Optimized machine learning model using Decision Tree for cancer prediction

Abstract

Decision making is used for the selection oriented process. But can we use it in the field of healthcare is the biggest question. The answer is yes here also we can have with machine learning (ML) classifiers like Decision Tree (DT), random forest (RF), support vector machine (SVM), and neural network (NN). In the case of breast cancer (BC) disease, we can predict using machine learning classifiers. For this, original BC samples experiment. Our model named clustering based feature selection (CFS) intends to increase the accuracy and to reduce the feature dimension. We experiment the classifiers in three stages: 1. all features of BC, 2. DT based cluster, and 3. Feature selection. In all feature, RF achieves 99.10% accuracy, in DT based cluster RF, achieves 99.37% accuracy, and finally in cluster-based feature selection (CFS) with only four features the RF achieves 99.50% accuracy. The experimental results of our CFS approach along with DT, RF, SVM, and NN classifier along with its performance are presented.