A Method for Automatically Extracting Domain Semantic Networks from Wikipedia

摘要

This paper describes a method for automatically extracting domain semantic networks of concepts connected by non-specific relations from Wikipedia. We propose an approach based on category and link structure analysis. The method consists of two main tasks: concepts extraction and relations acquisition. For each task we developed two different implementation strategies. Aiming to identify what strategies have the best performances we conducted different extractions for two domains and we analyze their results. From this evaluation we discuss the best approach to implement the extraction method.