Convolutional neural network with group theory and random selection particle swarm optimizer for enhancing cancer image classification

Kun Lan; Gloria Li; Yang Jie; Rui Tang; Liansheng Liu; Simon Fong; Kun Lan; Gloria Li; Yang Jie; Rui Tang; Liansheng Liu; Simon Fong

doi:10.3934/mbe.2021281

Mathematical Biosciences and Engineering

2021, Volume 18, Issue 5: 5573-5591. doi: 10.3934/mbe.2021281

Previous Article Next Article

Research article Special Issues

Convolutional neural network with group theory and random selection particle swarm optimizer for enhancing cancer image classification

1.
Department of Computer and Information Science, Faculty of Science and Technology, University of Macau, Macau 999078, China
2.
DACC Laboratory, Zhuhai Institutes of Advanced Technology of the Chinese Academy of Sciences, Zhuhai 519080, China
3.
Department of Management and Science and Information System, Faculty of Management and Economics, Kunming University of Science and Technology, Kunming 650093, China
4.
Department of Medical Imaging, First Affiliated Hospital of Guangzhou University of Chinese Medicine, Guangzhou 510405, China

Academic Editor: Kelvin K.L. Wong

Received: 19 April 2021 Accepted: 15 June 2021 Published: 21 June 2021

As an epitome of deep learning, convolutional neural network (CNN) has shown its advantages in solving many real-world problems. Successful CNN applications on medical prognosis and diagnosis have been achieved in recent years. Their common goal is to recognize the insights from the subtle details from medical images by building a suitable CNN model with maximum accuracy and minimum error. The CNN performance is extremely sensitive to the parameter tuning for any given network structure. To approach this concern, a novel self-tuning CNN model is proposed with a significant characteristic of having a metaheuristic-based optimizer. The most optimal set of parameters is often found via our proposed method, namely group theory and random selection-based particle swarm optimization (GTRS-PSO). The insights of symmetric essentials of model structure and parameter correlation are extracted, followed by the hierarchical partitioning of parameter space, and four operators on those partitions are designed for moving neighborhoods and formulating the swarm topology accordingly. The parameters are updated by a random selection strategy at each interval of partitions during the search process. Preliminary experiments over two radiology image datasets: breast cancer and lung cancer, are conducted for a comprehensive comparison of GTRS-PSO versus other optimization algorithms. The results show that CNN with GTRS-PSO optimizer can achieve the best performance for cancer image classifications, especially when there are symmetric components inside the data properties and model structures.

Keywords:

Citation: Kun Lan, Gloria Li, Yang Jie, Rui Tang, Liansheng Liu, Simon Fong. Convolutional neural network with group theory and random selection particle swarm optimizer for enhancing cancer image classification[J]. Mathematical Biosciences and Engineering, 2021, 18(5): 5573-5591. doi: 10.3934/mbe.2021281

Related Papers:

[1]	Jasmina Djedjibegovic, Aleksandra Marjanovic, Ilhana Kobilica, Amila Turalic, Aida Lugusic, Miroslav Sober . Lifestyle management of polycystic ovary syndrome: a single-center study in Bosnia and Herzegovina. AIMS Public Health, 2020, 7(3): 504-520. doi: 10.3934/publichealth.2020041
[2]	Carla Gonçalves, Helena Moreira, Ricardo Santos . Systematic review of mediterranean diet interventions in menopausal women. AIMS Public Health, 2024, 11(1): 110-129. doi: 10.3934/publichealth.2024005
[3]	Ehsan Shoohanizad, Ata Garajei, Aida Enamzadeh, Amir Yari . Nonsurgical management of temporomandibular joint autoimmune disorders. AIMS Public Health, 2019, 6(4): 554-567. doi: 10.3934/publichealth.2019.4.554
[4]	Felipe Montalva-Valenzuela, Antonio Castillo-Paredes, Claudio Farias-Valenzuela, Oscar Andrades-Ramirez, Yeny Concha-Cisternas, Eduardo Guzmán-Muñoz . Effects of exercise, physical activity, and sports on physical fitness in adults with Down syndrome: A systematic review. AIMS Public Health, 2024, 11(2): 577-600. doi: 10.3934/publichealth.2024029
[5]	Junji Takaya . Small for Gestational Age and Magnesium: Intrauterine magnesium deficiency may induce metabolic syndrome in later life. AIMS Public Health, 2015, 2(4): 793-803. doi: 10.3934/publichealth.2015.4.793
[6]	Ahmed A Mohsen, Hassan Fadhil AL-Husseiny, Xueyong Zhou, Khalid Hattaf . Global stability of COVID-19 model involving the quarantine strategy and media coverage effects. AIMS Public Health, 2020, 7(3): 587-605. doi: 10.3934/publichealth.2020047
[7]	Nandeeta Samad, Pranta Das, Segufta Dilshad, Hasan Al Banna, Golam Rabbani, Temitayo Eniola Sodunke, Timothy Craig Hardcastle, Ahsanul Haq, Khandaker Anika Afroz, Rahnuma Ahmad, Mainul Haque . Women's empowerment and fertility preferences of married women: analysis of demographic and health survey’2016 in Timor-Leste. AIMS Public Health, 2022, 9(2): 237-261. doi: 10.3934/publichealth.2022017
[8]	Musyoka Kinyili, Justin B Munyakazi, Abdulaziz YA Mukhtar . Mathematical modeling and impact analysis of the use of COVID Alert SA app. AIMS Public Health, 2022, 9(1): 106-128. doi: 10.3934/publichealth.2022009
[9]	Pooja Yadav, Shah Jahan, Kottakkaran Sooppy Nisar . Analysis of fractal-fractional Alzheimer's disease mathematical model in sense of Caputo derivative. AIMS Public Health, 2024, 11(2): 399-419. doi: 10.3934/publichealth.2024020
[10]	Yasir Rehman, Nadia Rehman . Association of climatic factors with COVID-19 in Pakistan. AIMS Public Health, 2020, 7(4): 854-868. doi: 10.3934/publichealth.2020066

