COMPARISON OF DATA MINING TECHNIQUES FOR PREDICTING DIABETES OR PREDIABETES BY RISK FACTORS

Data mining, extraction and analysis techniques play an important role in health and care. Because the analysis and diagnosis of any disease must contain a huge number of data. The important role of extracting and analyzing patterns of medical diagnosis is therefore evident. Diabetes is a group of metabolic diseases that contain a high percentage of blood sugar for a long time. In addition to the challenges of classify and forecasting Diabetes, there is another problem that health data may contain data loss or be incorrect. Because of these problems and circumstances that may hinder the process of processing and overcoming data, many previous studies have provided many automated learning methods for diagnosis, prediction, processing of potential data loss and problem solving. This research will analyze and compare different techniques of data extraction and analysis of diabetes. Recent data mining techniques commonly used in Bayesian, SVM, Decision Tree, etc. This paper represents the proposed framework with hybrid datamining techniques. Results showed that hybrid classification in proposed framework outperforms other classifiers with an accuracy rate of 94%.