中文摘要：

考虑到2个或多个数据集的显露序列对学习/知识迁移有用,提出一种特殊的显露序列模式,即共享显露序列模式（SESs）,并给出一个基于共享广义后缀树的框架来挖掘共享显露序列模式,同时在挖掘SESs的过程中应用2种新的剪枝策略。从3个方面进行实验评估：SESs挖掘算法的性能分析,SESs的负迁移分析,以及SESs用于提高协同分类准确性分析。研究结果表明：新提出的SESs在时间性能、负迁移影响、提高协同分类准确性上均取得较好的性能。

英文摘要：

A particular type of emerging sequences called shared emerging sequences （SESs） was introduced. A shared generalized suffix-tree based framework for mining SESs was employed, and two new pruning techniques were utilized in the SESs mining algorithm proposed in this paper. Experiments were conducted to evaluate the mining algorithm, to analyze the negative learning transfer of SESs, and to evaluate the usefulness of SESs for improving performance of co-classification. The results show that our proposed SESs have good advantages on time performance, negative learning transfer, and co-classification performance.