What does the Elusion Rate statistic measure during project validation?

The Elusion Rate statistic specifically measures the percentage of documents that have been coded as relevant but are found in the uncoded, low-ranking portion of a sample. In the context of project validation, this means it assesses how effectively the active learning process is identifying relevant documents. A lower Elusion Rate indicates that the model is accurately capturing most of the relevant documents, while a higher rate suggests there are many relevant documents that remain uncaptured, indicating potential areas for improvement in the document review or active learning strategy.

In contrast, the other options provide different metrics related to document relevancy and positive findings, but they do not specifically focus on the evaluation of uncoded, low-ranking documents. For instance, the first option addresses the overall relevancy percentage, the second option pertains to active learning efficacy, and the fourth option highlights true positives among the found documents. None of these capture the specific situation of relevant documents that are not yet prioritized or coded, which is what the Elusion Rate illustrates clearly.