SURVEY OF MACHINE LEARNING ALGORITHMS FOR DIABETES DETECTION

Diabetes can be simply defined as the disorder/ state in human body where body cannot effectively use insulin produced . It is always beneficial to recognize diabetes as soon as possible as it reduces the risk of ailments such as heart related diseases, kidney related problems, eye diseases and even brain stroke, etc. By year 2025 there will be about 299.1 millions of diabetics.Developing countries like India will contain about 80% of the diabetics. With the development of technology, it is now possible to predict such diseases. In our research, we have studied the use of machine learning technologies to predict diabetes. From India related to this topic there are approximately3068 papers. Diabetes prediction comes under classification and supervised machine learning techniques are used. Different models used for this purpose are hierarchical clustering, SVM, Naive Bayes, K Means, etc. Different models have different accuracies in the range 70% to 99%.