Applying data mining techniques to predict five years survival rate for hemodialysis patients

• Main Authors: Yuan-Wei Hsu, 許原維
• Other Authors: Bor-Wen Cheng
• Format: Others
• Language: zh-TW
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/pq7qz5

碩士 === 國立雲林科技大學 === 工業工程與管理系 === 103 === The growth in the numbers receiving hemodialysis has become an issue of concern in many countries, Taiwan has the highest incidence and prevalence rates of end stage renal disease (ESRD) in the world. According to National Health Insurance statistics, there are about 70,000 dialysis patients in Taiwan and this number increases year-on-year. Hemodialysis is thus widely used in Taiwan, but five year survival rates of those receiving hemodialysis is only about fifty percent. In the light of these statistics, this study uses data mining techniques such as decision tree, neural network and logistic regression, to identify the significant factors affecting the survival of dialysis patients. The results show that using a neural network model has the highest accuracy in prediction at 81.48%. Using a logistic regression model has the worst accuracy, at only 68.57%. Based on this result, the AI classification technology can serve as an important and useful references in diagnosis for physicians to enhance hemodialysis effectiveness. Keywords: Data mining, Neural network, Logistic regression, Hemodialysis, Survival rate