Critical Review on Breast Cancer Classification

In recent days, most common disease which causes death is breast cancer. Worldwide, it is a second common disease causing more death. Around 4.5 to 5% new cases are recorded every year and at worldwide, morbidity is increased. Patient’s survival rate can be increased by cancer’s early detection with proper treatment and diagnosis. Classification accuracy and feature selection are major issues in available techniques. Various available optimization and data mining techniques are analyzed for rectifying above mentioned issues. From breast cancer dataset, hidden knowledge are obtained using those techniques. For enhancing specified breast cancer dataset’s accuracy significantly, classification process, feature selection and pre-processing are performed in this work. Every algorithm’s shortcomings and advantages are also analyzed in this survey. Different classification methods like FE-BkCapsNet, Convolution Neural Network (CNN), K Nearest Neighbor (KNN), Support Vector Machine (SVM) are used for making comparative analysis. With respect to f-measure, sensitivity, specificity, accuracy metrics, better performance are provided by FE-BkCapsNet approach when compared with other techniques as indicated in experimental results.