Hello all, I want to do some kind of evaluation of my Lucene application, but the problem is that I'm indexing XML documents with a specific DTD and I have just a small collection. Is there a way to run a quality benchmark using the benchmark package in Lucene with my own collection an queries? I have seen too something about mean average precision, but I can't find any reading about how to implement it.
My XML documents have elements like: title, body, div (for sections), references, keywords, figures... and those are the fields I indexed in Lucene.
