Generating new knowledge from scientific databases, fusioning products information of business companies or computing an overlap between various data collections are a few examples of applications that require data integration. A crucial step during this integration process is the discovery of correspondences between the data sources, and the evaluation of their quality. For this purpose, the overall metric has been designed to compute the post-match effort, but it suffers from major drawbacks. Thus, we present in this paper two related metrics to compute this effort. The former is called post-match effort, i.e., the amount of work that the user must provide to correct the correspondences that have been discovered by the tool. The latter enables the measurement of human-spared resources, i.e., the rate of automation that has been gained by using a matching tool.
展开▼
