Automatic Content Quality Estimation Using Deep Neural Networks in Collaborative Encyclopedias on the Web

• Clara D. A. Marques CEFET-MG
• Nicolas G. Rezende CEFET-MG
• Daniel H. Dalip CEFET-MG
• Marcos A. Gonçalves CEFET-MG

Resumo

Wikipedia is based on user-generated content, in other words, anyone with internet access can write and make changes to its articles, hence the information quality of the encyclopedias is often criticized. Therefore, assigning the correct class of quality to Wikipedia articles is crucial for the experience of both authors and readers when using this large repository of information. In this paper, we present an approach relying on deep neural networks for this problem, our experiments consisted of testing three different models to achieve the best possible result. The first one using a conventional deep learning architecture and the other two using sets of semantically related quality indicators (aka, views) to better exploit their different properties, thus improving the final prediction. Finally, we compared our results with the state-of-the-art method (Support Vector Regression with views), achieving similar results with the possibility of improvement.

Palavras-chave: Quality Assessment, Wikipedia, Machine Learning, Neural Net- works