Abstract

1. Introduction

One of the most prevalent endocrine system illnesses, polycystic ovarian syndrome (PCOS), affects 5 to 10% of women [1]. The PCOS is one of the most prevalent causes of ovulatory failure [2]. Women ages 18-44 are affected. PCOS develops during adolescence and is brought on by hormonal imbalances. Follicles, or cysts, which are fluid-filled sacs, are found on the ovary's periphery. A polycystic ovary (PCO) is defined as having twelve or more follicles that are between two to nine mm in diameter [3]. Both health and the standard of living for women are impacted by PCOS. Some of the symptoms include weight, worry, depression, and stress [4]. Others include heart issues, ovarian failure and infertility, type 2 diabetes, late menopause, acne, hair loss, baldness, and hirsutism [5]. According to reports, the prevalence ranges from 2.2–26% globally [6]. According to community research conducted in the United Kingdom (UK), South Asians exhibit a prevalence of 52% when compared to Caucasians, who exhibit a prevalence of 22% [3]. Early identification and treatment can be used to control the symptoms and avoid long-term problems. Gynaecologists urge patients on clomiphene citrate to wait no longer than six months after beginning their ovulation cycles before starting gonadotrophin therapy. The ovulation stimulation approach should not depend on whether the clomiphene citrate and gonadotropin therapy cycles are identified and their therapeutic follow-up is taken care of, but whether the pregnancy does not develop over the course of the 912 treatment cycles. Thus, it is encouraged to use assisted reproductive techniques (ART), among which in vitro fertilization (IVF) is one [7]. It has become increasingly obvious over the past ten years that PCOS patients need better clinical and therapeutic care. The primary attempts for better controlling PCOS were the adoption of more precise technologies for detecting circulating androgens, comprehension of the impact of PCOS on risk factors and ultimately, pharmaceutical regimens based on individual-specific phenotypic requirements [8]. Mathematical modeling can be used to suggest controls for any infectious disease after accounting for the disease's mechanism of transmission. Different nonlinear therapeutic rates between diseases are possible. The nonlinear incidence and treatment rate can be very helpful in assisting health organizations identify effective treatments that will stop the spread of the disease [9].

Various operators have been presented in fractional calculus. The Caputo operator, which is defined on a power law kernel, is the most fundamental operator. Many scientists working in engineering [10], mathematical biology [11], [12], fluid mechanics [13], [14], and mathematical physics [15]–[17] by using fractional derivative. A study on the numerical and computational aspects of the physical system of a (1+1)-dimensional Mikhailov-Novikov-Wang (MNW) integrable equation was conducted by Khater et al. [18]–[20]. The analytical and approximate solutions of the caudrrey-dodd-gibbon (CDG) model were examined by Khater et al. [21], while the computational simulations of the propagation of tsunami waves across the ocean were covered in [22]. Particularly, fractional calculus has also been used to study the dynamics of cancer [23]. A fractional component cancer model was put forth by Naik et al. [24]. Atangana [25] put out a novel family of two-parameter derivatives in 2017. The order of the two parameters, where one reflects the exponential decay, power law, or Mittag-Leffler kernel operations, describes the fractal dimension. It is impossible to obtain certain nonlinear model parameters through experimentation. This work suggests fractal-fractional derivatives as a potential solution to these problems. Farman et al. proposed in [26] a Caputo Fabrizio fractional order model for glucose control in insulin therapy for diabetes. Researchers expanded hybrid fractal-fractional operators and discussed a whole new hybrid model of the coronavirus propagation in [27]. They also compared its results with past iterations of the fractal-fractional model. To understand how the Ebola virus spreads, Farman et al. [28] proposed a nonlinear time-fractional mathematical model of the disease. The fractional descement's stripping endothelial keratoplasty (DSEK) model was examined using fractional derivatives of the Atangana-Baleanu Caputo type [29]. The fractional Euler's numerical approach was derived using the Atangana-Baleanu Caputo (ABC) fractional derivative and was then applied to the fractional DSEK model. Researchers used fractal-fractional Atangana-Baleanu derivatives and integrals in the sense of Caputo to study the dynamics of Q (query) fever transmission in cattle and ticks as well as the bacterial load in the environment [30]. The freshly created Newton polynomial was utilized by the stated numerical approach. Researchers researched and observed the dynamical transmission of the illness under the impact of vaccination using a unique fractional order measles model, utilizing a constant proportional (CP) Caputo operator [31]. To express a set of fractional differential equations numerically, they used Laplace with the Adomian decomposition approach. A generalized fractional model was used to describe how HIV/AIDS spread throughout the Cape Verde Islands [32]. The model was successfully built using a two-step Lagrange polynomial interpolation, and the associated error analysis was investigated. In [33], a fractional order tubarculosis (TB) model with a fractal fractional operator was built using an expanded Mittag-Leffler kernel.

This is how the sections are organized: Literature evaluation and an introduction are provided in section one. Basic terms related to fractional derivatives are covered in section two. We spoke about the fractional-order PCOS model with the Mittag-Leffler kernal in section three. In section 3.1, we prove the detailed analysis of positiveness and boundedness, existence and uniqueness, reproductive number, and equilibrium point analysis. In section four, the equilibrium point stability analysis with Jacobian matrix and Lyaounov function is performed. In section five, numerical results are developed.

