期刊名称：International Journal of Intelligence Science
印刷版ISSN：2163-0283
电子版ISSN：2163-0356
出版年度：2017
卷号：07
期号：02
页码：25-53
DOI：10.4236/ijis.2017.72003
语种：English
出版社：Scientific Research Publishing
摘要：In order to improve the efficiency of elderly evaluation, an optimization method based on rough set is proposed. Compared with the traditional rough set attribute reduction, the redundant evaluation items are eliminated by items’ correlation. It avoids a big overhead of calculating the core of rough sets that have many attributes. A novel rule reduction method is proposed based on reliability and coverage, in order to solve the problem of rarely appeared rules and conflict rules in traditional rough set. A sorting algorithm based on coverage is used to optimize the traditional flat evaluation questionnaire model with a hierarchical order. By these optimizations, the number of items that need to evaluate is greatly reduced. The proposed approach is deployed in an elderly service company named Lime family. Real-life result shows that the method can reduce more than 40% items with over 90% accuracy prediction rate. Compared with decision tree and the method based on expert knowledge in reduction rate and accuracy rate, the method has same performance in one index, and 20% improvement on average in the other one.
