การสร้างออนโทโลยีแบบกึ่งอัตโนมัติเพื่อการจัดกลุ่มเอกสารโดยใช้อัลกอริทึมการจัดกลุ่มแบบลำดับชั้น

Abstract

Ontology design is complex and time-consuming tasks. It is difficult for domain experts to discover and design ontology from data or documents. This paper represents entity identification for ontology design using hierarchical clustering algorithm for reducing ontology design time. We used unstructured documents as dataset in longan pests domain. As results, this initial ontology expression provided the ontology designer as a guideline for classes and their hierarchical structure and they can further design the relationship between classes. A performance of this experiment gave the Purity is at 0.72, Rand index is 0.77, and F-measure is 0.68.