \left\{(x,y)\mid x\geq 0, y\geq0, x+y\leq A \right\}. \end{aligned} \end{equation*}$

The dynamics generated by system (2.2) and (2.1) in $\mathcal{D}$ are topological equivalent. Hence we only need to study system (2.2) in the region $\mathcal{D}$ with parameters satisfying the conditions (2.4).

Remark 2.2. In the rest of this paper, we study system $(2.2)$ in $\mathcal{D}$ with the assumption that the inequlities in $(2.4)$ hold.

2.2. Analysis of equilibria of system (2.2)

$E_0(0, 0)$ is always an equilibrium of system (2.2). The Jacobian matrix of system (2.2) at $E_0$ is

$J(E_0) = \left(\begin{array}{cc} -m & 0\\ n & -1 \end{array}\right),$

which has two negative eigenvalues $-m$ and $-1$ . Thus, $E_0(0, 0)$ is a stable hyperbolic node.

Denote by $E(\bar{x}, \bar{y})$ a positive equilibrium of system (2.2). Then

$\begin{equation*} \begin{aligned} \dfrac{x^p}{1+\beta x^p} (A-x-n x)-m x = 0,\,\,\,\bar{y} = n\bar{x} \end{aligned} \end{equation*}$

in the interval $(0, A/(n+1))$ , or

$\begin{equation} \begin{aligned} h(\bar{x}) = 0,\quad \bar{x}\leq \frac{A}{n+1}, \end{aligned} \end{equation}$

(2.5)

where

$\begin{equation} \begin{aligned} h(x) = (m \beta+ n+1) x^{p}-A x^{p-1}+ m. \end{aligned} \end{equation}$

(2.6)

A direct computation shows that

$\begin{equation} \begin{aligned} h'(x) = p(m \beta+n+1)(x-x_{*}) x^{p-2}, \end{aligned} \end{equation}$

(2.7)

where

$\begin{equation} \label{x_{*}} 0 < x_{*} = \frac{A (p-1)}{p(m \beta+n+1)}\leq \frac{A}{n+1}. \end{equation}$

(2.8)

It is obvious that $h'(x)$ has a unique positive root at $x = x_{*}$ . Therefore, $h(x)$ has no positive zero if $h(x_{*}) > 0$ , or has a unique positive zero at $x = x_{*}$ if $h(x_{*}) = 0$ , or has two positive zeros at $x = x_1$ , $x_2$ with $0 < x_1 < x_{*} < x_2$ if $h(x_{*}) < 0$ . Notice that $h(A/(n+1)) > 0$ , which means $x_2 < A/(n+1)$ . By calculation, $h(x_*) < 0$ is equivalent to $A > A_{*}$ , i.e., $\beta < \beta_{*}$ , or $n < n_{*}$ , where

$\begin{equation} \label{A_{*}} \begin{aligned} &A_{*} = p m^{\frac{1}{p}}\left(\frac{m \beta+n+1}{p-1}\right)^{1-\frac{1}{p}},\\ &\beta_{*} = \frac{1}{m}\left[\frac{p-1}{p}A^{\frac{p}{p-1}} (mp)^{-\frac{1}{p-1}}-n-1\right],\\ &n_{*} = \frac{p-1}{p}A^{\frac{p}{p-1}} (mp)^{-\frac{1}{p-1}}-m\beta-1. \end{aligned} \end{equation}$

(2.9)

Theorem 2.3. Let $x_{*}$ , $A_{*}$ , $\beta_{*}$ and $n_{*}$ are given by $(2.8)$ and $(2.9)$ . System $(2.2)$ always has a disease-free equilibrium $E_0(0, 0)$ . Moreover, for system $(2.2)$ ,

(I) if $A < A_{*}$ (i.e., $\beta > \beta_{*}$ , or $n > n_{*}$ ), then there is no positive equilibrium;

(II) if $A = A_{*}$ (i.e., $\beta = \beta_{*}$ or $n = n_{*}$ ), then there is a unique positive equilibrium at $E_*(x_{*}, y_{*})$ ;

(III) if $A > A_{*}$ (i.e., $\beta < \beta_{*}$ , or $n < n_{*}$ ), then there are two positive equilibria at $E_1(x_1, y_1)$ and $E_2(x_2, y_2)$ , where $0 < x_1 < x_{*} < x_2 < 2x_{*}$ .

Notice that $E_0(0, 0)$ is a stable node. According to the index theory, since $\mathcal{D}$ is positively invariant, system (2.2) has no limit cycle in $\mathcal{D}$ if it has no equilibrium. By Theorem 2.3(I), we have the following result.

Theorem 2.4. The disease-free equilibrium $E_0(0, 0)$ of system $(2.2)$ is globally asymptotical stable in $\mathcal{D}$ if $A < A_{*}$ (i.e., $\beta > \beta_{*}$ , or $n > n_{*}$ ).

By Theorem 2.4, we get the following corollary.

Corollary 2.5. The disease-free equilibrium $(b/d, 0, 0)$ of system $(1.5)$ is globally asymptotical stable in the interior $\mathbb{R}^3_+$ and the disease will die out if $A < A_{*}$ (i.e., $\beta > \beta_{*}$ , or $n > n_{*}$ ).

Remark 2.6. From $(2.3)$ and Theorem 2.4, we obtain that $A < A_{*}$ is equivalent to $\alpha > \alpha_0$ or $b < b_0$ or $\delta < \delta_0$ by calculation, where

$\begin{equation*} \begin{aligned} &\alpha_0 = \frac{k}{d+\mu}\left[(p-1)\left(\frac{b}{pd}\right)^{\frac{p}{p-1}} \left(\frac{k}{d+\mu}\right)^{\frac{1}{p-1}}-\frac{\mu }{d+\delta} -1\right],\\ &b_0 = \frac{d p}{k} (p-1)^{\frac{1}{p}-1} (d+\mu)^{\frac{1}{p}} \left[\alpha (d+\mu)+\frac{\mu k }{d+\delta} +k \right]^{1-\frac{1}{p}},\\ &\delta_0 = \mu\left[(p-1)\left(\frac{b}{pd}\right)^{\frac{p}{p-1}} \left(\frac{k}{d+\mu}\right)^{\frac{1}{p-1}}-\frac{\alpha(d+\mu)}{k} -1\right]^{-1}. \end{aligned} \end{equation*}$

Therefore, the disease will die out if either $\alpha > \alpha_0$ , or $b < b_0$ , or $\delta < \delta_0$ .

Remark 2.7. When $p = 2$ , it follows from Remark 2.6 that

$\begin{equation*} \begin{aligned} \alpha_0 = \frac{b^2k^2(d+\delta)-4d^2(d+\mu)(d+\mu+\delta)}{4d^2(d+\mu)^2(d+\delta)}, \end{aligned} \end{equation*}$

which coincides with the $\alpha_0$ in Remark 2.1 of ^[13] when $\beta = 0$ .

Next consider the positive equilibrium $E(\bar{x}, \bar{y})$ of system (2.2), whose coordinate satisfies

$\begin{equation*} \bar{y} = n\bar{x}, \,\,\,\, \bar{x}^{p-1} (A-\bar{x}-\bar{y}) = m \left(1+\beta \bar{x}^p\right). \end{equation*}$

The Jacobian matrix of system (2.2) at $E(\bar{x}, \bar{y})$ is

$J(E) = \left(\begin{array}{cc} J_{11} & J_{12}\\ n & -1 \end{array}\right),$

where

$\begin{align*} & J_{11} = \dfrac{\left[pm \left(1+\beta \bar{x}^p\right)-\bar{x}^p\right] -m \beta p \bar{x}^{p}}{1+\beta \bar{x}^p}-m = \frac{m(p-1)-(m\beta+1)\bar{x}^{p}}{1+\beta \bar{x}^p}, \\ & J_{12} = -\dfrac{\bar{x}^p}{1+\beta \bar{x}^p}, \end{align*}$

which gives

$\begin{equation*} \begin{aligned} \text{Det}\left(J(E)\right) = \dfrac{ p(m \beta+n+1)(\bar{x}-x_*) \bar{x}^{p-1}}{1+\beta \bar{x}^p} = \dfrac{ \bar{x} h'(\bar{x})}{1+\beta \bar{x}^p}, \end{aligned} \end{equation*}$

and its sign is determined by $h'(\bar{x})$ , where $h'(x)$ is defined in (2.7).

The trace of $J(E)$ is

$\begin{equation*} \begin{aligned} \text{Tr}\left(J(E)\right) = \dfrac{1}{1+\beta \bar{x}^p}S_T(\bar{x}), \end{aligned} \end{equation*}$

where

$\begin{equation} \begin{aligned} S_T(\bar{x}) = (m p-m-1)-(m \beta+\beta+1) \bar{x}^{p}. \end{aligned} \end{equation}$

(2.10)

To study the positive equilibria, let

$\begin{equation*} p_{1} = 1+\frac{1}{m},\,\,\, n_{1} = \frac{mp\beta+1}{mp-m-1},\,\,(p\neq p_1). \end{equation*}$

Note that $p > p_{1}$ if and only if $n_{1} > 0$ .

