Rule-based Approach for Automatic Ontology Population of Agriculture Domain

Abstract

Agricultural is one of the largest domain areas with rich diversified knowledge. However, most of the knowledge is hidden in unstructured textual documents such as books and web pages. Therefore, this knowledge need to be extracted and be presented so that can it be used in many applications. Ontology is one of the knowledge representation techniques that is seen suitable for modeling domain knowledge such as agriculture. Previous study has shown that automatic ontology population is possible, however, its application in the agriculture domain is yet to be explored. This study described the work on proposing a method to support ontology population for agricultural domain based on rules. A set of rules have been developed to automate the process of extracting the instances and their relationships between them. There are four steps involved in ontology population process: Document analysis, concept extraction using a Nearly-New Information Extraction System (ANNIE) processing resources, pattern analysis and lastly, the implementation of rule. Precision and recall have been used in this study. The study finding shows that the rule based approach is able to automate the process of ontology population in the agriculture domain. Key