| 摘要: |
| 首次对时间有序事务数据中聚簇频繁模式的挖掘问题进行研究. 为了解决Naive算法处理该问题时存在冗余运算的问题, 提出一种改进的聚簇频繁模式挖掘算法ICFPM (improved cluster frequent pattern mining). 该算法使用2种优化策略, 一方面可以利用定义的参数minCF, 有效减少挖掘结果的搜索空间, 另一方面可以参考(n–1)项集的判别结果加速聚簇频繁n项集的判别过程, 算法还使用了ICFPM-list结构来减少候选n项集的构建开销. 基于两个真实世界数据集的仿真实验证明了ICFPM算法的有效性, 与Naive算法相比, ICFPM算法在时间和空间效率方面得到了大幅度的提高, 是解决聚簇频繁模式挖掘的有效方法. |
| 关键词: 时间有序事务数据 聚簇 频繁模式 数据挖掘 向下闭包 |
| DOI：10.13328/j.cnki.jos.007209 |
| 分类号:TP311 |
| 基金项目:国家自然科学基金(62066034, 62262047) |
|
| Clustering Frequent Pattern Mining for Time-ordered Transaction Data |
|
WANG Shao-Peng1,2,3, NIU Chao-Yu1
|
|
1.College of Computer Science, Inner Mongolia University, Hohhot 010021, China;2.Engineering Research Center of Ecological Big Data (Inner Mongolia University), Ministry of Education, Hohhot 010021, China;3.Inner Mongolia Engineering Laboratory for Cloud Computing and Service Software (Inner Mongolia University), Hohhot 010021, China
|
| Abstract: |
| In this study, the problem of mining cluster frequent patterns in time-ordered transaction data is discussed for the first time. To deal with redundant operations when the Naive algorithm solves this problem, the improved cluster frequent pattern mining (ICFPM) algorithm is proposed. The algorithm uses two optimization strategies. On the one hand, it can use the defined parameter minCF to effectively reduce the search space of mining results; on the other hand, it can refer to the discriminative results of (n–1)-itemsets to accelerate the discriminative process of cluster frequent n-itemset. The algorithm also applies the ICFPM-list structure to reduce the overhead of the candidate n-itemsets construction. Simulation experiments based on two real-world datasets demonstrate the effectiveness of the ICFPM algorithm. Compared with the Naive algorithm, the ICFPM algorithm improves substantially in terms of time and space efficiency, which makes it an effective method for solving clustered frequent pattern mining. |
| Key words: time-ordered transaction data clustering frequent pattern data mining downward closure |
