Abstract:Sequential pattern mining is an important data mining problem with broad applications. However, it is also a difficult problem since the mining may have to generate or examine a combinatorially explosive number of intermediate subsequences. Most of the previously developed sequential pattern mining methods, such as GSP, explore a candidate generation-and-test approach [1] to reduce the number of candidates to be examined. However, this approach may not be efficient in mining large sequence databases having numerous patterns and/or long patterns. In this paper, we propose a projection-based, sequential pattern-growth approach for efficient mining of sequential patterns. In this approach, a sequence database is recursively projected into a set of smaller projected databases, and sequential patterns are grown in each projected database by exploring only locally frequent fragments. Based on an initial study of the pattern growth-based sequential pattern mining, FreeSpan [8], we propose a more efficient method, called PSP, which offers ordered growth and reduced projected databases. To further improve the performance, a pseudoprojection technique is developed in PrefixSpan. A comprehensive performance study shows that PrefixSpan, in most cases, outperforms the a priori-based algorithm GSP, FreeSpan, and SPADE [29] ( a sequential pattern mining algorithm that adopts vertical data format), and PrefixSpan integrated with pseudoprojection is the fastest among all the tested algorithms. Furthermore, this mining methodology can be extended to mining sequential patterns with user-specified constraints. The high promise of the pattern-growth approach may lead to its further extension toward efficient mining of other kinds of frequent patterns, such as frequent substructures.

Mining Sequential Patterns with Constraints in Large Databases

Constraint-based Sequential Pattern Mining: the Pattern-Growth Methods

Constraint-based sequential pattern mining

Prefix-Projection Global Constraint for Sequential Pattern Mining

Efficiently Mining Closed Subsequences with Gap Constraints

Mining Sequential Patterns by Pattern-Growth: the PrefixSpan Approach.

An Efficient Algorithm for Mining Frequent Sequence with Constraint Programming

Constraint-Based Graph Mining in Large Database

Self-adaptive nonoverlapping sequential pattern mining

Constraint-based sequence mining using constraint programming

On Top-K Closed Sequential Patterns Mining

Mining Sequential Patterns in Uncertain Databases Using Hierarchical Index Structure

Imcs: Incremental Mining Of Closed Sequential Patterns

Mining Frequent Induced Subtree Patterns with Subtree-Constraint

Efficient Mining of Gap-Constrained Subsequences and Its Various Applications

Finding Sequential Patterns from Large Sequence Data

Intelligent Sequential Mining Via Alignment: Optimization Techniques For Very Large Db

New Approach for the Sequential Pattern Mining of High-Dimensional Sequence Databases

Accelerated Frequent Closed Sequential Pattern Mining for Uncertain Data

Sequential Pattern Mining in Multi-Databases Via Multiple Alignment

Parallel Sequential Pattern Mining by Transaction Decomposition