Addressing class query word mining method and system

Abstract

The invention provides an addressing class query word mining method and system. The method comprises that: primary domain URLs with the same domain names recorded in a click log of a user are normalized so that corresponding primary domain names are generated, and a query word set of the primary domain names is generated according to corresponding query words of the primary domain URLs; word segmentation is performed on the query words in the query word set, occurrence frequency of the obtained word segments is counted, and the longest word segment in the word segments with the highest occurrence frequency is confirmed to the corresponding core word of the primary domain names; and the query word set comprises the core word through confirmation and the query word with the highest query frequency is a corresponding addressing class query word of the primary domain names. According to the technical scheme provided by the invention, an addressing class query word set can be automatically mined and generated, and addressing class Bad Case mining recall rate is enhanced.

Description

【Technical field】 [0001] The invention relates to the search technology in the Internet field, in particular to a mining method and system for addressing query words. 【Background technique】 [0002] Search engine query words can be divided into addressing query words, information query words and transaction query words. According to Andrei Broder's research, the proportions of these three query words are: 12.3%, 62% and 25.7%. Addressing query words refer to the query words that users provide when they need to query the address of a certain website, for example, Taobao.com, Mogujie, Ping An official website of China, etc. The user's search needs for such query words are very clear, that is, they hope to find The corresponding official website address, so the search engine needs to put the corresponding official website address in the front position of the search results, such as the first three. However, in actual situations, the following official website addresses may no...

AI Technical Summary

Problems solved by technology

However, in actual situations, the following official website addresses may not be ranked first and cannot meet the search needs of users:

[0003] 1. The official website address does not appear in the first place in the search results;

[0004] 2. The official website address does not appear on the home page;

[0005] 3. The official website address is not included by search engines;

[0006] 4. A false official website address appears;

The disadvantages of this mining method are: the set of addressing query words needs to be manually configured, which is a waste of manpower and has low efficiency. The result cannot meet the needs of users, requires frequent operations by users, and the search efficiency is low, which also increases the burden on search engines

Images