Theorem 2.8. If $A = A_{*}$ , then system $(2.2)$ has a unique equilibrium at $E_{*}(x_{*}, y_{*})$ in $\mathcal{D}$ .

(I) If (i) $1 < p\leq p_{1}$ , or (ii) $p > p_{1}$ , $n\neq n_{1}$ , then $E_{*}$ is a saddle-node.

(II) If $p > p_{1}$ and $n = n_{1}$ , then $E_{*}$ is a cusp of codimension two.

Proof. If $A = A_*$ , it follows from Theorem 2.3 that system (2.2) has a unique positive equilibrium $E_{*}(x_{*}, y_{*})$ , where $x_{*}$ is given by (2.8) satisfying $h(x_{*}) = h'(x_{*}) = 0$ , and $y_{*} = nx_{*}$ . This implies $\text{Det}\left(J(E_{*})\right) = 0$ . It follows from $h(x_{*}) = 0$ and (2.8) that

$\begin{equation} x_{*}^{p} = \frac{m}{Ax_{*}^{-1}-(m\beta+n+1)} = \frac{m(p-1)}{m\beta+n+1}. \end{equation}$

(2.11)

Substituting it into $S_T(x_{*})$ , we get

$\begin{equation*} \begin{aligned} S_T(x_{*}) = \dfrac{(mp-m-1)n-mp\beta-1}{m\beta+n+1}. \end{aligned} \end{equation*}$

If $mp-m-1\leq 0$ , i.e., $1 < p\leq p_{1}$ , then $S_T(x_{*}) < 0$ . Assume that $mp-m-1 > 0$ , i.e., $p > p_{1}$ , then $S_T(x_{*}) < 0$ if and only if $n < n_{1}$ , $S_T(x_{*}) > 0$ if and only if $n > n_{1}$ and $S_T(x_{*}) = 0$ if and only if $n = n_{1}$ , respectively.

(I) Suppose either $1 < p\leq p_{1}$ , or $p > p_{1}$ and $n\neq n_{1}$ . Then $\text{Tr}\left(J(E_*)\right)\not = 0$ and $\text{Det}\left(J(E_*)\right) = 0$ , which imply that $0$ and $J_{11}(E_*)-1 = \text{Tr}(J(E_*))\not = 0$ are eigenvalues of the matrix $J(E_*)$ . Taking the changes $X = x-x_{*}, \, \, Y = y-y_{*}$ and $\bar X = -n X+J_{11}(E_*)Y, \, \, \bar Y = -n X+Y, \, \, \tau = \text{Tr}\left(J(E_*)\right)t$ , system $(2.3)$ is written as (for simplicity, still denote $\bar X, \, \bar Y, \, \tau$ by $x, \, y, \, t$ , respectively)

$\begin{equation} \left\{ \begin{aligned} \dot{x} = &\eta_1x^2+\eta_2xy+\eta_3x^3+\eta_4x^2y+o(\mid(x,y)^3\mid) = P_2(x,y),\\ \dot{y} = &y+\eta_1x^2+\eta_2xy+\eta_3x^3+\eta_4x^2y+o(\mid(x,y)^3\mid) = y+Q_2(x,y), \end{aligned} \right. \end{equation}$

(2.12)

where $\eta_2, \eta_3, \eta_4$ are real numbers, and it follows from (2.11) that

$\begin{align*} \eta_1 = &\frac{p^2x_*^p(m\beta+n+1)\left\{\beta [\beta m p (p+1)+(n+1) (3 p-1)] x_*^p+(p-1) (-\beta m p+n+1)\right\}}{2n\left( \text{Tr}\left(J(E_*)\right)\right)^3A(p-1)^2(1+\beta x_*^p)^2},\\ = &\frac{p^2x_*^p(m\beta+n+1)(\beta mp+n+1)^2}{2n\left( \text{Tr}\left(J(E_*)\right)\right)^3A(p-1)(1+\beta x_*^p)^2}\not = 0. \end{align*}$

Solving the equation $y+Q_2(x, y) = 0$ , we get that $y(x) = -\eta_1x^2+\cdots$ , which implies that $P_2(x, y(x)) = \eta_1x^2+\cdots$ . By Theorem 7.1 of Chapter 2 in ^[14], the origin is a saddle-node of system (2.12). This proves the assertion (I).

(II) Suppose that $A = A_{*}$ and $n = n_{1}$ . Let $X = x-x_{*}, \, Y = y-y_{*}$ . Then system (2.2) takes the form (rewrite $X$ , $Y$ into $x, y$ , respectively)

$\begin{eqnarray} \begin{array}{ll} \dot{x} = x+a_1 y+a_2 x^2+a_3 xy+o(\mid(x,y)^2\mid), \\[6pt] \dot{y} = n_1 x-y, \end{array} \end{eqnarray}$

(2.13)

where

$\begin{equation*} \begin{aligned} a_1 = -\dfrac{1}{n_1} ,\,\, a_2 = \dfrac{(2+m-mp)p(n_1-\beta)^{\frac{1}{p}+1}}{2 n_1} ,\,\, a_3 = \dfrac{-p (n_1-\beta)^{\frac{1}{p}+1}}{n_1^2}. \end{aligned} \end{equation*}$

Let $X = x$ , $Y = x-y/n_1$ to get (rewrite $X$ , $Y$ into $x, y$ , respectively)

$\begin{eqnarray} \begin{array}{ll} \dot{x} = y+ (a_2+a_3 n_1) x^2-a_3 n_1 xy+o(\mid(x,y)^2\mid), \\[6pt] \dot{y} = (a_2+a_3 n_1) x^2- a_3 n_1 xy+o(\mid(x,y)^2\mid). \end{array} \end{eqnarray}$

(2.14)

Taking the change $z = y+ (a_2+a_3 n_1) x^2-a_3 n_1 xy+o(\mid(x, y)^2\mid), t = (1+ a_3n_1 x)\tau$ , system $(2.14)$ is reduced to the following system in the small neighborhood of $(0, 0)$

$\begin{eqnarray*} \begin{array}{ll} \dfrac{dx}{d\tau} = z(1+a_3n_1 x), \\[6pt] \dfrac{dz}{d\tau} = (1+a_3n_1 x)((a_2+a_3 n_1) x^2+(2 a_2+a_3 n_1) xz- a_3 n_1z^2+o(\mid(x,y)^2\mid). \end{array} \end{eqnarray*}$

Let $Y = z(1+a_3n_1 x)$ and rewrite $Y$ as $y$ . We have

$\begin{eqnarray} \begin{array}{ll} \dot{x} = y, \\[6pt] \dot{y} = (a_2+a_3 n_1) x^2+(2 a_2+a_3 n_1) xy+o(\mid(x,y)^2\mid), \end{array} \end{eqnarray}$

(2.15)

where

$\begin{equation*} \begin{aligned} a_2+a_3 n_1 = \dfrac{m(1-p)p(n_1-\beta)^{1+\frac{1}{p}}}{2 n_1},\,\, 2 a_2+a_3 n_1 = -\dfrac{p(1+\beta+m\beta)(n_1-\beta)^{\frac{1}{p}}}{n_1}. \end{aligned} \end{equation*}$

Since $n_1-\beta = (1+\beta+m\beta)/(mp-m-1)\neq 0$ , we have $a_2+a_3 n_1\neq 0$ and $2 a_2+a_3 n_1\neq 0$ . By the results in ^[15], $E_{*}(x_{*}, y_{*})$ is a cusp of codimension two.

Theorem 2.9. If $A > A_{*}$ (i.e., $\beta < \beta_{*}$ or $n < n_{*}$ ), system $(2.2)$ has two positive equilibria $E_1(x_1, y_1)$ and $E_2(x_2, y_2)$ , where $0 < x_1 < x_{*} < x_2 < 2x_{*}$ . Moreover, $E_1$ is always a saddle point, and $E_2$ is

(I) a stable focus (or node) if $S_T(x_2) < 0$ ; or

(II) a weak focus (or a center) if $S_T(x_2) = 0$ ; or

(III) an unstable focus (or node) if $S_T(x_2) > 0$ , where $S_T$ is given in $(2.10)$ .

Proof. The first half of the theorem holds by Theorem 2.3. To determine the types of $E_1$ and $E_2$ , consider the signs of $h'(x_1)$ , $h'(x_2)$ , $S_T(x_1)$ and $S_T(x_2)$ , where $h'(x)$ , $S_T(x)$ are given in $(2.7)$ and $(2.10)$ , respectively. Noting $0 < x_1 < x_* < x_2$ (cf. Theorem 2.3), we have $h'(x_1) < 0 < h'(x_2)$ . This follows that the Jacobian determinants at $E_1$ and $E_2$ satisfy $\text{Det}\left(J(E_1)\right) < 0$ and $\text{Det}\left(J(E_2)\right) > 0$ , respectively. Hence we obtain the types of $E_1$ and $E_2$ .

