An empirical analysis using Intelligent computational methods On risk factors of cervical cancer to Aid early diagnosis

Abstract

Cervical cancer is one of the most widespread diseases among the women that have a high mortality rate. Early detection through screening methods helps to reduce the mortality rate and extends the life of women. World Health Organization (WHO) estimates that, cervical cancer has become the fourth leading cancer that affects women. Surveys reveal that just about 90% of deaths due to this disease occur in low- and middle-income countries as there is lack of awareness about cancer and its prevention measures. The main idea is to find the important causes and risk factors for a woman to acquire cervical cancer and educate them to impart good quality life style for prevention from the disease. The revised cervical cancer staging classifies the carcinoma of the cervix uteri in four stages and treatments are generally dictated based on the stages. The survival rate of the patient is high if and only the incidence of cancer is found in the earlier stages. The 5-year survival rate of cervical cancer is above 90 percent if detected at the earlier stage. Upon contemplation of the above, this research work focuses on identifying the key risk factors of cervical cancer and thereby reduces the incidence and mortality of women in poor socio-economic background. Risk factors associated with cervical cancer are identified from the dataset as features and analyzed in terms of the target Biopsy as a preliminary work. Feature extraction using computational methods are implemented in identifying the feature which has a momentous contribution towards the possibility of cervical cancer. Chi-square analysis, Correlation coefficient method and Genetic Algorithm are used for feature selection and the best relevant features are extracted from the dataset. Support Vector Machine (SVM) Linear Classifier is used as a classifier model in predicting the possibility of cervical cancer newline