A chronic disease analysis system based on dirty data mining

Abstract

With the rapid progress in data mining techniques, more and more systems are facing to the analysis of chronic disease because of the convenience for doctors and patients. However, low-quality data seriously leads to low-quality analysis results which may cause one’s life lost. Even though many efforts have been made to enhance data quality, there always exists the data which we cannot get the exact value. Motivated by this, we develop a chronic disease analysis system adopting the mechanism that combines data cleaning and fault-tolerant data mining. In our system, we conduct a complete data mining of raw dirty data set and integrate the analysis of some kinds of chronic disease which is different to just analysis for a single disease. Moreover, our system also provides a platform for training and testing a new medical data set which is more convenient for users who do not know data mining well.