3. Bifurcation analysis

We are going to investigate various bifurcations in system $(2.2)$ in this section.

3.1. Saddle-node bifurcation

From Theorem 2.8, we know that the surface

$\begin{equation*} SN = \{(A,m,n,\beta)\mid A = A_*, \mathrm{either}\,\, 1 < p\leq p_{1}, \mathrm{or}\,\, p > p_{1}, n\neq n_{1}\} \end{equation*}$

is the saddle-node bifurcation surface under the assumption (2.4). On one side of this surface there is no equilibrium and on the other side there are two equilibria.

3.2. Bogdanov-Takens bifurcation

In this subsection, we study the Bogdanov-Takens bifurcation of codimension two for system (2.2).

Theorem 3.1. If $A = A_{*}$ , $n = n_1$ , $p > p_1$ and conditions $(2.4)$ hold, then the unique positive equilibrium $E_*(x_*, y_*)$ is a cusp of codimension two (i.e., Bogdanov-Takens singularity), and system $(2.2)$ undergoes Bogdanov-Takens bifurcation of codimension two in a small neighborhood of $E_*(x_*, y_*)$ when we choose $A$ and $n$ as bifurcation parameters. Hence, there are different parameter values such that system $(2.2)$ has an unstable limit cycle or an unstable homoclinic loop.

Proof. Consider the system

$\begin{equation} \begin{aligned} \dot{x} = &\dfrac{x^p}{1+ \beta x^p} (A_*+\lambda_1- x- y)- mx,\\ \dot{y} = &(n_1+\lambda_2) x- y, \end{aligned} \end{equation}$

(3.1)

in a small neighborhood of the equilibrium $E_*(x_*, y_*)$ , where $\lambda_1$ and $\lambda_2$ are small parameters. To convenience, in below the functions $P_1(x, y, \lambda_1, \lambda_2)$ and $Q_i(x, y, \lambda_1, \lambda_2), \, \, i = 1, 2, 3, 4$ , are $C^\infty$ functions at least of third order with respect to $(x, y)$ , whose coefficients depend smoothly on $\lambda_1$ and $\lambda_2$ , .

Taking the changes $X = x-x_*$ , $Y = y-y_*$ , system (3.1) is rewritten as (we still denote $X$ , $Y$ by $x$ , $y$ , respectively)

$\begin{eqnarray} \begin{array}{ll} \dot{x} = b_1+b_2 x- \dfrac{1}{n}y+b_3 x^2+ b_4 xy+P_1(x,y,\lambda_1,\lambda_2), \\[6pt] \dot{y} = b_5 +b_6 x- y, \end{array} \end{eqnarray}$

(3.2)

where

$\begin{equation*} \begin{aligned} b_1 = \dfrac{\lambda_1}{n_1},\, b_2 = 1- a_3\lambda_1,\, b_3 = a_2-\frac{p(n_1-\beta)^{1+\frac{2}{p}}(2p\beta+n_1-pn_1)}{2n_1^3}\lambda_1,\, b_4 = a_3,\\ \end{aligned} \end{equation*}$

$\begin{equation*} b_5 = (n_1-\beta)^{-\frac{1}{p}}\lambda_2,\, b_6 = n_1+\lambda_2. \end{equation*}$

Let

$\begin{equation*} \begin{aligned} X = x,\,\, Y = b_1+b_2 x- \dfrac{1}{n_1}y+b_3 x^2+ b_4 xy+P_1(x,y,\lambda_1,\lambda_2). \end{aligned} \end{equation*}$

System (3.2) takes the form (we rewrite $X$ , $Y$ as $x$ , $y$ , respectively)

$\begin{eqnarray} \begin{array}{ll} \dot{x} = y, \\[6pt] \dot{y} = c_1+ c_2 x+ c_3 y+ c_4 x^2+ c_5 xy+ c_6 y^2+ Q_1(x,y,\lambda_1,\lambda_2), \end{array} \end{eqnarray}$

(3.3)

with

$\begin{align*} c_1 = \frac{1}{n_1}\lambda_1-\frac{(n_1-\beta)^{-\frac{1}{p}}}{n_1}\lambda_2+O(|\lambda_1,\lambda_2|^3),&\; \; c_2 = -a_3\lambda_1+(a_3(n_1-\beta)^{-\frac{1}{p}}-\frac{1}{n_1})\lambda_2+O(|\lambda_1,\lambda_2|^2),\\ c_3 = O(|\lambda_1,\lambda_2|^2),&\; \; c_4 = a_2+a_3n_1+O(|\lambda_1,\lambda_2|),\\ c_5 = 2a_2+a_3n_1+O(|\lambda_1,\lambda_2|),&\; \; c_6 = -a_3n_1+O(|\lambda_1,\lambda_2|). \end{align*}$

Next let $dt = (1-c_6 x)d \tau$ . System (3.3) takes the form (still denote $\tau$ by $t$ )

$\begin{eqnarray} \begin{array}{ll} \dot{x} = y(1-c_6 x), \\[6pt] \dot{y} = (1-c_6 x)(c_1+ c_2 x+ c_3 y+ c_4 x^2+ c_5 xy+ c_6 y^2+ Q_1(x,y,\lambda_1,\lambda_2). \end{array} \end{eqnarray}$

(3.4)

Let $X = x$ , $Y = y(1-c_6 x)$ , and rewrite $X$ , $Y$ as $x$ , $y$ respectively. System (3.4) becomes

$\begin{eqnarray} \begin{array}{ll} \dot{x} = y, \\[6pt] \dot{y} = d_1+ d_2 x+ d_3 y+ d_4 x^2+ d_5 xy+ Q_2(x,y,\lambda_1,\lambda_2), \end{array} \end{eqnarray}$

(3.5)

where

$\begin{equation*} \begin{aligned} d_1 = c_1,\,\, d_2 = c_2-2 c_1 c_6,\,\, d_3 = c_3,\,\, d_4 = c_4-2 c_2 c_6+ c_1 {c_6}^2,\,\, d_5 = c_5-c_3c_6. \end{aligned} \end{equation*}$

If $\lambda_1 = \lambda_2 = 0$ , then by direct computation and the proof of Theorem 2.8, we get

$\begin{equation*} \begin{aligned} d_1 = 0,\,\, d_2 = 0,\,\, d_3 = 0,\,\, d_4 = a_2+a_3 n_1\neq 0,\,\, d_5 = 2a_2+a_3n_1\neq 0. \end{aligned} \end{equation*}$

Further, let $X = x+d_2/(2d_4), \, \, Y = y$ . System (3.5) becomes (we rewrite $X$ , $Y$ as $x$ , $y$ , respectively)

$\begin{eqnarray} \begin{array}{ll} \dot{x} = y, \\[6pt] \dot{y} = e_1+ e_2 y+ e_3 x^2+ e_4 xy+ Q_3(x,y,\lambda_1,\lambda_2), \end{array} \end{eqnarray}$

(3.6)

where

$\begin{equation*} \begin{aligned} e_1 = d_1-\dfrac{d_{2}^2}{4 d_4},\,\, e_2 = d_3-\dfrac{d_2 d_5}{2 d_4},\,\, e_3 = d_4,\,\, e_4 = d_5. \end{aligned} \end{equation*}$

Finally, taking the changes

$\begin{equation*} \begin{aligned} X = \dfrac{e_{4}^2}{e_3} x,\,\, Y = \dfrac{e_{4}^3}{e_{3}^2} y,\,\, \tau = \dfrac{e_3}{e_4}t, \end{aligned} \end{equation*}$

one obtains (we rewrite $X$ , $Y$ as $x$ , $y$ , respectively)

$\begin{eqnarray} \begin{array}{ll} \dot{x} = y, \\[6pt] \dot{y} = \mu_1+ \mu_2 y+ x^2+ xy+ Q_4(x,y,\lambda_1,\lambda_2), \end{array} \end{eqnarray}$

(3.7)

where

$\begin{equation*} \begin{aligned} \mu_1 = \dfrac{e_1 e_{4}^4}{e_{3}^3},\,\, \mu_2 = \dfrac{e_2 e_4}{e_3}. \end{aligned} \end{equation*}$

We can express $\mu_1$ and $\mu_2$ in terms of $\lambda_1$ and $\lambda_2$ as follows:

$\begin{eqnarray} \begin{array}{ll} \ \mu_1 = s_1\lambda_1 +s_2\lambda_2+o({\mid(\lambda_1,\lambda_2)\mid}), \\[6pt] \ \mu_2 = t_1\lambda_1 +t_2\lambda_2+o({\mid(\lambda_1,\lambda_2)\mid}), \end{array} \end{eqnarray}$

(3.8)

where

$\begin{align*} s_1 = &-\frac{8p(n_1-\beta)^{1+\frac{1}{p}}(mp-m-1)^4}{m^3(p-1)^3n_1^2},\, s_2 = \frac{8p(n_1-\beta)(mp-m-1)^4}{m^3(p-1)^3n_1^2},\\ t_1 = &\frac{2p(mp-m-1)^2(n_1-\beta)^{1+\frac{1}{p}}}{m^2(p-1)^2n_1^2},\, t_2 = -\frac{2(mp-m-1)(p+p\beta-1)}{m^2(p-1)^2n_1^2}.\\ \end{align*}$

Since

$\begin{equation*} \begin{aligned} \left|\dfrac{\partial(\mu_1,\mu_2)}{\partial(\lambda_1,\lambda_2)}\right|_{(\lambda_1,\lambda_2) = (0,0)} = -\frac{16p(mp-m-1)^5(1+\beta+m\beta)(n_1-\beta)^{\frac{1}{p}}}{m^5(p-1)^5n_1^3}\neq 0, \end{aligned} \end{equation*}$

for $p > p_1$ (i.e., $mp-m-1 > 0$ ), $n_1-\beta\neq 0$ , $p > 1$ , $m > n_1 > 0$ , $\beta > 0$ , the change (3.8) is a homeomorphism in a small neighborhood of $(0, 0)$ . According to the results in ^[16,17,18], system (3.7) (i.e., (3.1) or (2.2)) undergoes Bogdanov-Takens bifurcation if $(\lambda_1, \lambda_2)$ changes in a small neighborhood of the origin.

