Fuzzy Associative Memories Approach to Analyze the Oral Cancer

 

A. Felix, J. Ravi Sankar^*, G. Mokesh Rayalu

Department of Mathematics, School of Advanced Sciences, VIT University, Vellore, India.

 

ABSTRACT:

The purpose of this present research is to determine the threat and achievement factors of oral cancer using Fuzzy Associative Memories. Mouth Cancer is one of the many cancers that people are being affected all over the world. India has one of the highest levels of people being affected to these diseases. People’s perception on Oral Cancer is on just smoking, but smokeless tobacco usage among people has become so highly dangerous that it has been made to be a major health trouble. It is observed from the medical experts that Smoking, usage of Tobacco products and drinking Alcohol in temperance and refrains from indulge drinking are the reasons for Oral cancer. But recent medical diagnosis reported that imbalanced diet and also too much exposure to the sun has increased the danger of cancer. Since there is uncertain and complex situation to find the cause and effect of oral cancer, Fuzzy logic based model called Fuzzy Associative Memories used to analyze this issue.

 

 

 

1. INTRODUCTION:

Now a day’s cancers are the most common thing that causes death. Among those cancers, oral cancer takes third place in India by Kosko et al in  [1]. It occurs in any part of the mouth like lips, tongue, cheek lining, under the teeth, neck. Intake of extreme alcohol, tobacco(cigarette), smokeless tobacco, chewing of betel nut, family history,  intense sun exposure, diet these are the most regular peril factors of throat cancer. In India, south and Southeast Asian countries are highly affected by oral cancer compared to other countries in the paper of Rajesh Dikshit et al. [5]. A study of case control that was taken from India, Oral cancer is of considerable public health importance of India. Firstly, oral cancer is found at the later stage which conclusion in low treatment outcome. Secondly, rural areas are interconnected with less income.

 

Low economic class gives low nutrition, lack of health care, awful living conditions are the reason for the advance of oral cancer. Slavkin HC point out that cancers are genetic diseases that are because of cell growth abnormally as the effect of damaged DNA in [6]. From the above information about the income status, India is among the many low-income and middle income countries among many other countries that the expenditures makes the full family below poverty line as a diagnosis of cancer very often leads to very high a  person’s health expenditure[7]. The earlier cancers so far had been diagnosed, so the cure will be victorious. But final stage of the cancer is highly tough to cure. More than thousand people pass away every year because of increasing of the cancer as there is inadequacy of awareness and medical resources. Due to non availability of medical practitioner, people from rural background are failed to diagnose their disease. To assist them, experts based knowledge system  could be constructed to diagnosis any particular diseases. Recently many researchers constructed medical based diagnosis system for disease such breast cancer, diabetes, hypertension, etc [3,4,,8,9,10]. But there is drawback that many researchers have not focused on the cause effect of the disease. Since it is complex to obtain the intensity of the relationship between cause and effect factor of oral cancer, Fuzzy Associative Memories is used to analyze this problem.

 

2. PRELIMINARIES:

Some basic definitions of fuzzy sets, fuzzy numbers,  Linguistic terms and Fuzzy Associative Memories are provided in this section.

 

Fuzzy sets was proposed by Zadeh LA (1965) [11]to deal vagueness of  any problem. It yields lot of application in the field of medical diagnosis, Engineering, Many other decision making problem etc.,

 

Definition 2.1.

A fuzzy set is a subset of a universe of discourse X, which is characterized by a membership function  representing a mapping  The function value of  is called the membership value, which represents the degree of truth that x is an element of fuzzy set .

 

Definition 2.2.

A fuzzy set is fuzzy number if it has the following characteristics,

(i)   it is convex.

     

(ii)   it is normal if.

(iii)   it is piecewise continuous.

 

Definition 2.3.  

A triangular fuzzy number can be defined as a triplet (l, m, r) and the membership function is defined as:

 

Where l, m, r are real numbers and

 

Fig 2.1 Triangular Fuzzy Number

 

Definition 2.4.

A linguistic variable is a variable whose values are linguistic terms (Zadeh, 1975) [12].

 

2.1   Fuzzy Associative Memories:

Bart Kosko proposed Fuzzy Associative memories (1997)[2] to find the relationship between causes and effects of the problem.

 

Definition 2.1.1:

The n-dimensional unit hypercube is denoted as  Iⁿ = [0, 1] ⁿ = [0, 1]* . . . * [0, 1]. A fuzzy set defines a point in the cube Iⁿ and vertices of cube Iⁿ are considered as non- fuzzy sets. The n-dimensional unit hyper cube Iⁿ the fuzzy subsets are of the form  X = {x₁ ,x ₂,...,x _n}.

