CHAOTIC GIANT TRAVELLY OPTIMIZER BASED FASTER REGION BASED CONVOLUTIONAL NEURAL NETWORK FOR THE RISK PREDICTION OF DIABETES DISEASE IN NORTH KASHMIR

Povzetek

Diabetes is a chronic disease that occurs due to the rise of blood sugar level. The reason for diabetes varies with patients and in North Kashmir it rising now-a-days. To analyze the prevalence and risk of diabetes in North Kashmir we have conducted the study among 1023 people and collected the data for further processing. From the collected data we have predicted the risk of diabetes with innovative approaches. The prediction of disease by various approaches shows higher prediction loss, lower training speed with higher computational complexities. To overcome these issues, in this work we propose a deep learning based approach for Diabetes Disease Risk Prediction in North Kashmir. In this work, the data is pre-processed using Guided anisotropic difusion filtering (GADF). Subsequently, the feature extraction is performed using novel Young’s double slit experiment optimizer (YDO) based discrete cosine transform (DCT). Meanwhile, the risk prediction is carried out using Chaotic Giant Trevally Optimizer (CGTO) based Faster Region based Convolutional Neural Network (Faster-RCNN) approach. Experimental demonstration is carried out using Python and analyzed various performances and made an analogous study with state-of-art works. The statistical analysis shows that the proposed approach can be easily used for the prediction of diabetic disease from data.