Ontology In Association Rules Pre-Processing And Post-Processing

Data mining has emerged to address the problem of transforming data into useful knowledge. Although most data mining techniques, such as Association Rules, substantially reduce the search space, oftentimes one finds that the solution obtained surpasses the human ability to handle the resulting information. Furthermore, a good part of the information in repositories may be wrongfully dismissed due to the mining methods' inability to grasp the relationships between stored data from world knowledge that makes it possible to discover new valuable results, as well as eliminate irrelevant ones. This paper studies domain ontology as an instrument to enhance the mining results of Association Rules, which also acts to reduce the number of generated association rules. The adopted model is based on generalization and specialization processes in which the rules are filtered by metrics based on the coverage and confidence indicators. KEYWORDS Data Mining, Association Rules, Ontology, Preproce...

• Type: Conference
• Year: 2008
• Where: IADIS
• Authors: Inhaúma Neves Ferraz, Ana Cristina Bicharra Garcia
• Keywords: Association Rules, Data Mining, Data Mining Techniques, IADIS 2008, Internet Technology