By the results in ^[19], we obtain the local representation of the bifurcation curves as follows:

(i) The saddle-node bifurcation curve is

$\begin{align*} SN = &\{(\mu_1, \mu_2)| \mu_1 = 0, \mu_2 \neq 0\} = \{(\lambda_1,\lambda_2)|-\frac{8p(n_1-\beta)^{1+\frac{1}{p}}(mp-m-1)^4}{m^3(p-1)^3n_1^2}\lambda_1\\ &+\frac{8p(n_1-\beta)(mp-m-1)^4}{m^3(p-1)^3n_1^2}\lambda_2+o({\mid(\lambda_1,\lambda_2)\mid}) = 0,\mu_2 \neq 0\}. \end{align*}$

(ii) The Hopf bifurcation curve is

$\begin{align*} H = &\{(\mu_1, \mu_2)| \mu_2 = \sqrt {-\mu_1}, \mu_1 < 0\} = \{(\lambda_1,\lambda_2)|-\frac{8p(mp-m-1)^4(n_1-\beta)^{1+\frac{1}{p}}}{m^3(p-1)^3n_1^2}\lambda_1\\ &+\frac{8(mp-m-1)^3p(1+\beta+m\beta)}{m^3(p-1)^3n_1^2}\lambda_2+o(|\lambda_1,\lambda_2|) = 0, \mu_1 < 0 \}. \end{align*}$

(iii) The homoclinic bifurcation curve is

$\begin{align*} HL = &\{(\mu_1, \mu_2)| \mu_2 = \frac{5}{7}\sqrt {-\mu_1}, \mu_1 < 0\} = \{(\lambda_1,\lambda_2)| -\frac{200p(mp-m-1)^4(n_1-\beta)^{1+\frac{1}{p}}}{49m^3(p-1)^3n_1^2}\lambda_1\\ &+\frac{200(mp-m-1)^3p(1+\beta+m\beta)}{49m^3(p-1)^3n_1^2}\lambda_2+o(|\lambda_1,\lambda_2|) = 0, \mu_1 < 0\}. \end{align*}$

The Bogdanov-Takens bifurcation diagram and the phase portraits of system $(3.1)$ are shown in . The small neighborhood of the origin in the parameter $(\lambda_1, \lambda_2)$ -plane are divided into four regions (see ) by bifurcation curves $H$ , $HL$ , and $SN$ .

Figure 1. The bifurcation diagram and the phase portraits of system (3.1) for

$m = 2$ ,

$p = 2$ ,

$n = 1.4$ ,

$\beta = 0.1$ and

$A = 4.5607017$ . (a) Bifurcation diagram; (b) No positive equilibria for

$(\lambda_1, \lambda_2) = (0.1, 0.13)\in I$ ; (c) An unstable focus for

$(\lambda_1, \lambda_2) = (0.1, 0.1145)\in II$ ; (d) An unstable limit cycle for

$(\lambda_1, \lambda_2) = (0.1, 0.11345)\in III$ ; (e) An unstable homoclinic loop for

$(\lambda_1, \lambda_2) = (0.1, 0.11289)$ on

$HL$ ; (f) A stable focus for

$(\lambda_1, \lambda_2) = (0.1, 0.096)\in IV$ .

$E'_1$ and

$E'_2$ are the equilibria of system (3.1), near

$E_1$ and

$E_2$ , respectively.

DownLoad: Full-Size Img PowerPoint

(a) The unique positive equilibrium is a cusp of codimension two if $(\lambda_1, \lambda_2) = (0, 0)$ .

(b) There are no equilibria if $(\lambda_1, \lambda_2)\in I$ (see Figure 1(b)), implying the diseases die out.

(d) If the parameters $\lambda_1, \lambda_2$ cross $SN$ into the region II, the saddle-node bifurcation occurs, and there are two positive equilibria $E_1$ and $E_2$ which are saddle and unstable focus respectively (see Figure 1(c)).

(e) If the parameters $\lambda_1, \lambda_2$ cross $H$ into III, an unstable limit cycle will appear through the subcritical Hopf bifurcation around $E_2$ (see ). The focus $E_2$ is stable in region III, whereas it is an unstable weak focus of order one on $H$ .

(f) If $(\lambda_1, \lambda_2)\in HL$ , an unstable homoclinic orbit will occur through the homoclinic bifurcation around $E_1$ (see Figure 1(e)).

(g) If the parameters $\lambda_1, \lambda_2$ cross the $HL$ curve into IV, the relative location of one stable and one unstable manifold of the saddle $E_1$ will be reversed (compare Figure 1(c), (f)).

3.3. Hopf bifurcation

We study Hopf bifurcation around the equilibrium $E_2(x_2, y_2)$ in this subsection. Let

$\begin{align*} a_{*} = \dfrac{mp-m\beta-m-1}{\beta+1}. \end{align*}$

To simplify the computation, we follow our previous techniques in ^[20,21] and take the changes

$\begin{equation} \begin{aligned} \bar{x} = \frac{x}{x_2},\,\, \bar{y} = \frac{y}{y_2},\,\, \tau = x_2^p t, \end{aligned} \end{equation}$

(3.9)

under which system (2.2) is reduced to (we rewrite $\tau$ as $t$ )

$\begin{equation} \left\{ \begin{aligned} &\frac{d\bar{x}}{dt} = \dfrac{\bar{x}^p}{1+ \beta x_2^p \bar{x}^p} \left(\dfrac{A}{x_2}- \bar{x}- n \bar{y}\right)- \dfrac{m}{x_2^p} x, \\ &\frac{d\bar{y}}{dt} = \dfrac{1}{x_2^p}\left(\bar{ x}- \bar{y}\right). \end{aligned} \right. \end{equation}$

(3.10)

Setting

$\begin{equation} \begin{aligned} \bar{A} = \dfrac{A}{x_2},\,\,\bar{\beta} = \beta x_2^p, \,\, \bar{m} = \dfrac{m}{x_2^p},\,\, a = \dfrac{1}{x_2^p}, \end{aligned} \end{equation}$

(3.11)

and dropping the bars, system (3.10) becomes

$\begin{equation} \left\{ \begin{aligned} &\frac{dx}{dt} = \dfrac{x^p}{1+ \beta x^p} \left(A- x- n y\right)- m x, \\ &\frac{dy}{dt} = a\left(x- y\right). \end{aligned} \right. \end{equation}$

(3.12)

Since the equilibrium $E_2(x_2, y_2)$ of system (2.2) becomes the equilibrium $(1, 1)$ of system (3.12), we have

$\begin{equation} A = m \beta+ m+ n+1. \end{equation}$

(3.13)

Note that the positive equilibrium $E(x, y)$ of system (3.12) satisfies that $y = x$ and $h(x) = 0$ , where $h(x)$ is given by (2.6). From the discussion of $h(x)$ in Section $2$ , we have $h'(1) > 0$ since $1$ is the bigger positive root of $h(x)$ . By (2.7) and (3.13), one gets

$\begin{equation*} h'(1) = m \beta+ m+ n+1-pm. \end{equation*}$

Thus, we have the following condition

$\begin{equation} m \beta+ m+ n+1 > pm. \end{equation}$

(3.14)

This yields that the condition (2.4) becomes

$\begin{equation} \begin{aligned} \beta, A, m, n, a > 0,\,\,(p-\beta-1) m-1 < n < \dfrac{m}{a}. \end{aligned} \end{equation}$

(3.15)

Next letting $dt = (1+ \beta x^p)d\tau$ and substituting (3.13) into (3.12), one obtains (still denote $\tau$ by $t$ )

$\begin{equation} \left\{ \begin{aligned} &\frac{dx}{dt} = x^p \left(m \beta+ m+ n+1- x- n y\right)- m (1+ \beta x^p)x, \\ &\frac{dy}{dt} = a\left(x- y\right)(1+ \beta x^p), \end{aligned} \right. \end{equation}$

(3.16)

where $\beta, A, m, n, a$ satisfy (3.15). Since $1+ \beta x^p > 0$ holds in $\mathbb{R}^+_2 = \{(x, y) | x\geq 0, \ y\geq0\}$ , the topological structure of (3.16) is the same as (3.12).

