Abstract
The incorporation of temporal semantics into traditional data mining techniques has led to the development of a new field called temporal data mining. This is especially necessary for extracting useful knowledge from dynamic domains, which by nature are time-varying. However, in practical terms, this is a computationally intractable problem, and therefore, it poses more challenges to efficient processing than non-temporal techniques. In this paper, we present a tree-based structure and a handling algorithm, called TSET-Miner, for frequent temporal pattern mining from time-stamped datasets. The algorithm is based on mining inter-transaction association, and is mainly characterized by the use of a single tree-based data structure for generation and storage of all frequent sequences discovered by mining. Given the versatility involved in the use of a single data structure, it may be extended an adapted to extract other types of patterns with relative little effort. To demonstrate this, we also present TSET^m^a^x-Miner, an algorithm based on the TSET structure, designed to extract maximal frequent event-based sequences.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call