Author(s)

Weiping Li, Tong Mo, Xingzhang Ren, Jingbo Zhang, Zhonghai Wu

School of Software and Microelectronics, Peking University, Beijing, China.

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.

Rough Set, Attribute Reduction, Elderly Evaluation, Decision Rule

Li, W. , Mo, T. , Ren, X. , Zhang, J. and Wu, Z. (2017) A Rough Set Based Optimization Method for Elderly Evaluation. International Journal of Intelligence Science, 7, 25-53. doi: 10.4236/ijis.2017.72003.