Avaliação da qualidade da Wikipédia enquanto fonte de informação em saúde

Abstract

The Wikipedia is an online, free, multi language and collaborative encyclopaedia, currently one of the biggest data sources in the web. There, it is possible to find information from different areas, from technology to philosophy, including health. As an health related data source, it is not only used by general public, but also by professionals as well. One of the reasons for such, is that, apart from the content of the articles, it includes external links for additional data sources as well. The open nature of Wikipedia contributions, specifically in health context, raises safety concerns, as such data is being used to make decisions. Thus, and considering possible consequences, it is very relevant to evaluate the quality of the information herein. This subject has been previously addressed by several studies, considering different metrics. In this work, a set of predefined metrics will be used to evaluate que quality of the information, such as autorithy, completeness, complexity, informativeness, consistency, currency and volatility. An additional set of first level measurements, and metrics based on them will be used, focusing in the health area. The definition of these metrics resulted from the analysis of other previously defined ones, which have already been applied to Wikipedia. This set of measures and metrics was then evaluated based on a proposed dataset, consisting of health and medical, english, articles, previously evaluated by the WikiProject Medicine. In the last stage, with the objective of evaluating differences in the quality of the information, the proposed methodology was applied to other languages. Specifically, it was applied, when possible, to articles available in languages with over than one hundred million native speakers, and also in Greek, Italian, Korean, Turkish, Perse and Hebrew, for its historical tradition. As a result, this work contributes to the clarification of the role of Wikipedia in the access to health, specifically the access to health information in different languages. The proposal consists of a set of measurements and metrics to infer the quality of Wikipedia health related articles, plus an analysis regarding the differences in the quality between different languages available in Wikipedia.