Bootstrapping Method for Automatically Acquiring Instances and Properties in the Web Resources
Yusik Ri, Kim, Hyonchol, Song-il Cha · HAL (Le Centre pour la Communication Scientifique Directe) · 2018
With the rapid increase of Web resources, the Ontology has become an important research topic in the fields of Web information management and knowledge engineering. Nonetheless, ontology creation remains a daunting task. Bootstrapping method, which aims at automatically generating instances and their relations in a given domain, is a promising technique for ontology creation. In this paper we focus on an ontology instance and property generating method by using bootstrapping from the Web table resources. Most Web tables are designed for intuitive perception of human, thus, it has a certain limit to automatically interpret the table using only structural information. So, this work observes on the method for interpreting the table structure based on proper noun that is semantic core element of the table. This work propose algorithms to automatically acquire all instances or properties belonging to a given class, taking few terms belonging to class composed of proper noun as the seed. Empirical study of our novel approach has proved the effectiveness of ontology creation. Experimental results show that our approach can effectively acquire more accurate instances and properties from Web resources.