A Comparative Study on Breast Cancer Prediction using Optimized Algorithms

Abstract

Breast cancer is a type of cancer that develops in women as a result of abnormal cell growth in their breast and it is the second largest cause of mortality in women. Breast cancer detection and treatment have improved as a result of breakthroughs in the field; around 89 percentages of women who are diagnosed with breast cancer survive even after their diagnosis. Early detection of breast cancer is essential for the survival; therefore, a technology that can detect breast cancer in its early stages must be developed. Various machine learning algorithms from data mining have been used to identify breast cancer. The objective of this research is to assess the prediction accuracy of Divide and Conquer Kernal Support Vector Machine (DCKSVM) and Hybrid Radial Basis Function Neural Network machine learning algorithms for breast cancer disease using the Wisconsin Diagnostic Breast Cancer (W DBC) dataset. The assessment of the algorithms is made using three different parameters termed as Accuracy, Precision and Recall. In this research work the pre-processing of the dataset to fill the missing attribute values is done using the K-Nearest Neighbor algorithm. The result proves that the HRBFNN algorithm gives better accuracy when compared with the DCKSVM algorithm.