Diabetes Prediction Algorithm Using Deep Learning

Abstract

Diabetes has become a common disease in mankind of the young and the old person. According to the national survey 2017-18 on diabetes, overall, around 37.5 million people above 20 are suffering from this disease in the country.[1] Every 4th Pakistani is suffering from diabetes.[2] So, it is badly needed to develop a system that can effectively predict the diabetes using medical details. We can help reduce risk of Diabetes by giving suggestions to make changes in your lifestyle. Common risk factors include increased weight, blood pressure, cholesterol and triglyceride (blood fat) levels. Changing the habits of a lifetime isn’t easy, but it’s worth the effort. We propose a strategy for the prediction of diabetes using deep learning by training its attributes. We propose a strategy for the prediction of diabetes using deep learning by training its attributes. The idea of deep learning is fast-growing and it works quite like a human mind. It represents the data in multiple levels and able to solve the selectivity-invariance dilemma efficiently [3]. Deep learning techniques are used in a variety of forms in field of medical prognosis. Many research works prove that deep learning techniques provide a better outcome, declines the classification error rate and more robust to noise than other strategies. It can handle massive amount of data and have capability to decode a complex problem in an easy way. We are using Recurrent Neural Network (RNN). This system explores approaches to improve the accuracy using medical data with Deep learning algorithms