Assessing the quality of Thai Wikipedia articles using concept and statistical features

Abstract

The quality evaluation of Thai Wikipedia articles relies on user consideration. There are increasing numbers of articles every day therefore the automatic evaluation method is needed for user. Components of Wikipedia articles such as headers, pictures, references, and links are useful to indicate the quality of articles. However readers need complete content to cover all of concepts in that article. The concept features are investigated in this work. The aim of this research is to classify Thai Wikipedia articles into two classes namely high-quality and low-quality class. Three article domains (Biography, Animal, and Place) are testes with decision tree and Naïve Bayes. We found that Naïve Bayes gets high TP Rate compared to decision tree in every domain. Moreover, we found that the concept feature plays an important role in quality classification of Thai Wikipedia articles. © Springer International Publishing Switzerland 2014.