Evaluating and Aggregating Data Believability across Quality Sub-Dimensions and Data Lineage

Abstract

Data quality is crucial for operational efficiency and sound decision making. This paper focuses on believability, a major aspect of data quality. The issue of believability is particularly relevant in the context of Web 2.0, where mashups facilitate the combination of data from different sources. Our approach for assessing data believability is based on provenance and lineage, i.e. the origin and subsequent processing history of data. We present the main concepts of our model for representing and storing data provenance, and an ontology of the sub-dimensions of data believability. We then use aggregation operators to compute believability across the sub-dimensions of data believability and the provenance of data. We illustrate our approach with a scenario based on Internet data. Our contribution lies in three main design artifacts (1) the provenance model (2) the ontology of believability subdimensions and (3) the method for computing and aggregating data believability. To our knowledge, this is the first work to operationalize provenance-based assessment of data believability.