부산대학교 도서관

In this digital age, most of the applications need to be Internet of Things based for easy access. Biomedical research has made significant growth related to it. The visual pathological information provides comparatively better information as compared to signals that help physicians with further analysis and suggestions. Medical data often exhibit class imbalance, where the number of samples in different classes is not proportional. This imbalance can pose significant challenges to traditional classification algorithms. This report explores the application of Support Vector Machines (SVM) to imbalanced medical data classification, discussing the inherent issues, techniques for addressing class imbalance, and experimental results showcasing the effectiveness of SVM in handling imbalanced medical datasets. As a result, the retrieval process over the internet will be faster. In this paper, authors have considered three different cancerous image data of breast cancer, brain tumor, and liver cancer for analysis and detection using a machine learning approach. The data is collected from the UCI machine learning repository. The statistical features like class, radius, and area are considered and optimized using the Brain Storm Optimization algorithm for good accuracy. The optimized features are applied to the support vector machine classifier. Further, use for different kernels such as radial basis function (RBF), linear, polynomial, and sigmoid are used with SVM and verified for comparison. It is found that the RBF kernel provides SVM better accuracy as compared to other kernels.