Mining Knowledge from Textual Databases: An Approach using Ontology-based Context Vectors

A. Gonçalves (Brazil, UK), V. Uren (UK), V. Kern, and R. Pacheco (Brazil)

Keywords

Data mining, ontologybased context vector, semantic similarity, clustering

Abstract

We propose a semantic mining approach to knowledge discovery based on context vectors and ontology. The approach is illustrated using ontology and resumes from a Science & Technology database as inputs, and the involved phases in the proposed model are described emphasizing preprocessing and pattern generation. The main contribution of this paper is the proposal of a semantic component toward data mining. Initial results show a suitable cluster generation in terms of number and quality. The approach produced better classification when comparing the generated clusters against a set of vectors representing knowledge areas, thus allowing for improved knowledge discovery.

Important Links:



Go Back