A Proposed Method to Identify the Occurrence of Diabetes in Human Body Using Data Analysis

Abstract

Advanced machine-learning techniques are often used for reasoning-based diagnosis and advanced prediction system within the healthcare industry. The methods and algorithms are based on the historical clinical data and fact-based medicare evaluation. Diabetes is a global problem. Each year people are developing diabetes and due to diabetes, a lot of people are going for organ amputation. According to the World Health Organization (WHO), there is a sharp rise in number of people developing diabetes. In 1980, it was estimated that 180 million people with diabetes worldwide. This number has risen from 108 million to 422 million in 2014. WHO also reported that 1.6 million deaths in 2016 due to diabetes. Diabetes occurs due to insufficient production of insulin from pancreas. Several research show that unhealthy diet, smoking, less exercise, Body Mass Index (BMI) are the primary cause of diabetes. This paper shows the use of machine learning that can identify a patient of being diabetic or non-diabetic based on previous clinical data. In this article, a method is shown to analyze and compare the relationship between different clinical parameters such as age, BMI, Diet-chart, systolic Blood Pressure etc. After evaluating all the factors this research work successfully combined all the related factors in a single mathematical equation which is very effective to analyze the risk percentage and risk evaluation based on given input parameters by the participants or users.