Comparison of Accuracy in Evaluation of Obesity Utilizing Fuzzy Rule-based Systems and Multinomial Logistic Regression of Workers in Mae Moh Power Plant, Lampang Province
Wansiri Khooriphan, Watha Minsan, Suree Chooprateep, Pimpaka Thaninpong · Burapha Science Journal (วารสารวิทยาศาสตร์บูรพา ) · 2018
This study aimed to compare the accuracy in evaluation of obesity between utilizing Fuzzy Rule-based Systems and Multinomial Logistic Regression of workers in Mae Moh Power Plant, Lampang Province. The sample group consisted of 237 power plant workers in Mae Moh who attend the annual health check at Mae Moh Division of Medical and Health. The data were analyzed by using Factor Analysis and then using Fuzzy Rule-based Systems and Multinomial Logistic Regression. The accuracy in evaluation of obesity were checked. The results of Multinomial Logistic Regression analysis can be evaluated at 51%. The accuracy in group of obese, normal and overweight are 82 %, 52 % and 0 %, respectively. Fuzzy Rule-based Systems has 227 rule and can be evaluated at 78%. The accuracy in group of obese, normal and overweight are 100%, 72% and 48%, respectively. The results of evaluation of obesity utilizing Fuzzy Rule-based Systems are more accurate. Keywords : obesity, Fuzzy Logic, Fuzzy Rule-based Systems, Multinomial Logistic Regression