Early Prediction of Breast Cancer through Machine Learning with Minimal Features

Abstract

One of the wide and fast spreading diseases among the younger age groups of women is breast cancer. From recent survey it is revealed that once in four minutes a woman is diagnosed whereas a woman dies once in eight minutes due to late detection. If such cases are detected earlier, their lifetime could have been extended. Hence, the objective of the proposal is to predict the presence of breast cancer earlier through deep learning. Deep learning model is implemented in python programming language by using keras Application Programmable Interface and the accuracies of the popular machine learning models such as Logistic Regression, K Nearest Neighbours, Support Vector Machine (linear), Support Vector Machine (RBF), Gaussian (NB), Decision Tree and Random Forest are computed. Initially, the data set with 30 attributes are considered and then feature selection is carried out through heat map. The model consists of number of hidden layers which performs binary classification on the given dataset to predict whether a person is malignant or benign. The proposal exhibits its supremacy by demonstrating greater accuracy and almost similar confusion matrix and execution time in prediction with reduced attributes obtained through feature selection.