In what follows we study the Hopf bifurcation around $\tilde{E}_2(1, 1)$ in system (3.16), instead of $E_2(x_2, y_2)$ in system (2.2). We always assume that the conditions in (3.15) hold for (3.16).

Theorem 3.2. System $(3.16)$ has an equilibrium at $\tilde{E}_2(1, 1)$ . Moreover, it is

(I) an unstable node or focus if $a < a_{*}$ ;

(II) a stable node or focus if $a > a_{*}$ ;

(III) a weak focus or a center if $a = a_{*}$ .

Proof. The Jacobian matrix of system (3.16) at $\tilde{E}_2(1, 1)$ is

$\begin{align*} J(\tilde{E}_2) = \left(\begin{array}{cc} m p-m\beta-m-1 & -n\\ a (1+\beta) & -a (1+\beta) \end{array}\right). \end{align*}$

Then

$\begin{align*} \text{Det}(J(\tilde{E}_2)) = a(1+\beta)(m \beta+ m+ n+1-pm) = a(1+\beta)h'(1). \end{align*}$

By(3.14), $\text{Det}(J(\tilde{E}_2)) > 0$ . The trace of $J(\tilde{E}_2)$ is

$\begin{align*} \text{Tr}(J(\tilde{E}_2)) = mp-m\beta-m-1-a(\beta+1). \end{align*}$

By conditions (3.15), we have that $\text{Tr}(J(\tilde{E}_2)) = 0$ ( $> 0$ or $< 0$ ) if $a = a_{*}$ ( $a < a_{*}$ or $a > a_{*}$ ). This completes the proof.

Next we study Hopf bifurcation around $\tilde{E}_2(1, 1)$ in system (3.16). By Theorem 3.2, if Hopf bifurcation occurs, then

$\begin{equation} \begin{aligned} a = a_{*},\,\,\beta > 0,\,\,m > 0,\,\,1+\beta+\dfrac{1}{m} < p < 1+\beta+\dfrac{1+n}{m},\,\, \,\,0 < n < \dfrac{m}{a_{*}}. \end{aligned} \end{equation}$

(3.17)

Firstly the following inequality shows the transversality condition holds:

$\begin{align*} \dfrac{d}{da} \text{tr}(J(\tilde{E}_2))\big|_{a = a_{*}} = -(\beta+1) < 0. \end{align*}$

We are going to calculate the first Lyapunov constant for $\tilde{E}_2(1, 1)$ . Take the changes $X = x-1$ , $Y = x-1$ , and let $a = a_{*}$ . System (3.16) becomes (we rewrite $X, Y$ as $x, y$ , respectively)

$\begin{equation} \left\{ \begin{aligned} \dot{x} = &a_{10}x+a_{01}y+a_{20}x^2+a_{11}xy+a_{30}x^3+a_{21}x^2y +a_{40}x^4+a_{31}x^3y+a_{50}x^5+a_{41}x^4y\\ &+O(|x, y|^6),\\ \dot{y} = &b_{10}x+b_{01}y+b_{20}x^2+b_{11}xy+b_{30}x^3+b_{21}x^2y +b_{40}x^4+b_{31}x^3y+b_{50}x^5+b_{41}x^4y\\ &+O(|x, y|^6), \end{aligned} \right. \end{equation}$

(3.18)

where

$\begin{align*} &a_{10} = m p-m\beta-m-1,\ a_{01} = -n,\ a_{20} = \dfrac{p (m p-2 m\beta-m-2)}{2},\ a_{11} = -p n,\\ &a_{30} = \frac{p (p-1) (m p-3 m\beta-2 m-3)}{6}, \, a_{21} = -\dfrac{p (p-1) n}{2},\\ &a_{40} = \dfrac{p (p-1) (p-2) (m p-4 m\beta-3 m-4)}{24}, \,a_{31} = -\dfrac{p (p-1) (p-2) n}{6}, \\ &a_{50} = \frac{p(p-1)(p-2)(p-3) (mp-5m\beta-4m-5)}{120}, \,a_{41} = -\dfrac{p(p-1)(p-2)(p-3)n}{24},\\ &b_{10} = a_{10},\ b_{01} = -a_{10},\, b_{20} = \frac{p\beta a_{10} }{1+\beta},\, b_{11} = -b_{20},\, b_{30} = \frac{1}{2}(p-1)b_{20},\, b_{21} = -b_{30},\\ &b_{40} = \frac{1}{6}(p-1)(p-2)b_{20},\, b_{31} = -b_{40},\,b_{50} = \frac{1}{24}(p-1)(p-2)(p-3)b_{20}, b_{41} = -b_{50}. \end{align*}$

By the formula in ^[14], one gets the first Lyapunov coefficient with the aid of Maple-17 as follows

$\begin{equation} \begin{aligned} V_{3} = \dfrac{pn^2(mp\beta+1)(c_0+c_1 m)} {8(m\beta+m+n+1-pm)(1+\beta)^2}, \end{aligned} \end{equation}$

(3.19)

where

$\begin{align*} c_0 = (n+1)[(\beta-1)p+\beta+1],\,\,\,\,c_1 = 2p^2+(\beta+1)(\beta-3)p+(\beta+1)^2. \end{align*}$

The program for the computation of Lyapunov coefficient is available for noncommercial purpose via email to: gnsydyf@126.com.

By conditions in (3.17), the sign of $V_3$ is the same as that of

$\begin{equation} \varphi_1 = c_0+c_1 m. \end{equation}$

(3.20)

Now we investigate whether there exist some parameters such that $\varphi_1 = 0$ (i.e., $V_3 = 0$ ) under the conditions (3.17).

Lemma 3.3. If conditions in $(3.17)$ hold, then we have $c_1 > 0$ .

Proof. We will show that $c_1(p) > 0$ for all $p > 1$ . A straightforward calculation shows that

$\begin{align*} c_1'(p) = 4p+(\beta+1)(\beta-3),\,\,c_1''(p) = 4. \end{align*}$

Then $c_1'(p) > c_1'(1) = (\beta-1)^2\geq 0$ for all $p > 1$ . This implies that $c_1(p)$ is strictly monotonically increasing on $(1, +\infty)$ . Note that

$\begin{align*} c_1(1) = 2+(\beta+1)(\beta-3)+(\beta+1)^2 = 2\beta^2 > 0. \end{align*}$

Thus, $c_1(p) > c_1(1) > 0$ for all $p > 1$ . This completes the proof.

Denote

$\begin{align*} \tilde{m} = -\frac{c_0}{c_1},\,\,p_{*} = \frac{1+\beta}{1-\beta}. \end{align*}$

Note that $c_0 > 0$ if and only if either (i) $\beta\geq 1$ , or (ii) $\beta < 1$ and $p\leq p_{*}$ . Furthermore, we have $c_0 = 0$ if and only if $\beta < 1$ and $p = p_{*}$ , $c_0 < 0$ if and only if $\beta < 1$ and $p > p_{*}$ , respectively. By the above discussions, one obtains the following theorem.

Theorem 3.4. If conditions in $(3.17)$ hold, then the following statements hold.

(I) If either (i) $\beta\geq 1$ , or (ii) $\beta < 1$ and $p\leq p_{*}$ (i.e., $\varphi_1 > 0$ or $V_3 > 0$ ), then $\tilde{E}_2(1, 1)$ is an unstable weak focus of order one.

(II) If $\beta < 1$ and $p > p_{*}$ and

(II.1) $m > \tilde{m}$ (i.e., $\varphi_1 > 0$ or $V_3 > 0$ ), then $\tilde{E}_2(1, 1)$ is an unstable weak focus of order one;

(II.2) $m < \tilde{m}$ (i.e., $\varphi_1 < 0$ or $V_3 < 0$ ), then $\tilde{E}_2(1, 1)$ is a stable weak focus of order one;

(II.3) $m = \tilde{m}$ (i.e., $\varphi_1 = 0$ or $V_3 = 0$ ), then $\tilde{E}_2(1, 1)$ is a center or a weak focus of order at least two.

Define two hypersurfaces

$\begin{equation*} H_1:\,\, a = a_{*},\,\, \varphi_1 > 0,\,\,\,\,\,\mathrm{and}\,\,\,\,H_2:\,\, a = a_{*},\,\, \varphi_1 < 0. \end{equation*}$

By Theorems 3.2 and 3.4, we know that $\tilde{E}_2(1, 1)$ is an unstable focus (resp. a stable focus) when $\varphi_1 > 0$ and $a\leq a_{*}$ (resp. $\varphi_1 < 0$ and $a\geq a_{*}$ ), while $\tilde{E}_2(1, 1)$ is a stable focus (resp. an unstable focus) as $\varphi_1 > 0$ and $a > a_{*}$ (resp. $\varphi_1 < 0$ and $a < a_{*}$ ). Hence, if parameters pass from one side of the surface $H_1$ (resp. $H_2$ ) to the other side, system (3.16) can undergo a subcritical (resp.supercritical) Hopf bifurcation. An unstable limit cycle (resp. a stable limit cycle) can bifurcate from the small neighborhood of $\tilde{E}_2(1, 1)$ . The hypersurface $H_1$ (resp. $H_2$ ) is called the subcritical (resp. supercritical) Hopf bifurcation hypersurface of system.

