Search method and device for advertising system

Abstract

The invention provides search method and device for an advertising system. The search method includes: clustering key words in an advertising library to obtain key word clusters; searching received search strings in a search string analysis QA word list; if the search strings are absent in the QA word list, matching out key words corresponding to the search strings from the key word clusters, and pulling corresponding advertising information by the matched key words.

Description

technical field [0001] The invention relates to the technical field of search, in particular to a search method and device in an advertising system. Background technique [0002] For the convenience of understanding, the explanation of the following terms is first given: [0003] Search string (query): the search string submitted by the search engine user; [0004] Keyword (keyword): it can be called an auction word, and the same auction word can be purchased by multiple customers as different keywords; [0005] Advertisement: Unless otherwise specified, an advertisement refers to a promotional result containing advertising ideas and keywords that can be displayed on the page; [0006] Advertisement creativity: The actual display of "promotional results" according to the customer's willingness to launch and business rules, including the title of the ad, the first line of the ad description, the second line of the ad description, the target URL, the display URL and other in...

AI Technical Summary

Problems solved by technology

[0029] In the above method, the qa vocabulary excavated offline can theoretically account for 75% or more of qv, but in practical applications, the actual effect depends heavily on the quality of the excavated keywords associated with the query. As well as the real-time effectiveness of keywords, the statistical results of the matching system show that the actual qv ratio is far lower than the theoretical value

[0030] In addition, although the processing of the long-tail query in the above method can establish a relationship with the analyzed high-frequency query, it is difficult to guarantee the quality of the extended query, and often the associated query found will deviate in semantics, and then use The keywords found by the query whose semantics deviate also deviate, and finally match the advertisement with poor relevance

Images