Detection of Breast Cancer Using Machine Learning and Deep Learning Methods

The most common reason for women’s deaths worldwide is Breast Cancer, which can be controlled if detected at an early stage. There are numerous techniques based on Machine Learning as well as Deep Learning can be used for the diagnoses of the cancer. This paper employs Convolutional Neural Network (CNN), which also performs feature extraction for the Machine Learning algorithms in comparison with Support Vector Machine (SVM), Naïve Bayes classifier (NB), K-Nearest Neighbor (KNN), Logistic Regression (LR) and Random Forest (RF). The system is experimented on BreaK-Hist Dataset obtained from Kaggle. The train_test_split method is used to bifurcate the raw data to distinct sets in order to achieve appropriate results. Softmax activation function predicts the outcomes based on probabilities. The main objective is to correctly classify the data and asses the performance of the models in comparison based on precision and recall rate. Further it also asses based on f1- score and as well as accuracy. When experimented the results show that CNN gives maximum accuracy with the lower most loss/error rate.