Recall at k measures the fraction of all known relevant documents that appear among the first k retrieved results. Mean Reciprocal Rank (MRR) averages the reciprocal rank of the first relevant result for each query. Normalized Discounted Cumulative Gain (nDCG) rewards relevant results near the top of the ranking and can support graded relevance. Precision at k measures the fraction of the first k retrieved results that are relevant. Evaluation metrics depend on the quality and completeness of relevance judgments.
