Mining Top-K Sequential Rules

Mining sequential rules requires specifying parameters that are often difficult to set (the minimal confidence and minimal support). Depending on the choice of these parameters, current algorithms can become very slow and generate an extremely large amount of results or generate too few results, omitting valuable information. This is a serious problem because in practice users have limited resources for analyzing the results and thus are often only interested in discovering a certain amount of results, and fine-tuning the parameters can be very time-consuming. In this paper, we address this problem by proposing TopSeqRules, an efficient algorithm for mining the top-k sequential rules from sequence databases, where k is the number of sequential rules to be found and is set by the user. Experimental results on real-life datasets show that the algorithm has excellent performance and scalability.