Extracting Semantically Similar Frequent Patterns Using Ontologies

Many methods were proposed to generate a large number of association rules efficiently. These methods are dependent on non-semantic information such as support, confidence. Also work on pattern analysis has been focused on frequent patterns, sequential patterns, closed patterns. Identifying semantic information and extracting semantically similar frequent patterns helps to interpret the meanings of the pattern and to further explore them at different levels of abstraction. This paper makes a study on existing semantic similarity measures and proposes a new measure for calculating semantic similarity using domain dependent and domain independent ontologies. This paper also proposes an algorithm SSFPOA (Semantically Similar Frequent Patterns extraction using Ontology Algorithm) for extracting and clustering semantically similar frequent patterns. The case study which is illustrated in this paper shows that the algorithm can be used to produce association rules at high level of abstraction.