DIABETES MELLITUS PREDICTION SYSTEM USING DATA MINING

Now a days detection of patients with elevated risk of diabetes mellitus is developing critical to the improved prevention and overall health management of these patients. We aim to apply association rule mining to electronic medical records (EMR) to invent sets of risk factors and their corresponding subpopulations that represent patients which have high risk of developing diabetes. With the high dimensionality of EMRs, association rule mining generates a very large set of rules which we need to summarize for easy medical use. We reviewed four association rule set summarization techniques and conducted a comparative evaluation to provide guidance regarding their applicability, advantages and drawbacks. We proposed extensions to incorporate risk of diabetes into the process of finding an optimum summary. We evaluated these modified techniques on a real-world border line diabetes patient associate. We found that all four methods gives summaries that described subpopulations at high risk of diabetes with every method having its clear strength. In this extension to the BottomUp Summarization (BUS) [1] algorithm produced the most suitable summary. The subpopulations identified by this summary covered most high-risk patients, had low overlap and were at very high risk of diabetes.