Development of Score Based Smart Risk Prediction Tool for Detection of Type-1 Diabetes: A Bioinformatics and Machine Learning Approach

Abstract

In this study, a smart risk prediction tool has been demonstrated along with the algorithm, which works as a backend of the tool to detect Type-1 Diabetes. The algorithm was contrived by the weightage values that are articulated by analyzing the risk factors of Type-1 diabetes. The analysis takes place with a machine learning and statistical approach. Data were collected from a number of cases and control groups, which was preprocessed to be fit for the analysis. Risk factors were extracted by comparing two different approaches one is machine learning, and another is the statistical approach. A common regulatory pattern was found that leads to the design of an algorithm that gives a predictive result of the risk level of any user for Type-1 Diabetes. Elaborated results of different approaches have also been shown in this paper, which gives clear excogitation about risk factors and their ranking.