A Study of Data Mining Techniques for Diagnosis Diabetes

Diabetes Mellitus is a disease caused by metabolic disorders that is characterized by high blood sugar levels in the long term. Symptoms often include frequent urination, increased thirst and appetite. If left untreated, DM can cause many health problems. Data mining techniques can help diagnose DM early and prevent the progression of the disease as well as its complications such as cardiovascular disease, vision problems, leg ulcers, nerve damage and kidney disease. On the other hand, reviewing a large amount of this data requires the use of effective and efficient methods to find appropriate patterns of this data that the use of various techniques of machine learning and data mining, especially classification algorithms can be a significant help in this regard. In this review study, Google Scholar, Science Direct, and Scopus databases were reviewed with the aim of finding English articles published between ۲۰۰۵ and ۲۰۲۰. The search was based on relevant keywords and the obtained articles were evaluated based on the purpose of the study. The results of this study show that classification algorithms such as SVM, ANN, RF, DT, K-NN and NB are the most widely used machine learning and data mining techniques which they have been used to diagnose and predict the likelihood of developing DM. In fact, implementing a method that can correctly diagnose whether or not people have DM will be an important issue in controlling DM.

Keywords:

Data Mining , Diabetes Data Mining , Diabetes Mellitus , Machine Learning