Researcher Automates Wikipedia Article Quality Review

UvA researcher Xinyi Li has developed an automatic method to assess the quality of Wikipedia pages. Because Wikipedia runs on contributions from people with different opinions, it is sometimes difficult to determine whether the content is correct.

Li recently presented the model of the research group of which he is a part at the ‘European Conference on Information Retrieval’ in Vienna. One of the problems Wikipedia encounters is that of all Wikipedia articles, only a small part can be manually graded for quality. An automatic evaluation method is needed to be able to value all pages.

In addition to the content, the method Li presented uses information about who contributed to an article. Most articles are written by more than one person, but not everyone contributes equally. It also appears that the majority of authors only write about a limited number of subjects. In order to achieve the best results, information from existing manual article evaluations and Li’s software are combined, which the authors believe is a useful solution for article quality assessment.

By looking not only at the content of an article, but also at the expertise and the number of contributions of authors, Li’s method makes it easier to assess the quality of an article. In this way, automated warnings could be given. Now there are occasional warnings on articles if references are missing, if there is little text or if the writing style is not objective.