An Optimal Method for Result Cache Replacement

Abstract

The invention discloses an optimization method for result cache replacement and belongs to the technical field of computer networks. The method comprises the following steps: (1) taking a user ID (Identity) in a system searching log as a key and a request sent by a user as a key value to establish a hash table h2; (2) splicing all the requests sent by the same user ID to obtain a character string; carrying out a topic model analysis on a document set composed of the spliced character strings of all users to obtain a topic set M and the relevancy; (3) calculating appearing times of searching requests related to a certain topic and a specific value of the appearing times of all the requests in the same time period; calculating a weight contribution value of the request; (4) when obtaining one request q, finding a matched result cache and returning; updating the h2; if no matched result cache exists, finding the corresponding key value in the h2 to be spliced with the q; calculating the weight contribution value of the q; and (5) substituting a result item with a lower weight in the result cache and adding a new cache item. According to the optimization method for the result cache replacement, the efficiency of the cache hit ratio can be greatly improved.

Description

technical field [0001] The invention relates to a method for improving the hit rate of result cache in a search engine and a full-text retrieval system, specifically a method for improving the hit rate of result cache by using the result quality and semantic features of search requests, and belongs to the technical field of computer networks. Background technique [0002] Modern search engines have to process hundreds of millions to billions of search requests every day, so fast processing of search requests is a crucial issue. Result caching is a very important performance component in modern search engines. Therefore, how to effectively improve the efficiency of result caching is a problem that has attracted much attention in academia and industry. [0003] As a kind of cache, the result cache satisfies the basic assumption of the cache mechanism, that is, the principle of locality of search requests, that is, search requests that have appeared before may appear again in ...

AI Technical Summary

Problems solved by technology

[0006] These methods are very meaningful for improving the hit rate of the result cache. However, most of them only consider the characteristics of the search request itself, and do not consider the impact of the quality of the search results on the probability of the search request reappearing, nor the semantic characteristics of the search request. There is still a large distance between the hit rate of the result cache and the best hit rate that can be obtained

Images