Search

CN-116628556-B - Product label establishing method and system

CN116628556BCN 116628556 BCN116628556 BCN 116628556BCN-116628556-B

Abstract

A method for building up product label includes setting up corresponding relation between standard name of product and attribute of product, collecting data about product and carrying out text feature extraction on data to retain characteristic key words of data, utilizing characteristic key words to match standard name for realizing preliminary classification of product data, carrying out characteristic classification on characteristic key words divided into same class to obtain different characteristic features, supplementing characteristic features containing quantity of characteristic key words exceeding set threshold value into product attribute so as to build up new product attribute or perfecting existing product attribute to form product label. Therefore, the product attribute is updated according to the product data, so that the label is formed with a daily progression, and the label can be conveniently purchased and referenced by a consumer when aiming at the specific screening requirement of the consumer, and is also convenient to use as a recommendation algorithm.

Inventors

  • LI LIJUN
  • SUN SU
  • REN YUEFENG

Assignees

  • 上海桥创科技有限公司

Dates

Publication Date
20260505
Application Date
20230614

Claims (7)

  1. 1. A method of creating a product label comprising the steps of: s1, establishing a corresponding relation between a standard name of a product and a product attribute; The product attributes comprise category attributes, parameter attributes and performance attributes, wherein the category attributes refer to categories or types to which products belong, the parameter attributes refer to specific parameters of the products, and the performance attributes refer to performance performances of the products; s2, collecting data about the product, extracting text features of the data to preserve characteristic keywords of the data, and matching the characteristic keywords with standard names to realize preliminary classification of the product data; The characteristic keywords comprise name type characteristic keywords capable of knowing the names of the products and attribute type characteristic keywords capable of knowing the attributes of the products, and the name type characteristic keywords are matched with standard names to classify the product data into the standard names matched with the name type characteristic keywords; S3, classifying the characteristics of the characteristic keywords classified into the same class to obtain different characteristic features, and supplementing the characteristic features with the quantity of the characteristic keywords exceeding a set threshold value into the product attributes, thereby establishing new product attributes or perfecting the existing product attributes to form product tags; The characteristic keywords are processed by a characteristic embedding method or a text characteristic processing method and then subjected to similarity matching, classification or clustering, so that characteristic classification is realized to obtain different characteristic features; S4, selecting a plurality of characteristic features with the top ranking of the number of the characteristic keywords as core features, converting the core features into word vectors, and then carrying out nearest neighbor search; s5, dividing the product data into standard names with the core characteristics most similar to the characteristic keywords thereof through nearest neighbor search so as to reclassify the product data; s6, repeating the steps S4 and S5 until the classification is not changed any more.
  2. 2. The method according to claim 1, further comprising establishing a name correspondence between the standard name and the alias by integrating the preliminary names, performing dimension reduction processing on the information capable of obtaining the product name to obtain a name characteristic, matching the name characteristic with the standard name to perfect the name correspondence, and setting up a new standard name correspondingly if there is no standard name capable of matching with the name.
  3. 3. The method for creating a product label according to claim 1, wherein the name characteristics and the standard names are matched by using a similarity algorithm, a machine learning model or a manual auditing method.
  4. 4. A building system for implementing the method for building a product label according to any one of claims 1 to 3, comprising a corresponding module, a processing module, a first matching module and a second matching module, wherein the corresponding module is used for building a corresponding relation between a standard name of a product and a product attribute, the processing module is used for extracting characteristic keywords from data about the product, matching the characteristic keywords with the standard name to implement preliminary classification of the product data, classifying the characteristic keywords in the same class to obtain different characteristic features, and the first matching module supplements the characteristic features containing the characteristic keywords with the number exceeding a set threshold value into the product attribute, thereby building new product attribute or perfecting the existing product attribute; And the second matching module selects a plurality of characteristic features with the top ranking of the quantity of the contained characteristic keywords as core features, and divides the product data into standard names with the core features closest to the characteristic keywords through nearest neighbor search.
  5. 5. The system for creating a product label according to claim 4, further comprising a standard module, wherein the standard module is configured to create a name correspondence between a standard name and an alias, and to perform dimension reduction processing on information capable of obtaining a product name to obtain a name characteristic, match the name characteristic with the standard name to perfect the name correspondence, and set up a new standard name correspondingly if there is no standard name capable of matching with the name.
  6. 6. A computer readable storage medium having stored thereon computer instructions, which when executed by a processor, implement a method of creating a product label according to any of claims 1-3.
  7. 7. A terminal device comprising a memory and a processor, characterized in that the processor is configured to implement the method of creating a product label according to any of claims 1-3 when executed based on instructions stored in the memory.

