Classification of Breast Cancer: A Comparative Study using K-Means Clustering-Based Feature Extraction and Hybrid Classifier

In order to save a person’s life, it is essential to diagnose and identify cancer as early as possible. Breast cancer is one of the deadliest diseases women may have, and it’s a major focus of medical research. Breast cancer starts in the cells and tissues of the breast. A breast lump, along with other symptoms including an abrupt shift in breast size or shape, leaking fluid from the nipple, and a generally moist feeling to the skin, are all telltale markers of breast cancer. As the disease progresses, symptoms such as shortness of breath, bone discomfort, and swollen lymph nodes may become apparent. Breast cancer is the second biggest cause of death among women worldwide. It’s one of the most common forms of cancer among females. Since the root cause of the condition is unknown, early discovery, classification, and diagnosis are essential. If breast cancer is detected early, it is highly curable. In this work the breast cancer risk of patients is classified using a Gaussian Mixture Model (GMM), and the hybrid classifiers Firefly GMM, Cuckoo Search GMM, Harmonic Search GMM and PSO GMM the results are presented in terms of classification accuracy, error rate, sensitivity, specificity and Jaccard Metric. Among these five classifiers the hybrid classifier PSO GMM attained maximum accuracy of 94% with minimum error rate of 6%.