A 2-population classifier system
Yi‐Chang Chen, Shin-Ren Shen, Shan-Lin Chang · 2010 Sixth International Conference on Natural Computation · 2010
This study proposes a 2-population classifier system to increase the computing efficiency of classifier system. The system is applied to solve the Wisconsin Breast Cancer (WBC) problem. The system is compared to the traditional learning classifier system (LCS) and Wilson's extend classifier system (XCS) in terms of computing efficiency and prediction accuracy. On the WBC problem, the average execution time for 2-population classifier system is roughly 19.45% of XCS. Meanwhile, the 2-population classifier system is higher than LCS even to XCS according to the accuracy rate comparisons. Thus, this study presents the 2-population classifier system to achieve both higher prediction accuracy ability and lower execution time.