Incremental Mining of Frequent Serial Episodes Considering Multiple Occurrences

TL;DR

This paper introduces a new method for mining frequent serial episodes in data streams that considers multiple occurrences and series of itemsets, enhancing pattern recognition capabilities beyond existing presence-based approaches.

Contribution

It proposes a novel sequential pattern mining approach that accounts for multiple occurrences and series of itemsets, with efficient pruning strategies.

Findings

Effective in real and synthetic data

Improves pattern recognition over traditional methods

Reduces search space efficiently

Abstract

The need to analyze information from streams arises in a variety of applications. One of its fundamental research directions is to mine sequential patterns over data streams. Current studies mine series of items based on the presence of the pattern in transactions but pay no attention to the series of itemsets and their multiple occurrences. The pattern over a window of itemsets stream and their multiple occurrences, however, provides additional capability to recognize the essential characteristics of the patterns and the inter-relationships among them that are unidentifiable by the existing presence-based studies. In this paper, we study such a new sequential pattern mining problem and propose a corresponding sequential miner with novel strategies to prune the search space efficiently. Experiments on both real and synthetic data show the utility of our approach.