Research On Optimization Of Max-Score Query Processing Technology

The network resources keep growing explosively with the rapid development of the Internet. Facing the request of instant responding to high-concurrent queries from the massive data, how to improve its efficiency has become an essential problem to search engine. O ne possible way to relieve this tension is by optimizing the query performance of the stand-alone computer to enhance the efficiency of the whole system.In the beginning, some basic theories about query processing on inverted index are discussed, which include the structure of inverted index, ways to process query and dynamic index pruning.DAAT Max-Score is one of the classic algorithms in Top-k query processing. In order to solve the slow start threshold problem existed in Max-Score, an improved algorithm based on query terms partition and two-tiered index. With query terms partition, DAAT Max-Score according to the characteristics of the query words users submited, will choose candidate document and boost the starting threshold by using t he way TAAT algorithms quickly processing short query collections.using two-tiered index, DAAT Max-Score can sharply decrease the global max score of query terms in the lower index when constructing it, meanwhile the upper index can benefit the procedure of choosing candidate document and boosting start threshold. Both improved algorithm can effectively decrease the probability of non Top-k documents being chosen as candidate documents, improve the performance of query processing;In the end, on the basis of the query words divided mechanism and two-tiered index, integrate the two kinds of improved algorithm.