Search term recommendation method and device

Abstract

The invention relates to a search term recommendation method and device. The search term recommendation method comprises the steps that candidate search terms are determined according to user input terms; a prediction model of a first evaluation value of the search terms used in search behaviors is established according to the history search behavior data of a user, and the first evaluation value represents the trade information of commodities led by the search terms within a set time; the first evaluation value of the candidate search terms is predicated according to the prediction model and the user history search behavior data corresponding to the candidate search terms; according to the first evaluation value, the candidate search terms are filtered or ranked, and the candidate search terms after filtering or ranking are recommended to the user. Accordingly, the search term recommendation method and device achieve the purpose that the commodities corresponding to the recommended search terms can better clinch a deal, and increase cheating cost.

Description

technical field [0001] The present application relates to the field of electronic commerce, in particular to a method and device for recommending query words. Background technique [0002] In order to reduce the user's search path in the e-commerce website, during the user's query word input process and after the search is over, recommend possible candidate query words (Query) for the user to help the user clarify the search intention and realize the purchase behavior. Among them, the function of recommending candidate query words to the user during the process of user inputting query words is called the search drop-down prompt, and the function of recommending candidate query words to the user after the search is completed is called related search. [0003] The search drop-down prompt is the first service that users come into contact with when they log in to the search page. It can recommend relevant query words for users when they input search keywords. A common search dr...

AI Technical Summary

Problems solved by technology

[0008] The disadvantage of this method is that it pays too much attention to the similarity between the query words entered by the user and the recalled query words, resulting in the result of the recall being too similar to the query words entered by the current user; at the same time, the PV characteristics of the recalled words are relatively obvious, which can be artificially increased. The number of clicks on the search result pages of certain recalled query words is used to improve the ranking of recalled query words, and the cost of cheating is relatively low, making the recommendation results less accurate

Images