Coronavirus Algorithm for Features Optimization in Breast Cancer Classification

Abstract

Breast cancer is one of the most common medical problems that need early diagnosis. The early diagnosis helps on effective treatment of this disease; thus, techniques must be developed to assist clinicians in obtaining an accurate diagnosis. However, this task is challenging due to the magnitude of the problem and the variability of breast cancer prognostic data. This work aims to develop an approach that will increase the precision of breast cancer diagnosis. This aim was achieved by integrating the data feature optimization algorithm with classification algorithms. Firstly, to improve the features of data on breast cancer, the coronavirus algorithm was used as a reference to optimize the Wisconsin Diagnostic Breast Cancer (WDBC) dataset. Secondly, improved data scaling was performed before the classification process. Lastly, the outputs of the coronavirus algorithm were combined with those of machine learning algorithms Projective Adaptive Resonance Theory (PART) and Decision Tree (J48) algorithm. performed Integration of the coronavirus algorithm and a deep learning algorithm (CNN model). The proposed approach was implemented and evaluated on the WDBC dataset obtained from the University of California, Irvine, Machine Learning Repository. The evaluation of the model depended on the precision of classification, retrieval and measurement, and the proposed method was compared with different classification algorithms applied on the same dataset. Experimental results showed that the proposed classification approach exhibited competitive classification precision. The precision was 94.01% for the J48 algorithm, 94.18% for the PART algorithm and 100% for CNN. Feature optimization is very important to improve classification precision by using coronavirus algorithm