Discovery of Generalized Local Sequential Patterns
序列中的一般化局部序列模式发现

Previous work on pattern discovery in sequence data mainly considers finding global patterns, whereevery record in the temporal sequence contributes to support the patterns. However, local patterns, which arefrequent only in some time periods, are actually very common in practice and the efficient discovery of it ispotentially very useful. This paper presents a method for discovering generalized local sequential patterns with thestructure that supports efficiently locating and counting of the pattern instances and a two-phase method forefficiently mining of local patterns. Experimental results corresponded with the problem definition and verified thesuperiority of the approach.