Given a search engine benchmark (such as Medline dataset containing binary relevance judgements), explain clearly why you would prefer to use Mean Average Precision metric over Precision-At-Fixed-Recall metric to evaluate a ranked retrieval system based on Vector Space Model? The table below shows the output of an IR system on two queries. Only top 5 ranks are shown. Crosses correspond to documents which have been judged relevant by a human judge; circles correspond to irrelevant documents. There are no relevant documents in lower ranks (> 5). Compute the MAP. Rank Q1 Q2 1 O X 2 X O 3 X O 4 X O 5 O X Is MAP appropriate to evaluate a search engine that uses LSI? Justify. Is MAP appropriate to evaluate a search engine that uses graded/multi-level relevance measure in place of boolean/binary relevance measure? Justify.

Can u please by step by step in detail?