What is the Elusion Rate (Range)?

The Elusion Rate (Range) refers specifically to the percentage of previously uncoded documents that, when sampled from below the cutoff during Project Validation, are subsequently coded as positive. This metric is essential for assessing the performance of a predictive coding model, as it indicates how many relevant documents were inadvertently left uncoded.

When documents are reviewed and coded, there’s always a possibility of missing out on positive instances that fall beneath a certain cutoff threshold. By measuring the Elusion Rate, practitioners can gain insights into the efficacy and completeness of their document review process and understand the model's ability to find relevant information that may not have been specific enough to surpass the established cutoff during earlier coding phases.

Other options, while relevant to aspects of data and model performance, do not specifically define the Elusion Rate (Range). The success rate of the active learning process, the total documents reviewed, and the error rate of the active learning model pertain to different facets of model evaluation and performance metrics but do not capture the particular definition of Elusion Rate accurately. Hence, option B accurately aligns with the defined context of this metric.