Constructing Predictive Model to Predict important Factors for Malnutrition in hemodialysis patients

• Main Authors: TSAI, YU-TSUNG, 蔡佑璁
• Other Authors: CHENG, BOR-WEN
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/00624420443437971504

碩士 === 國立雲林科技大學 === 工業工程與管理系 === 104 === The growth of hemodialysis population has become an issue of concern in many countries. In the past, Taiwan has been the top three country of highest incidence and prevalence of end-stage kidney disease in the world. Kidney disease has also become one of the top ten causes of death in Taiwan. According to the statistics of the National Health Insurance Administration, Ministry of Health and Welfare, there were around 70,000 dialysis patients in Taiwan up to 2015 and the number is increasing year by year. The main method of dialysis in Taiwan is hemodialysis. From the data of National Health Insurance Administration, Ministry of Health and Welfare, we found that the nutritious status of the hemodialysis patients in the case hospital in Yunlin was worse than those in the case hospital in Taipei. Therefore, we collected the database from TSN-KiDiT, and have the patients treated in the case hospital in Yunlin from 2010 to 2015 as samples in this study. The sample size was 300, with the patients' age above 20 years old. We used data mining and statistics to analyze the data of hemodialysis patients. We hoped to find the important factors causing malnutrition in hemodialysis patients. Based on this result, the best predicted model is the Support Vector Machine mode 4, accuracy of training set is 99.18% and the testing set is 73.01%. This result can be a reference for the physicians in clinical medicine.