2. Basic concepts

2.1. Definition 1 [34]

Consider $x \in k (c, d), d > c, η \in [0,1]$ , for the Caputo derivative of arbitrary is given by

$D_{t}^{α} x (t) = \frac{L (α)}{1 - α} \int_{c}^{t} x^{'} (t) e x p [- α \frac{t - ς}{1 - α}] d ς,$ (1)

where L(α) is the normalization of the function that holds L(0) = L(1) = 1.

However, if $j \notin k (c, d)$ , we get

$D_{t}^{α} x (t) = α \frac{L (α)}{1 - α} \int_{c}^{t} (x (t) - x (ς)) e x p [- α \frac{t - ς}{1 - α}] d ς .$ (2)

Remark: If $κ = \frac{1 - α}{α} \in [0,1]$ , then we get the form

$D_{t}^{α} h (t) = \frac{R (κ)}{κ} \int_{c}^{t} x^{'} (t) e x p [- \frac{t - ς}{κ}] d ς,$ (3)

$G (0) = G (\infty) = 1$ . Moreover ${l i m}_{κ \to \infty} \frac{1}{κ} e x p [- \frac{t - ς}{κ}] = λ (ς - t) .$

2.2. Definition 2 [34]

The Caputo fractional derivative is given by c > 0, $q \in H_{1} (0, c)$ and $0 < η < 1$ ; thus

$^{C} D^{α} x (t) = \frac{1}{Γ (1 - α)} \int_{0}^{t} {(t - ς)}^{- α} (x^{'} (ς) d ς . t > 0$ (4)

2.3. Definition 3 [35], [36]

A power law kernel in the Riemann-Liouville concept is defined as;

$_{0}^{F F P} D_{t}^{α, υ} x (t) = \frac{1}{Γ (1 - α)} \frac{d}{d t^{υ}} \int_{0}^{t} {(t - ς)}^{- α} x (ς) d ς,$ (5)

with $0 \leq α, υ \leq 1$ ,

$\frac{d f (t)}{d t^{υ}} = \underset{t \to t_{1}}{l i m} \frac{f (t) - f (t_{1})}{t^{2 - υ} - t_{1}^{2 - υ}} (2 - υ) .$

The fractal-fractional integral corresponding power law kernel of order $(α, υ)$ is given as

$_{0}^{F F P} I_{t}^{α, υ} x (t) = \frac{1}{Γ (α)} \int_{0}^{t} {(t - ς)}^{α - 1} ς^{1 - υ} x (ς) d ς .$ (6)

3. Polycystic ovarian syndrome model with fractional derivative

Considering the classical order model given in [7], in our assumption, put second treatment factor zero for better understanding transmission of disease in society in different age groups. The model is predicated on the following hypotheses: the whole population, or N(t), is composed of five sub-populations: Women who are S(t) susceptible to infertility, I(t) suffering from PCOS, T(t) infertile women receiving treatment with gonadotropin and clomiphene citrate and R(t) women who have recovered from infertility. The following nonlinear fractional differential equations for the fractional order model are created using the fractal-fractional operator in the Caputo sense:

$\begin{array}{l} _{0}^{F F P} D_{t}^{α, υ} S (t) = Π - a S - b e S T \\ _{0}^{F F P} D_{t}^{α, υ} I (t) = b e S T - (a + k) I \\ _{0}^{F F P} D_{t}^{α, υ} T (t) = k c I - (r + a) T \\ _{0}^{F F P} D_{t}^{α, υ} R (t) = u r T - a R, \end{array}$ (7)

with initial state $S (0) = S_{0} \geq 0, I (0) = I_{0} \geq 0, T (0) = T_{0} \geq 0, R (0) = R_{0} \geq 0$ ,

where a represents the frequency at which patients visit the clinic for a disease diagnosis and treatment. Abortion and restarting the treatment cycle occur at a rate of b. The rate of therapy for women who got pregnant with medicine (gonadotropin and clomiphene citrate) is e. We display the treatment rate in the patient class with k as well as the k patient group who were receiving c medical therapy. We use r to display the group T's recovery rate. We represent the number of recoveries as urT, where u is the recovery rate of rT at time t. There were aR susceptible individuals who died and left the group. The number of infertile women who used a therapeutic therapy is represented by $Π = j N$ .

3.1. Analysis of proposed model

3.1.1. Solution boundedness and positivity

To demonstrate the positivity of the solutions since they reflect actual problems in the real world with positive values, in this subsection, we look at the circumstances in which the considered model's solutions satisfy the positivity requirement.

Theorem 3.1 $S (t), I (t), T (t)$ and R(t) are positive solutions to the system equation for all $t \geq 0$ if $S (0) \geq 0, I (0) \geq 0, T (0) \geq 0, R (0) \geq 0$ , respectively. All of the solutions to the system of equations $S (t), I (t), T (t)$ and R(t) are finite.

Let's begin with the S(t) group:

$\begin{array}{l} S (t) = Π - a S - b e S T, \forall t \geq 0 \\ S (t) \geq Π - a S - b e S T, \\ S (t) \geq - (a + b e T) S . \end{array}$ (8)

The norm defined as

${‖ h ‖}_{\infty} = s u p_{t \in D_{h}},$ (9)

$\begin{array}{l} S (t) \geq - (b e | T | + a) S \\ S (t) \geq - (a b e {‖ T ‖}_{\infty} + a) S . \end{array}$ (10)

This yields

$S (t) \geq S^{0} e^{- (a + b e {‖ T ‖}_{\infty})},$ (11)

$\begin{array}{l} I (t) = b e S T - (k + a) I, \forall t \geq 0 \\ \geq b e S T - (a + k) I \\ \geq - (k + a) I, \end{array}$ (12)

which then yields

$I (t) \geq I^{0} e^{- (a + k)} .$ (13)

Repeating the process for other classes finds the following inequalities:

$T (t) \geq T^{0} e^{- (r + a)}$ (14)

$R (t) \geq R^{0} e^{- a} .$ (15)

3.2. Positive solutions with nonlocal operator

All system (7) findings are favorable if the beginning conditions for nonlocal operators are satisfied.

