A REAL TIME DIABETIC DISEASE PREDICTION USING PATIENTS DATASET THROUGH KNN, SVM AND BPSO ALGORITHMS

Abstract

The evolution of information technology and the standardization of terminologies in the health area has generated large repositories of data that can be mined to enable discovery of knowledge to assist in the early identification of sick patients as well as cause and effect relationships. Among the diseases, Diabetes Mellitus (DM) stands out due to the increase in the number of cases. This, the
sooner it is discovered, the better and more economical its treatment becomes. Thus, finding a standard in the use of health plans would be useful for discovering and classifying patients affected by the disease, enabling treatments to be carried out in a shorter time, with improvement in the patient's quality of life and cost reduction. The technique applied in this study is the application of
algorithms that generate Binary Particle Swarm Optimization (BPSO), Support Vector Machine (SVM), K-Nearest Neighbor (KNN) and Naïve Bayes classifiers, through Data Mining, whose objective is the classification and selection of these patients for inclusion in preventive medicine programs.