Pattern Taxonomy Model (PTM) is a technique which uses Data Mining and Pattern Evolving approaches to discover useful representatives or patterns for addressing the problems encountered in the Knowledge Discovery field. However, these patterns are defined in the system only based on the statistical properties of pattern (e.g. frequency and probability). The conceptual information between patterns is unfortunately ignored at all. Therefore, the main purpose of this project is to construct semantic patterns using the information from the hierarchy of Lexical Ontology and then effectively apply them to the PTM system. This project will integrate several disciplines such as Ontology, Natural Language Processing, Data Mining, Information Retrieval, Information Filtering and Web Mining to construct our system. Furthermore, this project aims to develop an effective and efficient model, called the Ontology-based PTM model and build an OPTM-based Web Information Filtering system. The proposed model will be evaluated by conducting the real Web Mining tasks and the final experimental results will also be compared to other existing methods.