 

Definition 2.1.2: 

Fuzzy system defines mappings S: Iⁿ → I ^p, where n and p are finite positive integers. The n-dimensional unit hypercube consists of all the fuzzy subsets of the domain space X = {(x ₁, x₂ ,... ,x _n ) | x_i∈  R, i =1,2, ... ,n}. Similarly  I^p consists of all the fuzzy subsets of the range space  Y = {(y ₁,y ₂,... ,y _p ) | y_i∈  R,   i =1,..., p }.  Here X R_n and YR_p. The system maps similar inputs to similar outputs. These continuous fuzzy systems behave as an associative memory known as fuzzy associative memory. Thus fuzzy associative memories are transformations.  

 

Definition 2.1.3: 

The fuzzy set association (A_i, B_i) is named as a “rule”. The antecedent term Ai and the consequent term Bi in the fuzzy set association (Ai, Bi) are known as input associate and output associate respectively. The FAM system maps points A_j near A_i to points Bj near B_i.  If Aj is closer to Ai, then the point (A_j, B_j) is closer to (Ai, Bi) in the product space Iⁿ × I^p. Using the rule between the antecedent A_i and consequent Bi, the connection matrix M is obtained.

 

Definition 2.1.4: 

If the equilibrium state of a dynamical system is a unique state vector, then it is called a fixed point.

 

Definition 2.1.5:

If the state vector repeats in the form of then this kind of equilibrium point is called limit cycle.       

 

 

3. ADAPTATION OF THE PROBLEM TO THE MODEL:

More than fifteen patient of oral cancer are interviewed in Vellore, Tamil Nadu using the unsupervised method to identify the factor of oral cancer. The following attributes are chosen from their statement.

 

CAUSES(C)	EFFECTS(E)
C1- Smoking	E1-Both lungs are affected, gum diseases, affects digestive system, heart diseases, diabetes, pre matured weight loss/over weight babies, stained teeth.
C2- Smokeless tobacco	E2-Gums damaged, throat pain, sores, pancreas affected, stomach problem.
C3- Excessive consumption of alcohol	E3-Low immunity, brain affected, liver problem, liver damaged, pancreases damaged, heart problem.
C4- Family history of oral cancer	E4-By family history also oral cancer occurs.
C5- Excessive sun exposure	E5-Itchiness, redness, sunburn, dizziness, nausea, lips problem, head ache.

 

Then, with aid of 3 different medical experts the following relational matrices are constructed from the linguistic set L={Very Low (VL), Low (L), Medium (M), High (H), Very High (VH)}.

 

 

Table: 3.1 Linguistic Relational Matrix from the Expert-1

 

 

 

Table:3.2  Linguistic Relational Matrix from the Expert-2

 

 

Table:3.3 Linguistic Relational Matrix from the Expert-3

 

After constructing the relational Matrix, transform them into triangular fuzzy number. Then, triangular fuzzy value matrices are converted into single valued matrices  using the table.

 

Table: 3.4 Linguistic Fuzzy Scale

 

Table: 3.5 Fuzzy Relational Matrix from the Expert-1

 

Table: 3.6 Fuzzy Relational Matrix from the Expert-2

 

Table: 3.7 Fuzzy Relational Matrix from the Expert-3

 

Next, determine the average fuzzy relational matrix. It is considered as the dynamical system M.

 

Table 3.8 Dynamical system M

	E1	E2	E3	E4	E5
C1	0.9375	0.6458	0.6666	0.6666	0.3125
C2	0.7916	0.9375	0.7291	0.4375	0.125
C3	0.5	0.5833	0.875	0.5	0.1875
C4	0.4375	0.5208	0.5833	0.75	0.125
C5	0.0625	0.0625	0.0625	0.0625	0.7291

 

Case-1:Let us consider the patient's with his level of symptoms P₁= {( C1,VH)  (C2, H) (C3,M) (C4, NI)  (C5, L)}

 

Convert the linguistic term into fuzzy number to obtain a input vector P₁= {0.9375,  0.75, 0.5, 0, 0.25}. Then, pass the input vector into dynamical system using min-max composition.

X1M= (0.9375, 0.75, 0.7291, 0.6666, 0.7291) =Y1

Y1M^T= (0.9375, 0.7916, 0.7291, 0.6666, 0.7291) =X2

X2M= (0.9375, 0.7916, 0.7291, 0.6666, 0.7291) =Y2

Y2M^T= (0.9375, 0.7916, 0.7291, 0.6666, 0.7291) =X3=X2

X3M= (0.9375, 0.7916, 0.7291, 0.6666, 0.7291) =Y3=Y2

 

From the limit point, first two maximum value are chosen to identify the patient health condition.

 

Since the patient P₁ is addicted to C₁- Smoking and C₂- Smokeless tobacco, it leads to the chance of having  E₁-Both lungs are affected, gum diseases, affects digestive system, heart diseases, diabetes, pre matured weight loss/over weight babies, stained teeth. E₂-Gums damaged, throat pain, sores, pancreas affected, stomach problem.

