원문정보
초록
영어
This paper proposes a design of a Patient-customized Healthcare System based on the Hadoop with Text Mining (PHSHT) for an efficient Disease Management and Prediction. The PHSHT consists of a Medical Data Collection Module (MDCM), a Text Mining based Hadoop Module (TMHM), a Disease Rules Creation Module (DRCM), and a Disease Management and Prediction Module (DMPM). The MDCM stores the big data like healthcare and medical information in the Hbase. The TMHM converts the unstructured data collected by MDCM to the structured data and stores the result in the Hbase. The DRCM generates disease rules by using the Conditional Probability Set Theory (CPST) proposed in this paper and stores them in the Hbase. The DMPM not only provides an efficient patient-customized medical service and information by comparing the disease rules generated by the DRCM with a patient’s information, a patient’s current status, and family history but also predicts a patient’s disease according to his health status. In particular, because the CPST generates the disease rules by complementing the problems of an existing C4.5 algorithm which does not consider the relation between attributes, this paper guarantees that it excels the C4.5 algorithm.
목차
1. Introduction
2. Related Works
2.1. Text Mining
2.2 Hadoop
2.3. Hbase
2.4. RST (Rough Set Theory)
2.5. A C4.5 algorithm
3. A PHSHT design
3.1 A design of Medical Data Collection Module (MDCM)
3.2 A design of Text Mining based Hadoop Module (TMHM)
3.3 A design of Disease Rule Creation Module (DRCM)
3.4 A design of Disease Management and Prediction Module (DMPM)
4. Analysis
4.1 The accuracy estimation for disease rules
4.2 The analysis of risk index
5. Conclusion
References