In , we give the limit cycles arising from Hopf bifurcation around $\tilde{E}_2(1, 1)$ of system (3.16). In , we fix $p = 2$ , $\beta = 0.2$ , $m = 1.4$ and $n = 10$ , and get $a = 0.1$ from $\text{tr}(J(\tilde{E}_2)) = 0$ , then we obtain $V_3 = -1.62280702$ . Next perturb $a$ such that $a$ decreases to $0.099$ . So $\tilde{E}_2(1, 1)$ becomes an unstable hyperbolic focus, yielding to a stable limit cycle to appear around $\tilde{E}_2(1, 1)$ . In , we fix $p = 2$ , $\beta = 0.2$ , $m = 2$ and $n = 3$ . Using the same arguments as above, we obtain an unstable limit cycle around $\tilde{E}_2(1, 1)$ .

Figure 2. (a) A stable limit cycle bifurcated by the supercritical Hopf bifurcation of the system (3.16) with

$p = 2$ ,

$\beta = 0.2$ ,

$m = 1.4$ ,

$n = 10$ and

$a = 0.099$ . (b) An unstable limit cycle bifurcated by the subcritical Hopf bifurcation of the system (3.16) with

$p = 2$ ,

$\beta = 0.2$ ,

$m = 2$ ,

$n = 3$ and

$a = 0.53$ .

DownLoad: Full-Size Img PowerPoint

3.4. Degenerate Hopf bifurcation of codimension two

By (II.3) of Theorem 3.4, $V_3 = 0$ if $m = m_{*}$ . Using the formal series method in ^[14] and Maple-17, one gets the second Lyapunov constant

$\begin{equation} \begin{aligned} V_{5} = -\dfrac{pn^4(p+1)(2p-1)(p-\beta-1)\varphi_2} {288(1+\beta)^2c_1}, \end{aligned} \end{equation}$

(3.21)

where $c_1$ is given by (3.19) and

$\begin{align*} \varphi_2 = p\beta ( p\beta-p+\beta+1 ) n+ ( 1+\beta) ( p-1 ) ( p\beta-2p+\beta+1). \end{align*}$

Lemma 3.5. If $(3.17)$ and the condition (II.3) of Theorem 3.4 hold, then $V_5 > 0$ .

Proof. From the condition (II.3) of Theorem 3.4, we have $\beta < 1$ and $p\beta-p+\beta+1 > 0$ . This follows that $p \beta-2p+\beta+1 < 0$ . Thus, $\varphi_2 < 0$ , which implies $V_5 > 0$ .

Theorem 3.6. The equilibrium $E_2(x_2, y_2)$ of system $(2.2)$ is an unstable weak focus of order at most two. System $(2.2)$ can undergo degenerate Hopf bifurcation of codimension two in the small neibourhoof of $E_2(x_2, y_2)$ .

Proof. Since the equilibrium $\tilde{E}_2(1, 1)$ in system (3.16) corresponds to $E_2(x_2, y_2)$ in system (2.2), we are going to focus on $\tilde{E}_2(1, 1)$ of system (3.16) in this proof, instead of $E_2$ in system (2.2).

For convenience, denote $V_{1} = \text{tr}(J(\tilde{E}_2))$ . By Theorem 3.4 and Lemma 3.5, for any given parameters $(p_1, \beta_1, n_1, a_1, m_1)$ satisfying (3.17) and the condition (III.3) of Theorem 3.4, i.e.,

$\begin{equation*} a_1 = a_{*}, m_1 = \tilde{m}, \beta_1 > 0,1+\beta_1+\dfrac{1}{\tilde{m}} < p_1 < 1+\beta_1+\dfrac{1+n_1}{\tilde{m}},\,\, \,\,0 < n_1 < \dfrac{\tilde{m}}{a_{*}}, \end{equation*}$

we have $V_{1}(p_1, \beta_1, a_1, m_1) = V_{3}(p_1, \beta_1, n_1, m_1) = 0$ and $V_{5}(p_1, \beta_1, n_1) > 0$ . Therefore, $\tilde{E}_2(1, 1)$ in system (3.16) is a weak focus of order at most $2$ .

Now we show that two limit cycles can bifurcate from $\tilde{E}_2(1, 1)$ , which means that the Hopf cyclicity for the equilibrium $\tilde{E}_2(1, 1)$ is $2$ . We first perturb $m$ near $m_1$ such that $V_3V_5 < 0$ and adjust $a$ such that $V_1 = 0$ . Then the first limit cycle appears. The second limit cycle is obtained by perturbing $a$ such that $V_1V_3 < 0$ , see . Therefore, system (3.16) can undergo degenerate Hopf bifurcation of codimension two near $\tilde{E}_2(1, 1)$ .

Figure 3. Coexistence of two limit cycles around the focus

$\tilde{E}_2(1, 1)$ bifurcated by the degenerate Hopf bifurcation of codimension two of system (3.16) with

$p = 2.4$ ,

$\beta = 0.2$ ,

$n = 5.8$ ,

$m = 0.995$ and

$a = 1/6-0.002$ .

DownLoad: Full-Size Img PowerPoint

In the end of this section we give a numerical simulation in to show the coexistence of two limit cycles. Choosing $p = 2.4$ , $\beta = 0.2$ and $m = 1$ , we deduce that $n = 5.8$ , $a = 1/6$ by $V_3 = 0$ and $tr(J(\tilde{E}_2)) = 0$ , respectively, and $V_5 = 115.8683452$ , i.e., $\tilde{E}_2(1, 1)$ is an unstable weak focus of order two. Then we perturb $m$ and $a$ such that $m$ and $a$ decreases to $1-0.005$ and $1/6-0.002$ , respectively. An unstable limit cycle and a stable limit cycle occur around $\tilde{E}_2(1, 1)$ , see Figure 3.

Acknowledgments

This work is supported by the NNSF of China (No.11971495) and Guangdong-Hong Kong-Macau Applied Math Center (No.2020B1515310014).

Conflict of interest

The authors declare there is no conflict of interest.

