Hybrid Prediction Model for Type 2 Diabetes Using Pso-Gaussion Kernal Based Fuzzy C-Means and Pdf-Recurrent Neural Network
Abstract
Diabetes is one of the prevalent diseases in the world. The untreated and unidentified diabetes disease could cause instabilities in the sugar levels and in extreme cases, it damages the organs such as kidneys, eyes, and arteries in the heart. The Machine learning algorithms used to predict the disease automatically from the relevant datasets at an early stage could likely save human lives. The purpose of this research work to proposed better prediction model for the diabetes prediction. A proposed model based on a sequence of pre-processing procedures it included of two parts. First, the Particle Swarm Optimization with Gaussian kernel based fuzzy c-means algorithm (PSO-GKFCM) used for data reduction and, in second partthe Probability Density Function with Recurrent neural network algorithm (PDF-RNN) used for classification. As a result of the proposed model, performs better accuracy in overall process when compared with existing algorithm k-means, FCM and LR, CNN.