Quality assessment of Arabic web content: the case of Arabic Wikipedia

Abstract

With the huge size and large diversity of Arabic web content, machine assessment of document quality acquires added importance. Users are in dire need for quality rating of the material returned in response to their queries. The Wikipedia, with its large metadata, has been a topic of extensive research on document quality assessment. Criteria used include text properties and style parameters, contributor and edit characteristics and multimedia components. In this paper we report on our ongoing work to adapt existing document assessment approaches to Arabic content with concentration on the Arabic Wikipedia and present some of the results. We also try to augment that with features specific to Arabic as well as parameters like author expertise and social media presence. One of our goals is an aggregate measure integrating many of the features into a single document quality index. We plan to use Wikipedia article quality assessment results to train general content assessment methods that can be applied to general content that lacks major Wikipedia features.