References

[1]	S. Misono, S. Marmor, N. Roy, T. Mau, S. Cohen, Multi-institutional study of voice disorders and voice therapy referral, Otolaryngol. Head Neck Surgery, 155 (2016), 33-41. doi: 10.1177/0194599816639244
[2]	P. Bradley, Voice disorders: Classification, Otolaryngol. Head Neck Surgery, (2010), 555-562.
[3]	M. Behlau, M. L. S. Dragone, L. Nagano, The voice that teaches: The teacher and oral communication in the classroom, 2004.
[4]	A. E. Aronson, Clinical voice disorders, 3 ed., INC. New York: Thieme Medical Publishers, 1990, p. 3-11.
[5]	J. R. Spiegel, R. T. Sataloff, K. A. Emerich, The young adult voice, J. Voice, 11 (1997), 138-143. doi: 10.1016/S0892-1997(97)80069-0
[6]	L. O. Ramig, K. Verdolini, Treatment efficacy: Voice disorders, J. Speech Lang. Hear. Res., 41 (1998), 101-106.
[7]	J. Baker, The role of psychogenic and psychosocial factors in the development of functional voice disorders, J. Speech Lang. Pathol., 10 (2008), 210-230. doi: 10.1080/17549500701879661
[8]	S. T. Kasama, A. G. Brasolotto, Vocal perception and life quality, Pro. Fono., 9 (2007), 19-28.
[9]	L. P. Ferreira, J. G. Santos, M. F. B. Lima, Vocal sympton and its probable cause: Data colleting in a population, Rev. CEFAC, 11 (2009), 110-118. doi: 10.1590/S1516-18462009000100015
[10]	P. H. Dejonckere, P. Bradley, P. Clemente, G. Cornut G, L. C. Buchman, G. Friedrich, et al., A basic protocol for functional assessment of voice pathology, especially for investigating the efficacy of (phonosurgical) treatments and evaluating new assessment techniques, Eur. Arch. Otorhinolaryngol., 258 (2001), 77-82. doi: 10.1007/s004050000299
[11]	U. Cesari, G. De Pietro, E. Marciano, C. Niri, G. Sannino, L. Verde, Voice disorder detection via an m-Health system: Design and results of a clinical study to evaluate Vox4Health, BioMed. Res. Int., 2018 (2018), 1-19.
[12]	L. Verde, G. De Pietro, G. Sannino, Voice disorder identification by using machine learning techniques, IEEE Access, 6 (2018), 16246-16255. doi: 10.1109/ACCESS.2018.2816338
[13]	A. G. David, J. B. Magnus, Diagnosing parkinson by using artificial neural networks and support vector machines, Global J. Comput. Sci. Technol., (2009), 63-71.
[14]	Saarbruecken Voice Database—Handbook, Stimmdatenbank.coli.uni-saarland.de. [Online]. Available: http://www.stimmdatenbank.coli.uni-saarland.de/help_en.php4.
[15]	M. OpenCourseWare, Lab Database \| Laboratory on the Physiology, Acoustics, and Perception of Speech \| Electrical Engineering and Computer Science \| MIT OpenCourseWare, Ocw.mit.edu. [Online]. Available: https://ocw.mit.edu/courses/electrical-engineering-and-computer-science/6-542j-laboratory-on-the-physiology-acoustics-and-perception-of-speech-fall-2005/lab-database/
[16]	K. Daoudi, B. Bertrac, On classification between normal and pathological voices using the MEEI-KayPENTAX database: Issues and consequences, INTERSPEECH-2014, Sep 2014, Singapour, Singapore. ffhal-01010857
[17]	N. Sáenz-Lechón, J. I. Godino-Llorente, V. Osma-Ruiz, P. Gómez-Vilda, Methodological issues in the development of automatic systems for voice pathology detection, Biomed. Signal Process. Control, 1 (2006), 120-128.
[18]	A. Liberati, D. G. Altman, J. Tetzlaff, C. Mulrow, P. C. Gøtzsche, J. P. A. Ioannidis, et al., The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: Explanation and elaboration, BMJ, 339 (2009).
[19]	A. Al-Nasheri, G. Muhammad, M. Alsulaiman, Z. Ali, K. H. Malki, T. A. Mesallam, et al., Voice pathology detection and classification using auto-correlation and entropy features in different frequency regions, IEEE Access, 6, 6961-6974.
[20]	A. Al-Nasheri, G. Muhammad, M. Alsulaiman, Z. Ali, T. A. Mesallam, M. Farahat, et al., An investigation of multidimensional voice program parameters in three different databases for voice pathology detection and classification, J. Voice, 31 (2017), 113.e9-e18.
[21]	A. Al-Nasheri, G. Muhammad, M. Alsulaiman, Z. Ali, Investigation of voice pathology detection and classification on different frequency regions using correlation functions, J. Voice, 31 (2017), 3-15. doi: 10.1016/j.jvoice.2016.01.014
[22]	Z. Ali, M. Alsulaiman, G. Muhammad, I. Elamvazuthi, A. Al-Nasheri, T. A. Mesallam, K. H. Malki, et al., Intra- and inter-database study for Arabic, English, and German databases: Do conventional speech features detect voice pathology?, J. Voice, 31 (2017), 386.e1-e8.
[23]	E. S. Fonseca, R. C. Guido, S. B. Junior, H. Dezani, R. R. Gati, D. C. Mosconi Pereira, Acoustic investigation of speech pathologies based on the discriminative paraconsistent machine (DPM), Biomed. Signal Process. Control, 55 (2020).
[24]	J. A. Gómez-García, L. Moro-Velázquez, J. Mendes-Laureano, G. Castellanos-Dominguez, J. I. Godino-Llorente, Emulating the perceptual capabilities of a human evaluator to map the GRB scale for the assessment of voice disorders, Eng. Appl. Artific. Intell., 82 (2019), 236--251. doi: 10.1016/j.engappai.2019.03.027
[25]	V. Guedes, F. Teixeira, A. Oliveira, J. Fernandes, L. Silva, A. Junior, et al., Transfer Learning with AudioSet to Voice Pathologies Identification in Continuous Speech, Proced. Comput. Sci., 164 (2019), 662-669. doi: 10.1016/j.procs.2019.12.233
[26]	I. Hammami, L. Salhi, S. Labidi, Voice pathologies classification and detection using EMD- DWT analysis based on higher order statistic features, IRBM, 41 (2020), 161-171. doi: 10.1016/j.irbm.2019.11.004
[27]	D. Hemmerling, A. Skalski, J. Gajda, Voice data mining for laryngeal pathology assessment, Comput. Biol. Med., 69 (2016), 270-276. doi: 10.1016/j.compbiomed.2015.07.026
[28]	J. Moon, S. Kim, An approach on a combination of higher-order statistics and higher-order differential energy operator for detecting pathological voice with machine learning, 2018 International Conference on Information and Communication Technology Convergence (ICTC), 17-19 Oct. 2018, pp. 46-51.
[29]	K. Ezzine, M. Frikha, Investigation of glottal flow parameters for voice pathology detection on SVD and MEEI databases, 2018 4th International Conference on Advanced Technologies for Signal and Image Processing (ATSIP), 21-24 March 2018, pp. 1-6.
[30]	M. Markaki, Y. Stylianou, Using modulation spectra for voice pathology detection and classification, 2009 Annual International Conference of the IEEE Engineering in Medicine and Biology Society, 3-6 Sept. 2009, pp. 2514-2517.
[31]	M. Markaki, Y. Stylianou, Voice pathology detection and discrimination based on modulation spectral features, IEEE Transact. Aud. Speech Langu. Process, 19 (2011), 1938-1948. doi: 10.1109/TASL.2010.2104141
[32]	J. M. Miramont, J. F. Restrepo, J. Codino, C. Jackson-Menaldi, G. Schlotthauer, Voice signal typing using a pattern recognition approach, J. Voice, 2020.
[33]	G. Muhammad, M. Alsulaiman, Z. Ali, T. A. Mesallam, M. Farahat, K. H. Malki, et al., Voice pathology detection using interlaced derivative pattern on glottal source excitation, Biomed. Signal Process. Control, 31 (2017), 156-164. doi: 10.1016/j.bspc.2016.08.002
[34]	S. E. Shia, T. Jayasree, Detection of pathological voices using discrete wavelet transform and artificial neural networks, 2017 IEEE International Conference on Intelligent Techniques in Control, Optimization and Signal Processing (INCOS), 23-25 March 2017, pp. 1-6.
[35]	S. R. Kadiri, P. Alku, Analysis and detection of pathological voice using glottal source features, IEEE J. Select. Topics Signal Process., 14 (2020), 367-379. doi: 10.1109/JSTSP.2019.2957988
[36]	T. Zhang, Y. Shao, Y. Wu, Z. Pang, G. Liu, Multiple vowels repair based on pitch extraction and line spectrum pair feature for voice disorder, IEEE J. Biomed. Health Inform., 24 (2020), 1940-1951. doi: 10.1109/JBHI.2020.2978103
[37]	F. Teixeira, J. Fernandes, V. Guedes, A. Junior, J. P. Teixeira, Classification of control/pathologic subjects with support vector machines, Proced. Comput. Sci., 138 (2018), 272-279. doi: 10.1016/j.procs.2018.10.039
[38]	J. P. Teixeira, P. O. Fernandes, N. Alves, Vocal acoustic analysis—classification of dysphonic voices with artificial neural networks, Proced. Comput. Sci., 121 (2017), 19-26. doi: 10.1016/j.procs.2017.11.004
[39]	G. Muhammad, M. Melhem, Pathological voice detection and binary classification using MPEG-7 audio features, Biomed. Signal Process. Control, 11 (2014), 1-9. doi: 10.1016/j.bspc.2014.02.001
[40]	J. Nayak, P. S. Bhat, R. Acharya, U. V. Aithal, Classification and analysis of speech abnormalities, ITBM-RBM, 26 (2005), 319-327. doi: 10.1016/j.rbmret.2005.05.002
[41]	Z. Ali, I. Elamvazuthi, M. Alsulaiman, G. Muhammad, Automatic voice pathology detection with running speech by using estimation of auditory spectrum and cepstral coefficients based on the all-pole model, J. Voice, 30 (2016), 757.e7-e19.
[42]	R. Amami, A. Smiti, An incremental method combining density clustering and support vector machines for voice pathology detection, Comput. Electr. Eng., 57 (2017), 257-265. doi: 10.1016/j.compeleceng.2016.08.021
[43]	J. D. Arias-Londoño, J. I. Godino-Llorente, N. Sáenz-Lechón, V. Osma-Ruiz, G. Castellanos- Domínguez, An improved method for voice pathology detection by means of a HMM-based feature space transformation, Patt. Recogn., 43 (2010), 3100-3112.
[44]	M. K. Arjmandi, M. Pooyan, M. Mikaili, M. Vali, A. Moqarehzadeh, Identification of voice disorders using long-time features and support vector machine with different feature reduction methods, J. Voice, 25 (2011), e275-e289. doi: 10.1016/j.jvoice.2010.08.003
[45]	R. R. A. Barreira, L. L. Ling, Kullback-leibler divergence and sample skewness for pathological voice quality assessment, Biomed. Signal Process. Control, 57 (2020), 101697. doi: 10.1016/j.bspc.2019.101697
[46]	C. R. Francis, V. V. Nair, S. Radhika, A scale invariant technique for detection of voice disorders using Modified Mellin Transform, 2016 International Conference on Emerging Technological Trends (ICETT), 21-22 Oct. 2016, pp. 1-6.
[47]	H. Cordeiro, J. Fonseca, I. Guimarães, C. Meneses, Hierarchical classification and system combination for automatically identifying physiological and neuromuscular laryngeal pathologies, J. Voice, 31 (2017), 384.
[48]	H. T. Cordeiro, C. M. Ribeiro, Spectral envelope first peak and periodic component in pathological voices: A spectral analysis, Proced. Comput. Sci., 138 (2018), 64-71. doi: 10.1016/j.procs.2018.10.010
[49]	S. H. Fang, Y. Tsao, M. J. Hsiao, J. Y. Chen, Y. H. Lai, F. C. Lin, et al., Detection of pathological voice using cepstrum vectors: A deep learning approach, J. Voice, 33 (2019), 634-641. doi: 10.1016/j.jvoice.2018.02.003
[50]	G. Muhammad, Voice pathology detection using vocal tract area, 2013 European Modelling Symposium, 20-22 Nov. 2013, pp. 164-168.
[51]	H. Ghasemzadeh, M. Tajik Khass, M. Khalil Arjmandi, M. Pooyan, Detection of vocal disorders based on phase space parameters and Lyapunov spectrum, Biomed. Signal Process. Control, 22 (2015), 135-145. doi: 10.1016/j.bspc.2015.07.002
[52]	J. I. Godino-Llorente, R. Fraile, N. Sáenz-Lechón, V. Osma-Ruiz, P. Gómez-Vilda, Automatic detection of voice impairments from text-dependent running speech, Biomed. Signal Process. Control, 4 (2009), 176-182.
[53]	M. Hariharan, K. Polat, R. Sindhu, S. Yaacob, A hybrid expert system approach for telemonitoring of vocal fold pathology, Appl. Soft Comput., 13 (2013), 4148-4161. doi: 10.1016/j.asoc.2013.06.004
[54]	A. Mahmood, A solution to the security authentication problem in smart houses based on speech, Proced. Comput. Sci., 155 (2019), 606-611. doi: 10.1016/j.procs.2019.08.085
[55]	J. Mekyska, E. Janousova, P. Gomez-Vilda, Z. Smekal, I. Rektorova, I. Eliasova, et al., Robust and complex approach of pathological speech signal analysis, Neurocomputing, 167 (2015), 94-111. doi: 10.1016/j.neucom.2015.02.085
[56]	G. Muhammad, M. Melhem, Pathological voice detection and binary classification using MPEG-7 audio features, Biomed. Signal Process. Control, 11 (2014), 1-9. doi: 10.1016/j.bspc.2014.02.001
[57]	J. Nayak, P. S. Bhat, R. Acharya, U. V. Aithal, Classification and analysis of speech abnormalities, ITBM-RBM, 26 (2005), 319-327. doi: 10.1016/j.rbmret.2005.05.002
[58]	P. Henriquez, J. B. Alonso, M. A. Ferrer, C. M. Travieso, J. I. Godino-Llorente, F. Diaz-de- Maria, Characterization of healthy and pathological voice through measures based on nonlinear dynamics, IEEE Transact. Audio Speech Lang. Process., 17 (2009), 1186-1195.
[59]	P. Salehi, Using patient's speech signal for vocal ford disorders detection based on lifting scheme, in 2015 2nd International Conference on Knowledge-Based Engineering and Innovation (KBEI), 5-6 Nov. 2015, pp. 561-568.
[60]	N. Sáenz-Lechón, J. I. Godino-Llorente, V. Osma-Ruiz, P. Gómez-Vilda, Methodological issues in the development of automatic systems for voice pathology detection, Biomed. Signal Process. Control, 1 (2006), 120-128.
[61]	C. M. Travieso, J. B. Alonso, J. R. Orozco-Arroyave, J. F. Vargas-Bonilla, E. Nöth, A. G. Ravelo- García, Detection of different voice diseases based on the nonlinear characterization of speech signals, Expert Systems Appl., 82 (2017), 184-195.
[62]	T. A. Mesallam, F. Mohamed, K. H. Malki, A. Mansour, A. Zulfiqar, A. N. Ahmed, et al., Development of the arabic voice pathology database and its evaluation by using speech features and machine learning algorithms, J. Healthc. Eng., (2017), 1-13.
[63]	K. Uma Rani, Mallikarjun S Holi, A comparative study of neural networks and support vector machines for neurological disordered voice classification, Inter. J. Eng. Res. Techol., 3 (2014).
[64]	J. Godino-Llorente, P. Gómez-Vilda, N. Sáenz-Lechón, M. Blanco-Velasco, F. Cruz-Roldán, M. Ferrer-Ballester, Support vector machines applied to the detection of voice disorders, Nonlin. Analy. Algor. Speech Process., (2006), 219-230.
[65]	S. Huang, N. Cai, P. P. Pacheco, S. Narrandes, Y. Wang, W. Xu, Applications of support vector machine (svm) learning in cancer genomics, Cancer Genom. Proteom., 15 (2018).
[66]	S. Yue, P. Li, P. Hao, SVM classification: Its contents and challenges, Appl. Math. J. Chinese Univer., 18 (2003), 332-342. doi: 10.1007/s11766-003-0059-5
[67]	D. Reynolds, Gaussian Mixture Models, In: S. Z. Li, A. Jain (eds), Encyclopedia of Biometrics, Springer, Boston, MA, 2009.
[68]	L. Breiman, J. Friedman, C. J. Stone, R. A. Olshen, Classification and regression trees, Boca Raton, FL: CRC press, 1984.
[69]	L. Breiman, Bagging predictors, Mach. Learn., 24 (1996), 123-140.
[70]	S. Indolia, A. Goswami, S. Mishra, P. Asopa, Conceptual understanding of convolutional neural network- A deep learning approach, Proced. Computer Sci., 132 (2018), 679-688. doi: 10.1016/j.procs.2018.05.069
[71]	R. Yamashita, M. Nishio, R. Do, K. Togashi, Convolutional neural networks: An overview and application in radiology, Insights Imag., 9 (2018), 611-629. doi: 10.1007/s13244-018-0639-9
[72]	V. Parsa, D. G. Jamieson, Identification of pathological voices using glottal noise measures, J. Speech Langu. Hear. Res., 43 (2000), 469-485. doi: 10.1044/jslhr.4302.469
[73]	D. D. Deliyski, H. S. Shaw, M. K. Evans, Influence of sampling rate on accuracy and reliability of acoustic voice analysis, Logoped. Phoniatr. Vocol., 30 (2005), 55-62. doi: 10.1080/1401543051006721
[74]	Y. Horii, Jitter and shimmer in sustained vocal fry phonation, Folia Phoniatr., 37 (1985), 81-86. doi: 10.1159/000265785
[75]	J. L. Fitch, Consistency of fundamental frequency and perturbation in repeated phonations of sustained vowels, reading, and connected speech, J. Speech Hear. Disord., 55 (1990), 360-363. doi: 10.1044/jshd.5502.360
[76]	T. Mesallam, M. Farahat, K. Malki, M. Alsulaiman, Z. Ali, A. Al-nasheri, et al., Development of the arabic voice pathology database and its evaluation by using speech features and machine learning algorithms, J. Healthc. Eng., 2017, 1-13.
[77]	P. Harar, Z. Galaz, J. Alonso-Hernandez, J. Mekyska, R. Burget, Z. Smekal, Towards robust voice pathology detection, Neural Comput. Appl., 2018.
[78]	D. D. Mehta, R. E. Hillman, Voice assessment: Updates on perceptual, acoustic, aerodynamic, and endoscopic imaging methods, Curr. Opin. Otolaryngol. Head Neck Surg., 16 (2008), 211. doi: 10.1097/MOO.0b013e3282fe96ce

