A Regression-Based Temporal Pattern Mining Scheme for Data Streams

This chapter develops a regression-based algorithm, frequent temporal patterns of data streams (FTP-DS), to mine frequent temporal patterns for data streams. Algorithm FTP-DS has two major features, namely, one data scan for online statistics collection and regression-based compact pattern representation, which are designed to address the time and the space constraints respectively in a data stream environment. With these features, algorithm FTP-DS is able to not only conduct mining with variable time intervals but also perform trend detection effectively. The experimental results show, while, allowing one data scan to meet the time and space constraints in a data stream environment, FTP-DS is able to obtain the mining results of very good quality. Moreover, sensitivity analysis on the lift of support threshold has also been conducted to provide more insights into algorithm FTP-DS. The chapter also develops the techniques of the segmentation tuning and segment relaxation to enhance the functions of FTP-DS.