Optimizing query evaluation in n-gram indexing

In n-gram indexing widely used in Asian languages, because a long query word is divided into more than one n-gram, proximity constraints among ngrams need to be checked. And this checking degrades the retrieval speed. This paper proposes two optimizations of query evaluation: one selects a combination of n-grams whose processing cost is the minimum among all the possible combinations, and the other uses additional n-grams in finding potential documents whose proximity constraints need to be checked.