Combining Multiple Features for Web Data Sources Clustering

The numbers of web data sources grow significantly, and as a sequence, crucial data management issues should be addressed. Clustering is one of the issues that many researchers have focused on. Clustering has been proposed to improve the information availability. To this end, in this paper, we propose a feature-based clustering approach for clustering web data sources without any human intervention and based only on features extracted from the source schemas. In particular, we combine linguistic and structure features of each data source to enhance computation of schema similarity. We experimentally demonstrate the effectiveness of the proposed approach in terms of both the clustering quality and runtime.