标题:
阈值自适应关联分类方法研究Research on Associative Classification Method Based on Threshold Adaptive Technology
作者:
汤伟, 刘敬学
关键字:
自适应, 关联分类, 规则提取, 类型判别Self-Adaptive; Associative Classification; Rule Extraction; Class Being Distinguished
期刊名称:
《Hans Journal of Data Mining》, Vol.2 No.4, 2012-11-01
摘要:
本文采用评估反馈控制机制，对关联分类方法的阈值设定自适应调节进行了研究。首先建立了对分类规则集进行判优评估的数学模型；然后基于此模型提出了运用区间迭代逼近求解关联分类最小支持度，并以此进行阈值自适应调节的方法；最后的仿真实验表明，该方法比一般的关联分类方法具有更好的自适应性和有效性。
Based on the mechanism of evaluation feedback and control, this paper researches for adaptive adjustment of setting threshold of associative classification method. Firstly, the mathematical model of evaluating the set that consist of classification rules is established. Then, based on this model, a method that uses the interval iterative approximation for getting the approximate optimal solution of associative classification minimum support in order to achieve the adaptive threshold settings is put forward. Finally, the simulation experiment shows that the method has better adaptability and effectiveness comparing with some general ones.