论文部分内容阅读
LEM2算法是一种有效的规则提取算法,但是效率有待提高.针对效率问题,文章对LEM2算法作了一些改进:在计算属性值对的同时获取单条件属性的确定规则,并且可以去掉某些冗余属性;定义了决策分配势矩阵作为启发信息,简化了LEM2算法内部的多重循环.最后给出了两种算法的复杂度分析和仿真对比试验,结果表明改进算法能获取更为简洁的规则,并能有效的提高算法的效率.
LEM2 algorithm is an effective rule extraction algorithm, but the efficiency needs to be improved.For the efficiency problem, the article makes some improvements to the LEM2 algorithm: obtaining certain rules of the one-condition attribute while calculating attribute value pairs, and eliminating some redundant I define the decision distribution matrix as the heuristic information, which simplifies the multiple loops inside the LEM2 algorithm.Finally, the complexity analysis and simulation comparison of the two algorithms are given.The results show that the improved algorithm can obtain more concise rules, And can effectively improve the efficiency of the algorithm.