Cervical Cancer Diagnosis Using a Chicken Swarm Optimization Based Machine Learning Method

Abstract

Cervical cancer occurs in the cells of the cervix and in the early stages it generally produces no symptoms or signs. The detection of the cervical cancer as early as possible is still a research challenge and the most common four tests are: Hinselmann, Schiller, Citology and Biopsy. The correlation of the features, the small number of data samples and the high class imbalance are still research challenges that must be considered in the cervical cancer diagnosis. In this article the cervical cancer diagnosis is approached using a machine learning method in which the features are selected using linear correlation and the data is classified using the Support Vector Machines (SVM) algorithm. The hyperparameters of the SVM are selected using Chicken Swarm Optimization (CSO). The method is tested and validated on the publicly available Cervical cancer (Risk Factors) Data Set from the UCI Machine Learning Repository.