一种改进的关联分类算法

Improved Algorithm of Classification Based on Association

关联分类算法是数据挖掘技术中一种主要分类方法,但传统关联分类算法仅根据置信度构造分类器,影响分类精度。提出一种改进算法,在选择高置信度构造分类器的基础上,优先考虑短规则分类。实验结果表明,该改进算法在分类精度和分类器大小上均优于传统分类算法。

Classification Based on Association（CBA） algorithm is one of the main methods in data mining. However, this algorithm only uses confidence parameter to process classification problem which may impact the classification accuracy. In this paper, a new improved method of CBA is presented. On the basis of CBA method, shorter rules are selected firstly for classification. Experimental results show this algorithm has higher classification accuracy and fewer numbers of classification rules than traditional ones.