一种高效的频集挖掘算法

频集挖掘是关联规则挖掘的关键步骤,它对强规则、相关分析和时间序列有着重要的意义.常用的频集算法包括Apriori和FP-G rowth.为了提高算法效率,提出了一种基于D iffset的混合算法———D iffsetHybrid,该算法根据数据集的稀疏程度决定采用D iffset的某种形式来挖掘频集,减少了存储空间,提高了算法效率.试验表明,该算法对于稀疏数据集和稠密数据集都有良好的计算性能.

An efficient frequent itemset mining algorithm

Frequent itemset mining is a critical step of association rule mining and it plays an important role in many data mining tasks including strong rules, correlations and sequential rules. The ordinary frequent itemset algorithms include Apriori and FP -Growth. In order to improve the efficiency of algorithm, this article advanced a hybrid algorithm based on Diffset DiffsetHybrid. Depending on the degree of sparsity of dataset, this algorithm adopts different ways of Diffset to mine frequent itemsets, in this way, the storage space can be decreased and algorithm efficiency can be improved. The tests indicate that the new algorithm show good performance with both sparse datasets and dense datasets.