Breast Cancer Prediction Model Using Machine Learning

Abstract

Breast cancer requires early detection, hence it can be prevented earlier or treated more optimally. This
article aims to demonstrate predictive modelling of breast cancer and evaluate the accuracy of its
predictions using a machine learning approach. This study uses secondary data from the Wisconsin
Breast Cancer Dataset (BCWD) which consists of predictive factors for breast cancer and labels for
benign or malignant cancers that result. Modelling with machine learning is done by selecting three
candidate algorithms, namely Random Forest, Support Vector Machine, and Logistic Regression.
Evaluation of the classification performance of each algorithm is carried out by analysing its sensitivity,
specificity, and accuracy. The experimental results show that Random Forest has better prediction
accuracy (99.6%) followed by Support Vector Machine (98.7%), and Logistic Regression (93.9%).