Using web data provenance for quality assessment

TLDR

The Web of Data cannot be a trustworthy source unless an approach for evaluating its quality is established and integrated into the data publication and access process. The paper proposes using provenance information to assess Web data quality and trustworthiness, and suggests associating certainty values with calculated quality values to address missing provenance. The authors present a provenance model and an adaptable assessment method, demonstrating its use for evaluating Web data timeliness and proposing certainty-value association to handle missing provenance.

Abstract

The Web of Data cannot be a trustworthy data source unless an approach for evaluating the quality of data on the Web is established and integrated as part of the data publication and access process. In this paper, we propose an approach of using provenance information about the data on theWeb to assess their quality and trustworthiness. Our contributions include a model for Web data provenance and an assessment method that can be adapted for specific quality criteria. We demonstrate how this method can be used to evaluate the timeliness of data on the Web, to reflect how up-to-date the data is. We also propose a possible solution to deal with missing provenance information by associating certainty values with calculated quality values.