Breast Cancer Diagnosis (BCD) Model Using Machine Learning

In the recent years, breast cancer research has made a significant growth however there is still a scope of advancement. Breast cancer increases the statistics of mortality among women. In concern to this issue, treatment of cancer should be started at the earlier stage, to increase the chances of survival of the patient. Thus, there is a need to diagnose breast cancer at the early stage using the features from the mammograms. This paper proposes an efficient BCD model to detect breast cancer by using Support Vector Machine (SVM) with 10-fold cross validation. The complexity of the problem increases if there are many input features for the diagnosis of cancer. Thus, Principal Component Analysis (PCA) is used to reduce the feature space from a higher dimension to a lower dimension. Experiment result shows that the PCA increases the accuracy of the model. The proposed BCD model is compared with other supervised learning algorithms like Decision trees (DT), Random Forest, k- Nearest Neighbors(k-NN), Stochastic Gradient Descent (SGD), AdaBoost, Neural Network (NN), and Naïve Bayes. Evaluation parameters like F1 measure, ROC curve, Accuracy, Lift curve and Calibration Plot proves that proposed BCD model outperforms and gives the highest accuracy among other compared algorithms.