Pushing Constraints to Generate Top-K Closed Sequential Graph Patterns

Abstract

In this paper, the problem of finding sequential patterns from graph databases is investigated. Two serious issues dealt in this paper are efficiency and effectiveness of mining algorithm. A huge volume of sequential patterns has been generated out of which most of them are uninteresting. The users have to go through a large number of patterns to find interesting results. In order to improve the efficiency and effectiveness of the mining process, constraints are more essential. Constraint-based mining is used in many fields of data mining such as frequent pattern mining, sequential pattern mining, and subgraph mining. A novel algorithm called CSGP (Constraint-based Sequential Graph Pattern mining) is proposed for mining interesting sequential patterns from graph databases. CSGP algorithm is revised to mine topk closed patterns and named as TCSGP (Top-k Closed constraint-based Sequential Graph Pattern mining). General Terms Data mining, Graph mining, Constraint-based mining.