Abstract:

Abstract In this paper, we propose a new evolutionary classifier that is capable of achieving better prediction results than the conventional classifiers and the earlier evolutionary classifiers that are available in the literature. To reduce the severity level of diabetes and also predict the disease levels as Type-1 and Type-2. Moreover, a new evolutionary classifier incorporating a diabetic monitoring system is proposed to monitor the diabetic disease by using the newly proposed brain storming classification algorithm that applies the existing Enhanced Fireworks Algorithm and Brian Storm Optimization method. This work improves the structure of the evolutionary classifier by enhancing the classification accuracy. Feature selection is necessary to handle the huge datasets. So, the proposed model extracts the necessary features first by applying the newly proposed Learning Automata and Fireworks Algorithm based Feature Selection Method to identify the most important features that are helpful to enhance the prediction accuracy. This work is evaluated by using the different diabetic datasets from the UCI Repository and hospital datasets by conducting various experiments and is also proved to be better than other models by considering the evaluation metrics, namely precision, recall value, f-measure value, and decision accuracy.

Keywords: