id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
work_zco22hvrvzfdjivaorp37e7ije	Andreas Broschart	Efficient query processing and index tuning using proximity scores	2012.0	259	.pdf	application/pdf	95622	11259	77	of term pair index lists based on information from a query log. We propose to extend the index with additional term pair lists that maintain proximity scores. We allow a selective materialization of term pair index lists based on information from a query log To reduce computation costs, de Kretser and Moffat restrict the evaluation of aggregated proximity scores to locations where query terms appear. all pairs of query terms are considered if a document is to be scored, they generate an Figure 7.1: Score-ordered term, proximity, and combined index lists which can be used We then evaluated retrieval quality with pruned (term and combined) index lists, if we use only the BM25 scores from the pruned CLs. To this end, we devise another list structure called CTL (combined term index into term pair lists with document-position related scores although the index may take 7.1 Score-ordered term, proximity, and combined index lists which can be	./cache/work_zco22hvrvzfdjivaorp37e7ije.pdf	./txt/work_zco22hvrvzfdjivaorp37e7ije.txt