基于主成分分析的模糊频繁项集合挖掘方法

Mining method of fuzzy frequent item set based on principal component analysis

传统模糊频繁项集合挖掘方法的挖掘范围较大,且无法预处理模糊频繁项集合,导致存在运行内存过大、应用有效性低的问题。因此提出新的间隔约束条件下的模糊频繁项集合挖掘方法。方法预处理模糊频繁项集合,修补缺损数据,并引入主成分分析法完成数据的降维,并为数据添加间隔约束条件,缩小挖掘范围。利用蚁群算法获取最优爬行路径,挖掘模糊频繁项集合。实验结果表明,与传统方法相比,所提方法运行内存小,挖掘有效性更理想。

The traditional mining methods of fuzzy frequent item-sets have many defects, such as wide mining range, lack of preprocessing process of fuzzy frequent item-sets, which lead to the problems of too large running memory and low application effectiveness. Based on this, a novel mining method of fuzzy frequent item-sets under interval constraints is reported. Firstly, the set of fuzzy frequent terms was preprocessed in order to repair the defective data. Secondly, principal component analysis was used to reduce dimension data. Next, interval constraints were added to the data to narrow the mining scope. Finally, an ant colony algorithm was introduced to obtain the optimal crawling path and mine the fuzzy frequent item set. The experimental results show that compared with the traditional methods, the running memory and mining effectiveness of this method are better than those of the traditional methods.