An Efficient and Interactive Approach for Association Rules Generation by Integrating Ontology and Filtering Technique

Abstract

Association rule mining identifies the correlation among the set of items provided in the database. Although Apriori, frequent pattern mining, and other algorithms are proposed in the literature for association rule generation, these are statistical methods. In such cases, mining is completely uncontrolled because once data is supplied to algorithm; it produces results according to the predetermined methodology. Many times generated rules lack user’s expectations and hence need arises for methodologies with traditional algorithms. To overcome the aforesaid drawback, we propose the usage of ontology and filters along with frequent pattern tree mining algorithm for getting the desired results. Graphical structures are used for generation of ontologies. This paper thus proves and indicates the use of ontology and filters, and their proper implementations to obtain optimum and desired results through utilization of the above mentioned improved technique for data mining.