A Pattern Growth-Based Technique for Mining All Sequential Patterns from a Large Data Set in a Short Period of Time

Abstract

This paper presents a new technique for mining all sequential patterns from a large data set in a short period of time. The main idea of the technique is to use a pattern growth approach to store and update the discovered patterns in the data set. A novel pattern tree structure is proposed to store the patterns, and a set of efficient pruning techniques is developed to reduce the search space and time complexity. Experiments on real-world data sets demonstrate that the proposed technique can efficiently and effectively mine all sequential patterns in a large data set in a fraction of the time required by existing techniques.