Disambiguation of Text Field Data in a Database for Automatic Generation of Concept Hierarchy
Jaeki Song, Sangno Lee · 한국비즈니스리뷰 · 2010
In this paper, we propose the disambiguation of text field data in a database using semantic networks for automatic generation of concept hierarchy. One of critical obstacles for the automatic generation of concept hierarchy is polysemy of a word, or multiple meanings of a word. Extant research mainly generates a concept hierarchy for a text field with only one word in a database, and does not expand to multiple words in a field due to the challenge of polysemy. We suggest semantic networks method for disambiguating senses of multiple words in text field. The method first extracts the relations among concepts and constructs semantic networks from external knowledge resources. The resultant semantic networks, then, applied to disambiguate words in a text field for the generation of concept hierarchy. The proposed method is tested in a book title field of a database in a library. The automatic disambiguation of text field data turns out to save efforts of experts or designers who need to identify senses of words for making a concept hierarchy.