Description

Product label establishing method and system Technical Field The invention relates to the technical field of information processing, in particular to a method and a system for establishing labels for various products. Background Along with the rapid development of the Internet, various industries select to conduct electronic commerce transactions through a network and rely on the network to establish communication for various product information, wherein the communication comprises evaluation after selling products and discussion, evaluation and the like of linked or unlinked products, the evaluation, the discussion, the evaluation and the like can be used as data of related products, and tag information in aspects of products, performance and the like of corresponding products can be obtained through processing and integrating the data, and the tag information has great significance for purchasing of consumers or updating and improving products of merchants. In addition, a recommendation algorithm may be performed by using the tag information, or a user representation may be performed on a user using such a product. However, at present, a manual mode is basically adopted for processing and integrating product data, a mode of checking after-sales evaluation by after-sales customer service is mainly adopted in the manual mode, so that the product type or performance label corresponding to the sold product is obtained. Based on this, there is a need to provide a method that facilitates the processing of mass product data to create accurate product labels. Disclosure of Invention The invention aims to provide a method and a system for establishing a product label, wherein characteristic keywords in product data are extracted and processed to obtain characteristic features, the characteristic features with the quantity of the characteristic keywords exceeding a set threshold value are added into product attributes, so that new product attributes are established or existing product attributes are perfected, consistency of product attributes and product data updating is ensured, relevance and accuracy of product standard names and product attributes are ensured, and the product label which is in progress at the moment is established. In order to achieve the above purpose, the present invention adopts the following technical scheme. A method for establishing a product label comprises the following steps: s1, establishing a corresponding relation between a standard name of a product and a product attribute; s2, collecting data about the product, extracting text features of the data to preserve characteristic keywords of the data, and matching the characteristic keywords with standard names to realize preliminary classification of the product data; S3, classifying the characteristics of the characteristic keywords classified into the same class to obtain different characteristic features, and supplementing the characteristic features with the quantity of the characteristic keywords exceeding a set threshold value into the product attributes, thereby establishing new product attributes or perfecting the existing product attributes to form product tags; therefore, the product attribute is updated according to the product data, so that the label is formed with a progressive property, and the label can be conveniently selected and referred by consumers and used as a recommendation algorithm when specific screening requirements of the consumers are met; further, in the step S1, the method further comprises the steps of establishing a name corresponding relation between the standard name and the alias through integration, performing dimension reduction processing on the information capable of knowing the product name to obtain a name characteristic, matching the name characteristic with the standard name to perfect the name corresponding relation, and correspondingly establishing a new standard name if the standard name capable of being matched with the name characteristic does not exist; further, matching the name characteristics with the standard names by adopting a similarity algorithm, a machine learning model or a manual auditing mode; Further, in step S1, the product attributes include a category attribute, a parameter attribute and a performance attribute, where the category attribute refers to a category or a kind to which the product belongs, the parameter attribute refers to a specific parameter of the product, and the performance attribute refers to a performance of the product; further, in step S2, the characteristic keywords include a name type characteristic keyword capable of obtaining a product name and an attribute type characteristic keyword capable of obtaining a product attribute, and the product data is classified into a standard name matched with the name type characteristic keyword by matching the name type characteristic keyword with the standard name; Further, in step S3, the feature keywords are