A method and system for optimizing e-commerce product search

Abstract

A system for optimizing e-commerce commodity search, including a front-end processing module, a back-end processing module, a commodity entity lexicon, an attribute search domain, an entity search domain, a scoring module, and a commodity information database, wherein the front-end processing module calls the commodity entity lexicon to Product search information for data processing; the back-end processing module invokes the commodity entity lexicon to process data on the existing commodity information in the e-commerce website, and after processing by the back-end processing module, an entity search based on all commodities in the e-commerce website will be formed domain and attribute search domain; the product information database: refers to the data source of e-commerce product search, used to store the product description data of existing products on the website, and the product description data is provided and maintained by the merchant; the scoring module is based on the front-end processing module and The information matching of the back-end processing module calculates the corresponding matching score; the hit score of the product search term in the search field is counted, and the search results are sorted and output.

Description

technical field [0001] The invention belongs to the field of e-commerce search, in particular to commodity search technology, in particular to a method and system for optimizing e-commerce commodity search. Background technique [0002] In the e-commerce environment, product search is a frequently used tool, through which product sellers promote their products to users, and users use it to quickly find the products they need. With the rapid growth of product types and quantities, in the face of countless and inaccurate product information, product search tools sometimes give unsatisfactory search results. For example, when a user enters two keywords, the product description that may be searched It does contain these two keywords, but the problem is that this product is not what users want. [0003] The realization principle of the traditional product search method is to completely match the search term with the product information. Through word segmentation, the descriptio...

AI Technical Summary

Problems solved by technology

[0006] (2) The quality of product information reduces the accuracy of search results

Product information on e-commerce websites is generally published and maintained by merchants. In order to make their products quickly searchable, they will add a lot of words that do not match or are not related to the product when describing the product. The quality of product information varies. The unevenness restricts the realization of accurate recall by traditional search methods, and it is easy to lead to "because the product information contains search terms, even unnecessary information will be recalled" or "because the product information does not contain search terms, it is obviously the information that users need Phenomena such as "cannot be recalled" occur frequently

Images