Publications

Abstract : The advancement in technology is a root cause for the significant decrease in physical activities. Moreover, changes in food habits and increases in mental stress cause most common health disorders such as diabetes and high blood pressure. A statistical report on diabetes, available on the World Health Organization’s website, shows an exponential increase in the past few years. Across the world, there is a significant increase in the number of diabetic patients, from 108 million in 1980 to 422 million in 2014. The International Diabetes Federation claims that 425 million people currently suffer from diabetes. As diabetes adversely affects people from all age groups, it has become a cause of concern to predict this disorder well in advance. This motivated the authors to focus on diabetes prediction. In this chapter, the authors present the state of art in the arena of diabetes prediction. They identify challenges in existing techniques, namely Naïve Bayes, decision tree, and support vector machine, and they propose effective solutions for these. The proposed model performs an analysis of publicly available data gathered from diabetic patients listing the causing factors of diabetes, most affected age groups, work style, and food habits. The model applies artificial neural networks for detecting diabetes and identifying its type. It is efficient in predicting the survival rate of diabetic patients. The prediction proves useful in preventing other health disorders such as retinopathy, nephropathy, and cardiovascular disorders that may arise due to diabetes. For evaluation of the proposed model, the authors use the “Pima Indian Diabetes” dataset. The dataset includes the medical history of 768 patients. It considers nine different symptoms as parameters for occurrence of diabetes. The highest accuracy of 85.09% proves the efficacy of the proposed work.