基于BIDE的多核并行闭合序列模式挖掘

摘要

基于经典的BIDE算法,提出一种多核并行闭合序列模式挖掘算法—MT_BIDE.该算法在频繁序列扩展判断前进行剪枝,在扩展过程中动态调整频繁序列及其伪投影数据集,平衡不同线程间挖掘闭台序列模式的计算量差异.实验结果表明,该算法具有较高的运行效率和加速比.%Based on the classical BIDE algorithm, this paper presents a multi-core parallel closed sequential patterns mining parallel algorithm, MTBIDE. Through pruning before frequent sequential patterns expansion and reassigning them and their pseudo-projected datasets during the expansion process, MTBIDE achieves the workload balancing which is always influenced by different calculation of different threads mining closed sequential patterns. Experimental results show that the algorithm has higher operating efficiency and speedup ratio.