With regard to the power law kernel and the fractal-fractional operator, we obtain

$\begin{array}{l} S (t) \geq E_{α} S (0) (- q^{1 - υ} (a + b e {‖ T ‖}_{\infty}) t^{α}), \forall t \geq 0 \\ I (t) \geq E_{α} I (0) (- q^{1 - υ} (a + k) t^{α}), \forall t \geq 0 \\ T (t) \geq E_{α} T (0) (- q^{1 - υ} (a + r) t^{α}), \forall t \geq 0 \\ R (t) \geq E_{α} R (0) (- q^{1 - υ} (a) t^{α}), \forall t \geq 0. \end{array}$ (16)

where q is time component.

3.3. Positively invariant region

Theorem 3.1 The PCOS model's suggested solution is separate and limited in $R_{+}^{4}$ under straight-line circumstances.

Proof. We will look at the positive solution of the system (7), which is as follows:

$\begin{eqnarray} _0^{FFP}D_t^{\alpha,\upsilon} S(t)|_{S=0} &=& \Pi \geq 0,\nonumber\\ _0^{FFP}D_t^{\alpha,\upsilon} I(t)|_{I=0} &=& b e S T \geq 0,\nonumber\\ _0^{FFP}D_t^{\alpha,\upsilon} T(t)|_{T=0} &=& k c I \geq 0 ,\\ _0^{FFP}D_t^{\alpha,\upsilon} R(t)|_{R=0} &=& u r T \geq 0.\nonumber \end{eqnarray}$

(17)

The solution is unable to escape the hyperplane if $(S (0), I (0), T (0), R (0)) \in R_{+}^{4}$ . The vector field on each hyperplane around the nonnegative orthant also points into the domain $R_{+}^{4}$ , making it a positivity invariant set.

3.4. Existence and uniqueness analysis

In this part, we used the fixed-point theory to discuss the presence and distinctiveness of the proposed system. Schauder's fixed-point theorem guarantees the model's existence, while Banach's contraction theorem guarantees its singularity. By applying a fractional derivative in the Caputo sense for $0 < α \leq 1$ , system (7) can be generalized as illustrated below:

$\begin{array}{l} _{0}^{F F P} D_{t}^{α, υ} S (t) = Π - a S - b e S T \\ _{0}^{F F P} D_{t}^{α, υ} I (t) = b e S T - (a + k) I \\ _{0}^{F F P} D_{t}^{α, υ} T (t) = k c I - (a + r) T \\ _{0}^{F F P} D_{t}^{α, υ} R (t) = u r T - a R, \end{array}$ (18)

with initial state $S (0) = S_{0}, I (0) = I_{0}, T (0) = T_{0}, R (0) = R_{0}$ . Let

$\begin{array}{l} Δ_{1} (S, I, T, R, t) = Π - a S - b e S T \\ Δ_{2} (S, I, T, R, t) = b e S T - (a + k) I \\ Δ_{3} (S, I, T, R, t) = k c I - (a + r) T \\ Δ_{4} (S, I, T, R, t) = u r T - a R . \end{array}$ (19)

Using initial condition and fractional integral, we have

$S (t) = S_{0} + \frac{1}{Γ (α)} \int_{0}^{t} {(t - ζ)}^{α - 1} Δ_{1} (S, I, T, R, t) d ζ,$ (20)

$I (t) = I_{0} + \frac{1}{Γ (α)} \int_{0}^{t} {(t - ζ)}^{α - 1} Δ_{2} (S, I, T, R, t) d ζ,$ (21)

$T (t) = T_{0} + \frac{1}{Γ (α)} \int_{0}^{t} {(t - ζ)}^{α - 1} Δ_{3} (S, I, T, R, t) d ζ,$ (22)

$R (t) = R_{0} + \frac{1}{Γ (α)} \int_{0}^{t} {(t - ζ)}^{α - 1} Δ_{4} (S, I, T, R, t) d ζ .$ (23)

Let

$Λ (t) = (\begin{matrix} S (t) \\ I (t) \\ T (t) \\ R (t) \end{matrix}, Λ_{0} = (\begin{matrix} S_{0} \\ I_{0} \\ T_{0} \\ R_{0} \end{matrix}, ℵ (ζ, Λ (ζ) = (\begin{matrix} Δ_{1} (S, I, T, R, t) \\ Δ_{2} (S, I, T, R, t) \\ Δ_{3} (S, I, T, R, t) \\ Δ_{4} (S, I, T, R, t) . \end{matrix}$ (24)

Thus, system (20) becomes

$Λ (t) = Λ_{0} + \frac{1}{Γ (α)} \int_{0}^{t} ℵ (ζ, Λ (ζ) (t - ζ)^{α - 1} d ζ .$ (25)

Now, consider a Banach space $℘ [0, T] = ϖ$ with a norm

$‖ S, I, T, R ‖ = \underset{t \in [0, T]}{m a x} [| S, I, T, R |] .$ (26)

Let a mapping be defined as $\diamondsuit:\varpi \rightarrow \varpi$ , then

$\begin{equation} \diamondsuit \Lambda(t)=\Lambda_0 + \frac{1}{\Gamma(\alpha)} \int_0^t (t-\zeta)^{\alpha - 1} \aleph(\zeta , \Lambda(\zeta) d\zeta. \end{equation}$

(27)

In addition, we impose the following hypothesis on a nonlinear function:

P1) There exist constants $ρ_{m}, ρ_{m} > 0$ , such that

$| ℵ (t, Λ (t) | \leq ρ_{m} | Λ (t) | + ρ_{m} .$ (28)

P2) There exists a constant $L_{m} > 0$ for each $Λ, \bar{Λ} \in ϖ$ such that

$| ℵ (t, \bar{Λ}) - ℵ (t, {\bar{Λ}}_{1}) | \leq L_{m} | \bar{Λ} - {\bar{Λ}}_{1} | .$ (29)

Theorem 3.2 The system (18) has at least one solution if the assumptions (P1) are true.

