On multiview-based meta-learning for automatic quality assessment of wiki articles

Abstract

The Internet has seen a surge of new types of repositories with free access and collaborative open edition. However, this large amount of information, made available democratically and virtually without any control, raises questions about its quality. In this work, we investigate the use of meta-learning techniques to combine sets of semantically related quality indicators (aka, views) in order to automatically assess the quality of wiki articles. The idea is inspired on the combination of multiple (quality) experts. We perform a thorough analysis of the proposed multiview-based meta-learning approach in 3 collections. In our experiments, meta-learning was able to improve the performance of a state-of-the-art method in all tested datasets, with gains of up to 27% in quality assessment. © 2012 Springer-Verlag.