 

Case-2:  Let us consider the patient's with his level of symptoms P₂= {(C1,VH)  (C2,H) (C3,M) (C4,L) (C5,NI)}

Convert the linguistic term into fuzzy number to obtain a input vector P2={0.9375,  0.75, 0.5, 0.25, 0}. Then, pass the input vector into dynamical system using min-max composition.

X1M= (0.9375, 0.75, 0.5, 0.25, 0) =Y1

Y1M^T= (0.9375, 0.75, 0.7291, 0.6666, 0.3125) =X2

X2M= (0.9375, 0.7916, 0.5833, 0.6666, 0.3125) =Y2

Y2M^T= (0.9375, 0.7916, 0.7291, 0.6666, 0.3125) =X3

X3M= (0.9375, 0.7916, 0.7291, 0.6666, 0.3125) =Y3

Y3M^T= (0.9375, 0.7916, 0.7291, 0.6666, 0.3125) =X4=X3

X4M= (0.9375, 0.7916, 0.7291, 0.6666, 0.3125) =Y4=Y3

 

By choosing the first two maximum value from the limit point,  it is identified the patient health condition. The P₂ is also addicted to tobacco and smoking. Therefore P2 will have chance of affecting the disease same as P₁.

 

Suggestion to prevent  Oral Cancer:

Always brush and floss your teeth regularly:

A harmful mouth reduces our protected system and inhibits remains ability to fight rotten potential cancers.

 

Do not burn tobacco manufactured goods:

We are smoker, even with a casual habit, make the resolution to stop.

 

Regular Exercise:

An energetic daily life is recognized to improve the unaffected system and help charge off tumor.

 

 

 

Limit Alcohol:

Try lesser measures. Rather than sticking to pints, try sipping halve, go for a bottled beer or if we are drinking wine, opt for a small glass. Go watered down. Try a more diluted alcoholic drink such as a splitter or candy.

 

Bound your spotlight to the sun:

Stay out of the sun, especially between 10 am to 4 pm when sunlight is strongest. Always use lip balm and particular lower lip.

 

4.  CONCLUSION:

In this Paper, cause and effect of oral cancer is analyzed using  Fuzzy Associative Memories used. Also two different two patients of oral cancer were taken to find the cause and effect of the oral cancer. This system confirms the actual causes and effects of the health condition.

 

5. REFERENCES:

1.     Kosko, B.(1987), Adaptive Bidirectional Associative Memories. Applied Optics, 26, 4947-4960.

2.     Manisha Sharma, Manas Madan, Mridu Manjari, Tejinder Singh Bhasin, Spriha Jain, Saumil Garg, "Prevalence of Head and Neck Squamous Cell Carcinoma (HNSCC) in our population: The clinicpathological and morphological description of 198 cases".

3.     Pankaj Srivastava, Neeraja Sharma, Richa Singh, “Soft Computing Diagnostic System for Diabetes”, International Journal of Computer Applications (0975 – 888) Volume 47– No.18, June 2012.

4.     Pankaj Srivastavaand Neeraja Sharma, “A Spectrum of Soft Computing Model for Medical Diagnosis”, Appl. Math. Inf. Sci. 8(3), 1225-1230, 2014.

5.     Rajesh Dikshit et.al, "Cancer mortality in India: a nationally representative survey". www.thelancet.com Published online March 28, 2012 DOI:10.1016/S0140-6736(12)60358-4 http://environmentportal.in/files/file/Cancer%20mortality%20in%20India.pdf.

6.     Slavkin HC. “The human genome, implications for oral health and diseases,and dental education”. J Dental Education, 65: 463–479, 2001.

7.     Sree Vidya Krishna Rao, Gloria Mejia, Kaye Roberts-Thomson,Richard Logan, "Epidemiology of Oral Cancer in Asia in the Past Decade- An Update (2000-2012)”.

8.     Srivastava, Pankaj and Srivastava, Amit, Spectrum of Soft Computing Risk Assessment Scheme for Hypertension, International Journal of Computer Applications, 44(17), 23-30, 2012.

9.     Victor Devadoss A., Felix, A, A new Fuzzy DEMATEL method in an Uncertain Linguistic Environment, Advances in Fuzzy Sets and Systems, 16(2), 93-123, 2013.

10.   Victor Devadoss A., Dhivya A D, “A Study on T2DM Using Linguistic Fuzzy Technique”, International Journal of Applied Engineering Research, 10(79), 309-314, 2015.

11.   Zadeh L.A., Fuzzy sets, Inform. and Control, 8, 338-353, 1965.

12.   Zadeh L.A., The concept of a linguistic variable and its application to approximate reasoning, Inform.Sci.8,199-249(I), 301-357(II), 1975.

 

 

 

Accepted on 17.11.2017        © RJPT All right reserved

Research J. Pharm. and Tech 2018; 11(2):523-526.