An Ontology-Based Approach to Extracting Semantic Relations from Descriptive Text.

Da Huang,Wei Hu

DOI: https://doi.org/10.1007/978-3-642-54025-7_3

2013-01-01

Abstract:Linked Data have advantages over plain text, as data are organized in relations between information, which is convenient for learning and reasoning. However, most plain text with valuable information has not been converted into Linked Data form. Thus, we propose an ontology-based method to extract semantic relations from descriptive text about entities. Moreover, we conduct our experiment on the DBpedia dataset and design an automatic methodology to evaluate our ontology-based method as well as an intuitive method. As a result, we find out that our ontology-based method performs better than the intuitive one in general. At last, we analyze the results, and put forward our opinions on the difference between the two methods’ performance.