Diabetes Mellitus Prediction and Classification Using Firefly Optimization Based Support Vector Machine

The Diabetes mellitus is the most dangerous disease all over the world and it is deadly disease determined by the rapid growth levels of the glucose in a blood. The Machine Learning (ML) approaches supports in the prediction and detection of the diabetes. This research aims to propose a Firefly Optimized-based Support Vector Machine (FSVM) for the prediction and classification of diabetes. The datasets utilized for diabetic prediction are PIMA Indian Diabetic dataset. The pre-processing is done by using the data normalization, cleaning as well as data splitting techniques. The Principal Component Analysis (PCA)-based feature selection approach is used for selecting the features in a acquired dataset. Finally, the ML-based algorithms are used for the classification of the diabetes disease. The effectiveness of proposed method is evaluated by PIMA dataset with the various performance metrices such as accuracy, precision, recall, and F1-score of values about 98.45%, 98.69%, 99.37%, and 98.25% respectively when compared to previous approaches such as Naïve Bayes, Integrated kernel as well as K-Nearest Neighbour.