Modelling Wikipedia’s Information Quality using Informativeness, Reliability and Authority

Abstract

Wikipedia is the largest collaborative encyclopedia published on the internet. Due to its ‘open source' model, Wikipedia faces many issues regarding its Information Quality (IQ). Due to this reason, Wikipedia is generally not recommended for academic and research activities. However, hybrid approach which utilizes both content and metadata statistics of Wikipedia articles provide good insights in measuring the underlying IQ. Therefore, aligning with this hybrid approach, this study presents a simple yet precise model to assess the IQ of Wikipedia. The model comprises three IQ dimensions (1) Informativeness, (2) Reliability and (3) Authority, and 23 IQ features. The proposed model was tested with 1000 articles extracted from five WikiProjects Medicine, Politics, Sports, History, and Science. A Selenium-based web scraping technique was used to extract the data from articles automatically. The model received a classification accuracy of 79% and a clustering accuracy of 84%. Thus, this extensive experiment validates the effectiveness of the proposed model. Accordingly, the methodology, analysis and results, implications of the findings to theoretical discourse and practical applications, limitations, and futuristic directions are discussed in this paper.