A Comparative Analysis of Regression Algorithms with Genetic Algorithm In The Prediction of Breast Cancer Tumors

Abstract

Over the last few decades, breast cancer has become a major health concern worldwide, particularly in the women community, as its root cause is not always known and most times it is diagnosed in advanced stages which leads to high death rate. In more recent times, machine learning techniques have been employed as computer-aided diagnosis tools for breast cancer prediction These machine learning techniques have the capacity to classify and predict this cancer into benign or malignant. The main contribution of this study is to find a model which is most suitable for predicting this kind of tumor cell. Genetic Algorithm is applied as Feature Selection method to the Wisconsin Breast Cancer dataset to select the subsets of input features that are most relevant to the target variable. We compared five machine learning regression classifiers were considered, Linear regression, Logistic regression, Random Forest, Decision Tree and Support Vector Regression. The Random Forest classifier obtained the best precision and performance accuracy. This study contributes towards the enhancement of medical technology for prediction of breast cancer, which will not only improve the well-being and health of the female community but also reduce mortality rate related to breast cancer.