Performance evaluation of immune-inspired support vector machine

Immune-inspired approach designs ensembles of medical images for classification problems using neural network. Immune SVM is a classification algorithm that replaces the traditional SVM by optimising the parameters of SVM. Among the additional attributes provided by the SVM, the immune algorithm invokes automatic control to the population size along the search, improves the convergence speed and maintains the diversity of the antibody population. In this paper, performance of the immune SVM classifier is analysed by optimising SVM parameters. The experimental result shows that brain tumour detection using immune SVM provides greater recognition accuracy. It also shows good performance and promising results to assist surgeons and medical practitioners in detecting tumour.