The Applications for Diabetes Prediction by Machine Learning Algorithms

Since diabetes is a very widespread disease causing serious symptoms, accurate diagnosis becomes more and more important as the very first step of effective treatments dealing with diabetes. To relief doctors from their burden of diagnosis workloads, which require them to make medical estimates based on experience, computers and algorithms contribute a lot in the field of medical diagnosis with the development of new technology nowadays. However, due to the diversity and complexity of real-world data, usual statistical methods are often unable to handle or produce precise results. In this paper, we remove most of the outliers of Pima Indians Diabetes dataset for the subsequent classification through complex data pre-processing and data cleaning. The final experimental results show that SVM and Random Forest algorithm perform best, but XGboost algorithm also performs well