Abstract
Over recent years, the Internet has become people's main source of information, with many databases and web pages being added and accessed every day. This continued growth in the amount of information available has led to frustration and difficulty for those attempting to find a specific piece of information. As such, many techniques are widely used to retrieve useful information and to mine valuable data; indeed, these techniques make it possible to discover hidden relations and patterns. Most of the above-mentioned techniques have been used primarily to process and analyse English text, but not Arabic text. Limited Arabic resources (e.g. datasets, databases, and ontologies), also make analysing and processing Arabic text a difficult task. As such, in this paper, we propose a framework for building an Arabic ontology from multiple resources. Thus, we will first extract and build an Arabic ontology from a publicly available directory, following which, we will enhance this ontology with rich data from the Internet. We will then use an Arabic online directory to construct a multi-disciplinary ontology that provides a hierarchical representation of topics in a conceptual way. Following this, we introduce an enhanced technique to enrich these ontologies with sufficient information and proper annotation for each concept. Finally, by using common information retrieval evaluation techniques, we confirm the viability of the proposed approach.