Mining multi-level time-interval sequential patterns in sequence databases

Mining sequential patterns is an important issue in data mining and has many applications. An extended work of sequential pattern mining, called time-interval sequential pattern mining, is proposed to retrieve time-interval information between successive items. However, previous work only considers single-level time-interval in pattern extraction, which means sequential patterns with cross-level time-intervals are completely ignored. Therefore, this study first defines multi-level time-interval sequential patterns and then presents a novel algorithm, named MLTI-PrefixSpan, for discovering the complete set of multi-level time-interval sequential patterns. Experimental results show that the proposed algorithm is effective on the test dataset.