This article has been cited by:

1.	Gaoyang She, Fengqi Yi, Stability and Bifurcation Analysis of a Reaction–Diffusion SIRS Epidemic Model with the General Saturated Incidence Rate, 2024, 34, 0938-8974, 10.1007/s00332-024-10081-z
2.	Wenzhe Cui, Yulin Zhao, Saddle-node bifurcation and Bogdanov-Takens bifurcation of a SIRS epidemic model with nonlinear incidence rate, 2024, 384, 00220396, 252, 10.1016/j.jde.2023.11.030
3.	Yantao Luo, Yunqiang Yuan, Zhidong Teng, ANALYSIS OF A DEGENERATED DIFFUSION SVEQIRV EPIDEMIC MODEL WITH GENERAL INCIDENCE IN A SPACE HETEROGENEOUS ENVIRONMENT, 2024, 14, 2156-907X, 2704, 10.11948/20230346
4.	Burcu Gürbüz, Aytül Gökçe, Segun I. Oke, Michael O. Adeniyi, Mayowa M. Ojo, Dynamical behavior and bifurcation analysis for a theoretical model of dengue fever transmission with incubation period and delayed recovery, 2025, 03784754, 10.1016/j.matcom.2025.03.008

Reader Comments

Your name:*

Email:*
© 2020 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)