Proof. To demonstrate that $\diamondsuit$ is bounded, let $Φ = {Λ \in Φ | ρ \geq ‖ Λ ‖}$ , where

$ρ \geq \underset{t \in [0, T]}{m a x} \frac{Λ_{0} + (\frac{ρ_{m} T^{α}}{Γ (α + 1)})}{1 - (\frac{ρ_{m} T^{α}}{Γ (α + 1)})}$ (30)

is a closed convex subset of $ϖ$ . Now,

$\begin{eqnarray} \Vert \diamondsuit \Lambda \Vert &=& \max_{t \in [0,\mathbf{T}]}\vert \Lambda_0 + \frac{1}{\Gamma(\alpha)} \int_0^t (t-\zeta)^{\alpha - 1} \aleph(\zeta , \Lambda(\zeta) d\zeta \vert \nonumber\\ &\leq & \vert \Lambda_0 \vert + \frac{1}{\Gamma(\alpha)} \int_0^t (t-\zeta)^{\alpha - 1} \vert \aleph(\zeta , \Lambda(\zeta)\vert d\zeta \nonumber\\ &\leq & \vert \Lambda_0 \vert + \frac{1}{\Gamma(\alpha)} \int_0^t (t-\zeta)^{\alpha - 1} [\rho_m \vert \Lambda (t) \vert + \varrho_m ] d\zeta \\ &\leq & \vert \Lambda_0 \vert + [\rho_m \Vert \Lambda \Vert + \varrho_m ] \frac{\mathbf{T}^{\alpha}}{\Gamma(\alpha + 1)} \leq \rho.\nonumber \end{eqnarray}$

(31)

Since $\Lambda \in \Phi \Rightarrow \diamondsuit (\Phi) \subseteq \Phi$ , it shows that $\diamondsuit$ is bounded. Let $t_{1} < t_{2} \in [0, T]$ to prove that $\diamondsuit$ is completely continuous and take

$\begin{eqnarray} \Vert \diamondsuit \Lambda(t_2) - \diamondsuit \Lambda(t_1) \Vert &=& \vert \frac{1}{\Gamma(\alpha)} \int_0^{t_2} (t_2-\zeta)^{\alpha - 1} \aleph(\zeta , \Lambda(\zeta) d\zeta \nonumber\\ & - & \frac{1}{\Gamma(\alpha)} \int_0^{t_1} (t_1-\zeta)^{\alpha - 1} \aleph(\zeta , \Lambda(\zeta) d\zeta \vert \\ &\leq & [\rho_m \Vert \Lambda \Vert + \varrho_m ] \frac{[t_2^{\alpha} - t_1^{\alpha}]}{\Gamma(\alpha + 1)}.\nonumber \end{eqnarray}$

(32)

This shows that $\Vert \diamondsuit \Lambda(t_2) - \diamondsuit \Lambda(t_1) \Vert \rightarrow 0$ as $t_{2} \to t_{1}$ . The Arzela-Ascoli theorem proves that the operator $\diamondsuit$ is entirely continuous. According to Schauder's fixed point theorem, there is at least one solution for the given system (18). The system (18) was subsequently shown to have an exclusive solution using the Banach fixed point theorem.

Theorem 3.3 Suppose that if the requirements (P2) are met, system (18) has a unique (one) solution.

Proof. Let $\bar{\Lambda}, \bar{\Lambda}_1 \in \varpi$ and take

$\begin{eqnarray} \Vert \diamondsuit (\bar{\Lambda}) - \diamondsuit (\bar{\Lambda}_1) \Vert &=& \max_{t \in [0,\mathbf{T}]} \vert \frac{1}{\Gamma(\alpha)} \int_0^{t} (t_2-\zeta)^{\alpha - 1} \aleph(\zeta , \bar{\Lambda}(\zeta) d\zeta \nonumber\\ &-& \frac{1}{\Gamma(\alpha)} \int_0^{t} (t_1-\zeta)^{\alpha - 1} \aleph(\zeta , \bar{\Lambda}_1(\zeta) d\zeta \vert \\ &\leq & \frac{T^{\alpha}}{\Gamma(\alpha + 1)} L_m \vert \bar{\Lambda} - \bar{\Lambda}_1 \vert.\nonumber \end{eqnarray}$

(33)

Hence, the $\diamondsuit$ is the contraction. By the Banach fixed point (BFP) theorem, system (18) has a unique solution.

