A method and device for generating a store category recognition model and identifying a store category

Abstract

The present application discloses a store category recognition model generation method and a store category recognition method and device, wherein, the store category recognition model generation method includes: obtaining sample product information of a sample store and a multi-level model corresponding to the sample product information Business category; determine multiple identification dimensions for store category identification, at least one branch node corresponding to each identification dimension, and category identification constraint information corresponding to each identification dimension; based on sample product information and multi-level information corresponding to sample product information Business categories, generate target identification information of sample stores under each identification dimension; based on the impact factors of each identification dimension and at least one corresponding branch node, construct a preset tree structure corresponding to store category identification; based on target identification information The store category recognition training is carried out on the preset tree structure with the category recognition constraint information, and the store category recognition model is obtained. The application can improve the accuracy of store category identification and reduce the risk of manual operation errors.

Description

technical field [0001] The present application relates to the field of data processing, in particular to a method and device for generating a shop category recognition model and shop category recognition. Background technique [0002] The traditional store category identification method manually marks the store category, relies on the business personnel to manually click on the store link, and subjectively defines the store category according to the number of commodities and commodity categories under the store. [0003] For the traditional method of manually labeling store categories, due to the large amount of product data and the complex and varied product level classification system under the store, manually clicking on the store link not only requires a lot of manpower, but also leads to inefficient category annotation. Artificial subjectivity is prone to produce wrong labeling results. Based on the low efficiency and high risk shown by the method of manually labeling ...

AI Technical Summary

Problems solved by technology

[0003] For the traditional method of manually labeling store categories, due to the large amount of product data on sale and the complex and varied product level classification system under the store, manually clicking on the store link not only requires a lot of manpower, but also leads to inefficient category annotation. Artificial subjectivity, easy to produce wrong labeling results

Images