A method suitable for product labeling in e-commerce Chinese websites

Abstract

A method suitable for labeling products on e-commerce Chinese websites, the steps of which include a method for constructing a word segmentation lexicon, a method for collecting labels, and a method for labeling commodities; the so-called method for building a word segmentation lexicon refers to Count the frequency of each product keyword in different product descriptions, retain product keywords with a frequency greater than 3, and filter out keywords with a product keyword number of less than or equal to 5 as thesaurus data; the so-called tag collection method refers to The word segmentation lexicon, through the reverse maximum matching word segmentation algorithm to segment all commodity names in the e-commerce Chinese website; after the word segmentation processing of the maximum reverse matching algorithm, select the last word formed by the commodity after word segmentation processing as the commodity of this commodity Labels; in the end, all these labels form a label data set; the method of labeling products refers to finding the relationship between product attributes and labels by using text mining algorithms.

Description

technical field [0001] The invention belongs to the field of computer internet, and in particular relates to a method suitable for labeling commodities on e-commerce Chinese websites. Background technique [0002] On e-commerce Chinese websites, when users use keywords to search for products, they usually directly retrieve the basic information of the products. information, but still can not avoid the emergence of two types of problems: one is the problem of product information cheating, in order to provide the exposure rate and frequency of appearance of their products in the product search process, the merchants make the released products eye-catching, so that product buyers More products can be searched, and they abuse brand names or keywords that are not related to the product when describing the product, so that product buyers cannot accurately find the product they need; the second is that the product information is not comprehensive The problem is that merchants omit...

AI Technical Summary

Problems solved by technology

[0002] On e-commerce Chinese websites, when users use keywords to search for products, they usually directly retrieve the basic information of the products. information, but still can not avoid the emergence of two types of problems: one is the problem of product information cheating, in order to provide the exposure rate and frequency of appearance of their products in the product search process, the merchants make the released products eye-catching, so that product buyers More products can be searched, and they abuse brand names or keywords that are not related to the product when describing the product, so that product buyers cannot accurately find the product they need; the second is that the product information is not comprehensive The problem is that merchants omit key information about product descriptions when describing products, including missing important information such as product titles, pictures, and descriptions. The lack of information will cause the website to fail to return more relevant product search results when users search for products.

[0003] For the problem of merchants cheating on product information, e-commerce websites usually set up rules to solve the problem, and lower the rights of those cheating products that do not meet the rules. However, there are certain defects in the rules, and strict rules may lead to the reduction of rights of non-cheating products; Loose rules may make the anti-cheating effect not obvious enough; in order to solve the problem of incomplete information filled by merchants, in order to ensure that as many related products as possible are recalled, e-commerce websites choose to expand the retrieval range of commodity information at the expense of retrieval quality, That is, multiple product information fields are matched, and sometimes even fields with a large amount of data but poor quality such as "product description" are selected. Although more products can be recalled in this way, the recalled products cannot be recalled. Satisfied users, resulting in significant traffic loss

Images