3.5. Equilibrium points analysis

This section provides a comprehensive analysis of equilibrium points. For equilibrium points, we solve the system shown below:

$\begin{array}{l} Π - a S^{*} - b e S^{*} T^{*} = 0 \\ b e S^{*} T^{*} - (a + k) I^{*} = 0 \\ k c I^{*} - (a + r) T^{*} = 0 \\ u r T^{*} - a R^{*} = 0. \end{array}$ (34)

The disease free points are

$E^{0} (\frac{Π}{a},0,0,0) .$

Thus, if $I^{*} = 0$ , then we have $S^{*} = \frac{Π}{a}, T^{*} = 0, R^{*} = 0$ , which are free disease equilibrium points. The endemic equilibrium points at this time are

$E^{*} (S^{*}, I^{*}, T^{*}, R^{*}),$

$E^{*} (\frac{(a + k) (a + r)}{b k e c}, \frac{Π b k e c - a (a + k) (a + r)}{b k e c (a + k)}, \frac{Π b k e c - a (a + k) (r + a)}{b e (a + k) (r + a)}, \frac{u r (Π b k e c - a (k + a) (a + r))}{a b e (a + k) (a + r)}) .$

3.6. Reproduction number

Consider the following equation to get the reproduction number:

$\begin{array}{l} _{0}^{F F P} D_{t}^{α, υ} I (t) = b e S T - (a + k) I \\ _{0}^{F F P} D_{t}^{α, υ} T (t) = k c I - (a + r) T . \end{array}$ (35)

The next generation matrix technique must now be used to compute the matrices F and V⁻¹ as follows:

$F = [\begin{matrix} 0 & \frac{Π b e}{a} \\ 0 & 0 \end{matrix}], V^{- 1} = [\begin{matrix} \frac{1}{(a + k)} & 0 \\ \frac{k c}{(a + k) (a + r)} & \frac{1}{(a + r)} \end{matrix}],$ (36)

then the reproduction number is obtained as

$R_{0} = \frac{Π b k e c}{a (a + k) (a + r)} .$ (37)

Figure 1. Analysis of different parameters.

DownLoad: Full-Size Img PowerPoint

4. Stability analysis

4.1. Local stability of the disease-free equilibrium point

Theorem 4.1 If the effective reproduction number is R₀ < 1, at equilibrium point $E^{0} (\frac{Π}{a},0,0,0)$ of model (7) is locally asymptotically stable, and if R₀ > 1, it is unstable.

Proof. The Jacobian matrix J(E⁰) of model (7) with respect to (S, I, T, R) at the without equilibrium point $E^{0} (\frac{Π}{a},0,0,0)$ is as follows:

$J (E^{0}) = [\begin{matrix} - a & 0 & - \frac{Π b e}{a} & 0 \\ 0 & - (a + k) & \frac{Π b e}{a} & 0 \\ 0 & k c & - (a + r) & 0 \\ 0 & 0 & u r & - a \end{matrix}] .$

When it happens, the associated characteristic equation is

$(- a - Υ_{1}) (- (a + k) - Υ_{2}) (- (a + r) - Υ_{3}) (- a - Υ_{4}) = 0,$ (38)

$Υ_{1} = - a$ or $Υ_{2} = - (a + k)$ or $Υ_{3} = - (a + r)$ or $Υ_{4} = - a .$

As a result, every root is negative. Hence, the disease-free equilibrium point is unstable if the effective reproduction number is R₀ < 1 and is locally asymptotically unstable otherwise.

Table 1. Comparative analysis.

Year	Authors	Contribution
2023	Alamoudi et al. [2]	In this paper, a data set containing an ultrasound image of the ovary and clinical information about a patient classified as either PCOS or non-PCOS is presented.
2023	A. Chaudhuri [5]	Women are frequently advised by society to conceal physical issues like PCOS. This study covers the aetiology, diagnosis, causes, symptoms, and potential treatments for PCOS, including medication, herbal remedies, acupuncture, and bariatric surgery.
2022	A. S. Chauhan [4]	Obesity, inactivity, and a family history of PCOS are risk factors for PCOS. Treatment for PCOS may include changes to one's lifestyle, such as frequent exercise and weight loss. Many of the consequences can be prevented with early identification and treatment.
2019	S. Hafezi et al. [7]	The authors of this paper provide a dynamic model designed to forecast the outcome of treatment for infertile women who have polycystic ovarian syndrome.

| Show Table

DownLoad: CSV

5. Numerical simulation

In this part, solutions are obtained using the MATLAB software and the fractal fractional (FF) operator in the Caputo sense. The proposed system's parameter values with initial conditions are as follows: $b = 0.24, e = 0.47, c = 0.34, r = 0.9, u = 0.04$ and $S_{0} = 15, I_{0} = 30, T_{0} = 27, R_{0} = 20, N_{0} = 104$ , as given in [7]. To assess the model's qualitative behavior, the influence of its system characteristics and potential controls with different values of parameters are shown in Figures 1 to 3 at different fractional order values. Simulation of the system at $k = 0.47, a = 0.3$ and $j = 0.5$ is in Figure 1 of different compartments. Here, we can observe better results when shot for a short time and no response in treatment at higher ages given in [7], having a better understanding of the impact of medication on patients. The simulation of the system at $k = 0.47, a = 0.16, j = 0.23$ and $k = 0.1, a = 0.16, j = 0.23$ , respectively, is covered in Figures 2-3 and shows the impact of treatment, where we can observe the susceptible population increase and the infection rate decrease when treatment continues with time. In recent decades, medical therapy for women's difficulties has included the use of gonadotropin and clomiphene citrate. Gonadotrophin or clomiphene citrate are two oral medications that can be used to trigger ovulation in anovulatory PCOS individuals who wish to get pregnant. Drugs called gonadotropin enhance the release of gonadotrophin hormones and cause ovulation, while clomiphene lowers estrogen levels and raises gonadotrophin-releasing hormone (GnRH) levels. With different values of influencing parameters in several age groups, the impact of treating compartments is shown in Figures 2(a-d). Figures 3(a-d) shows better fixing of fractional and fractal parameters, while infection compartments in Figure 3(b) bounded to zero show an improved recovered population at fractional value. The persistence of the disease in the population affects how mathematical models are analyzed. This model assumes that the population volume is stable throughout time and offers a reliable estimate for brief periods with similar patient conditions in early ages. Here, we can also observe the memory effect of fractal and fractional order derivatives in simulations as compared to ordinary derivatives in this model.

