Search recall method and device

Abstract

The invention discloses a search recall method and device, and the method improves the category word mining accuracy through processing of a to-be-recognized commodity title through a phrase segmentation model and a filtering strategy, greatly improves the mining efficiency through an automatic process, and also greatly reduces the labor cost.

Description

technical field [0001] This application relates to but not limited to computer technology, especially a search and recall method and device. Background technique [0002] The category word entity includes the name of the corresponding commodity entity described in the e-commerce title or query, such as: mp3player, cat eye, sport shoe, scarf, etc. [0003] The method for mining category word entities in the related art mainly relies on the attribute information actively filled in by the seller. On the one hand, these attribute values ​​are filled in by the seller actively, and the quality is uneven. This method of obtaining the category word entity by extracting the attribute value describing the product type from the attribute relies heavily on the information filled in by the seller, and the seller does not fill in it. The part that cannot be covered, resulting in incomplete and inaccurate category word entity mining; on the other hand, the extracted words need a lot of ma...

AI Technical Summary

Problems solved by technology

On the one hand, these attribute values ​​are filled in by the seller actively, and the quality is uneven. This method of obtaining the category word entity by extracting the attribute value describing the product type from the attribute relies heavily on the information filled in by the seller, and the seller does not fill in it. The part that cannot be covered, resulting in incomplete and inaccurate category word entity mining; on the other hand, the extracted words need a lot of manual review to get the really needed category word entities, which virtually increases the cost

Moreover, for e-commerce platforms serving Southeast Asian buyers and sellers, there are five main languages, including English, Thai, Indian, Vietnamese, and Philippine. Each language will have a corresponding search entry on the platform. When shopping, buyers usually give priority to using languages ​​they are familiar with, such as Thai, Indonesian, etc., to input search keywords. When searching and displaying, they need to return relevant products according to each language. That is to say, searching and recalling in a multilingual environment is a have certain problems

Images