Figure 2. Simulation of the PCOS model compartments at different parameter values k = 0.47, a = 0.3, j = 0.5.

DownLoad: Full-Size Img PowerPoint

Figure 3. Simulation of the PCOS model compartments at different parameter values k = 0.47, a = 0.16, j = 0.23.

DownLoad: Full-Size Img PowerPoint

Figure 4. Simulation of the PCOS model compartments at different parameter values k = 0.1, a = 0.16, j = 0.23.

DownLoad: Full-Size Img PowerPoint

6. Conclusion

PCOS patients should first follow a healthy eating regimen. Clinicians should create customized dietary formulas, which include macronutrient ratios, micronutrient intake, and total calorie limitation, based on a comprehensive review of the patient's prior dietary composition. Second, exercise helps patients lose weight and also modifies their body composition. PCOS patients require a customized workout regimen that considers both muscle gain and fat loss, depending on their specific muscle fat ratios. In this article we proposed a fractional order model to predict the outcomes of treatment for infertile women with PCOS as a contributing factor through medication. Using the power law kernel, we developed a fractal-fractional model for observing the treatment impact and transmission of disease in society. We verified the important properties of the epidemic models, such as their positivity, boundedness, positive invariant region, equilibrium points, existence, and uniqueness of their solutions. The analysis of the local stability was done using the Jacobian matrix approach. The examination of the Lyapunov function for global stability was supported by the first and derivative tests. To analyze the effects of the fractional operator with numerical simulations, two-step Lagrange polynomial solutions were constructed. This demonstrates the impact of the sickness on women due to the effect of the many factors involved. According to numerical simulations that support analytical solutions, the dynamics of the development of PCOS are influenced by the fractal-fractional derivatives and they can reduce the spread of the condition in the population. Infertility is a frequent concern for PCOS women. Women with PCOS should be inspired by the possibility of a healthy pregnancy and successful conception. For PCOS-afflicted women should get the right advice and support while seeking medical attention. In this work, we contain a model by using Caputo operator lies properties of kernel which is local and non-singular. This can not implement for a nonlocal operator. In future work we proposed a new fractional model by using nonlocal kernel and optimal control stability and treatment plans that will also include other factors, including food, exercise and everyday activities.

Use of AI tools declaration

The authors declare they have not used Artificial Intelligence (AI) tools in the creation of this article.

References

[1]	World Health Organization, World health statistics 2020: monitoring health for the SDGs, sustainable development goals, 2020.
[2]	J. D. Li, G. Chen, M. Wu, Y. Huang, W. Tang, Downregulation of CDC14B in 5218 breast cancer patients: A novel prognosticator for triple-negative breast cancer, Math. Biosci. Eng., 17 (2020), 8152-8181. doi: 10.3934/mbe.2020414
[3]	Y. Ouyang, Z. Zhou, W. Wu, J. Tian, F. Xu, S. Wu, et al., A review of ultrasound detection methods for breast microcalcification, Math. Biosci. Eng., 16 (2019), 1761-1785.
[4]	M. Zhang, Y. Zhou, Y. Zhang, High expression of TLR2 in the serum of patients with tuberculosis and lung cancer, and can promote the progression of lung cancer, Math. Biosci. Eng., 17 (2020), 1959-1972.
[5]	G. Sun, T. Zhao, Lung adenocarcinoma pathology stages related gene identification, Math. Biosci. Eng., 17 (2020), 737-746. doi: 10.3934/mbe.2020038
[6]	J. Gao, Q. Jiang, B. Zhou, D. Chen, Convolutional neural networks for computer-aided detection or diagnosis in medical image analysis: an overview, Math. Biosci. Eng., 16 (2019), 6536-6561. doi: 10.3934/mbe.2019326
[7]	B. Sahiner, H. P. Chan, N. Petrick, D. Wei, M. A. Helvie, D. D. Adler, et al., Classification of mass and normal breast tissue: a convolution neural network classifier with spatial domain and texture images, IEEE T. Med. Imaging, 15 (1996), 598-610.
[8]	Q. Li, W. Cai, X. Wang, Y. Zhou, D. D. Feng, M. Chen, Medical image classification with convolutional neural network, in 2014 13th international conference on control automation robotics & vision (ICARCV), IEEE, (2014), 844-848.
[9]	T. Kooi, G. Litjens, B. Van Ginneken, A. Gubern-Mérida, C. I. Sánchez, R. Mann, et al., Large scale deep learning for computer aided detection of mammographic lesions, Med. Image. Anal., 35 (2017), 303-312.
[10]	J. Snoek, O. Rippel, K. Swersky, R. Kiros, N. Satish, N. Sundaram, et al., Scalable bayesian optimization using deep neural networks, in International conference on machine learning, PMLR, (2015), 2171-2180.
[11]	M. Z. Alom, T. M. Taha, C. Yakopcic, S. Westberg, P. Sidike, M. S. Nasrin, et al., A state-of-the-art survey on deep learning theory and architectures, Electronics, 8 (2019), 292.
[12]	X. He, K. Zhao, X. Chu, AutoML: A Survey of the State-of-the-Art, Knowl-Based. Syst., 212 (2021), 106622.
[13]	L. Li, A. Talwalkar, Random search and reproducibility for neural architecture search, in Uncertainty in Artificial Intelligence, PMLR, (2020), 367-377.
[14]	P. Liu, M. D. El Basha, Y. Li, Y. Xiao, P. C. Sanelli, R. Fang, Deep evolutionary networks with expedited genetic algorithms for medical image denoising, Med. Image. Anal., 54 (2019), 306-315. doi: 10.1016/j.media.2019.03.004
[15]	P. Esfahanian, M. Akhavan, Gacnn: Training deep convolutional neural networks with genetic algorithm, preprint, arXiv: 1909.13354.
[16]	K. Pawełczyk, M. Kawulok, J. Nalepa, Genetically-trained deep neural networks, in Proceedings of the Genetic and Evolutionary Computation Conference Companion (GECCO), ACM, (2018), 63-64.
[17]	G. L. F. da Silva, T. L. A. Valente, A. C. Silva, A. C. de Paiva, M. Gattass, Convolutional neural network-based PSO for lung nodule false positive reduction on CT images, Comput. Meth. Prog. Bio., 162 (2018), 109-118.
[18]	P. R. Lorenzo, J. Nalepa, M. Kawulok, L. S. Ramos, J. R. Pastor, Particle swarm optimization for hyper-parameter selection in deep neural networks, in Proceedings of the genetic and evolutionary computation conference (GECCO), ACM, (2017), 481-488.
[19]	K. Lan, D. T. Wang, S. Fong, L. S. Liu, K. K. Wong, N. Dey, A survey of data mining and deep learning in bioinformatics, J. Med. Syst., 42 (2018), 1-20. doi: 10.1007/s10916-017-0844-y
[20]	T. Serizawa, H. Fujita, Optimization of convolutional neural network using the linearly decreasing weight particle swarm optimization, preprint, arXiv: 2001.05670.
[21]	J. Tóth, H. Toman, A. Hajdu, Efficient sampling-based energy function evaluation for ensemble optimization using simulated annealing, Pattern. Recogn, 107 (2020), 107510.
[22]	L. R. Rere, B. A. Wardijono, Y. I. Chandra, A comparison study of three single-solution based metaheuristic optimisation for stacked auto encoder, in Journal of Physics: Conference Series, IOP Publishing, (2019), 012066.
[23]	G. Rosa, J. Papa, A. Marana, W. Scheirer, D. Cox, Fine-tuning convolutional neural networks using harmony search, in Iberoamerican Congress on Pattern Recognition, (2015), 683-690.
[24]	R. Gens, P. M. Domingos, Deep symmetry networks, Adv. Neur. Inf., 27 (2014), 2537-2545.
[25]	V. Badrinarayanan, B. Mishra, R. Cipolla, Understanding symmetries in deep networks, preprint, arXiv: 1511.01029.
[26]	K. Lan, L. Liu, T. Li, Y. Chen, S. Fong, J. A. L. Marques, et al., Multi-view convolutional neural network with leader and long-tail particle swarm optimizer for enhancing heart disease and breast cancer detection, Neural. Comput. Appl., 32 (2020), 15469-15488.

mbe-18-05-281- supplementary.pdf

This article has been cited by:

1.	Kamal Shah, Shabir Ahmad, Aman Ullah, Thabet Abdeljawad, Study of chronic myeloid leukemia with T-cell under fractal-fractional order model, 2024, 22, 2391-5471, 10.1515/phys-2024-0032
2.	Tawakalt A. Ayoola, Amos O. Popoola, Morufu O. Olayiwola, Adedapo I. Alaje, Mathematical modeling of chickenpox transmission using the Laplace Adomian Decomposition Method, 2024, 15, 26667207, 100436, 10.1016/j.rico.2024.100436
3.	Esam Y. Salah, Bhausaheb Sontakke, Mohammed S. Abdo, Wasfi Shatanawi, Kamaleldin Abodayeh, M. Daher Albalwi, Md. Kamrujjaman, Conformable Fractional-Order Modeling and Analysis of HIV/AIDS Transmission Dynamics, 2024, 2024, 1687-9651, 1, 10.1155/2024/1958622
4.	Morufu Oyedunsi Olayiwola, Adedapo Ismaila Alaje, Mathematical analysis of intrahost spread and control of Dengue Virus: Unraveling the crucial role of antigenic immunity, 2024, 7, 27731863, 100117, 10.1016/j.fraope.2024.100117
5.	Hasanen A. Hammad, Manuel De la Sen, Existence of a mild solution and approximate controllability for fractional random integro-differential inclusions with non-instantaneous impulses, 2025, 111, 11100168, 306, 10.1016/j.aej.2024.10.017
6.	Muhammad Imran Liaqat, Adnan Khan, Hafiz Muhammad Anjum, Gregory Abe-I-Kpeng, Emad E. Mahmoud, S. A. Edalatpanah, A Novel Efficient Approach for Solving Nonlinear Caputo Fractional Differential Equations, 2024, 2024, 1687-9120, 10.1155/2024/1971059
7.	K. Muthuvel, K. Kaliraj, Kottakkaran Sooppy Nisar, V. Vijayakumar, Relative controllability for ψ−Caputo fractional delay control system, 2024, 16, 26667207, 100475, 10.1016/j.rico.2024.100475
8.	Munkaila Dasumani, Binandam S. Lassong, Isaac K. Adu, Fredrick A. Wireko, Stephen E. Moore, Jesus Manuel Munoz-Pacheco, Fractional Derivative Technique for Modeling the Dynamics of Social Media Impacts, 2024, 2024, 1026-0226, 10.1155/2024/5578416
9.	Savita Rathee, Yogeeta Narwal, Komal Bansal, Homan Emadifar, Sensitivity analysis of fractional order SVEIR Lumpy Skin Disease model, 2025, 119, 11100168, 609, 10.1016/j.aej.2025.01.076

Reader Comments

Your name:*

Email:*
© 2021 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)