Commodity attribute-based automatic classification method

Abstract

The present invention provides a commodity attribute-based automatic classification method. The method includes the following steps of: step S1, information acquisition; step S2, information processing; step S3, information analysis; and step S4, commodity classification. According to the commodity attribute-based automatic classification method, face image feature information of individuals is acquired; processing and classification are carried out according to corresponding face image feature information; and commodities are classified. Comparison with the prior art, the classification-based recommendation method makes the services of e-commerce enterprises more meticulous and thoughtful, makes marketing more targeted and can satisfy the one-stop shopping and information needs of a user.

Description

technical field [0001] The invention is an automatic classification method according to commodity attributes, belonging to the Internet field. Background technique [0002] Traditional brick-and-mortar stores (supermarkets) classify products according to product attributes, such as food, cosmetics, clothing, electrical appliances, etc., and each large category will also have small categories. The purpose of classification and placement is to facilitate users (customers) to find the products they need. When customers need several different products, they only need to choose to buy according to the order of classification. [0003] The product classification of existing e-commerce websites also follows the principle of physical stores. Amazon, the originator of e-commerce in the United States, and China’s JD.com, Dangdang, Yihaodian, etc. are all classified according to product attributes. Similarly, shopping on an e-commerce site is similar to visiting a brick-and-mortar sto...

AI Technical Summary

Problems solved by technology

However, most of the existing algorithms have several main problems: 1. Poor accuracy, because the less you know about user information (crowd attributes), the worse the recommendation accuracy will be; If you log in to an e-commerce website, if you cannot obtain relevant information, you cannot make recommendations for him; 3. The recommendation information lags behind, and a user who has already purchased a certain product still recommends it. After entering the website, relevant product recommendations are still made, but what users need now is follow-up services such as mobile phone instructions or repair point addresses

[0005] On the other hand, existing e-commerce websites are still unable to recommend information for users, especially personalized information.

Similarly, whether it is a comprehensive portal website or a vertical information website, users can only search for the information they need according to the standard classification labels of the website, and cannot provide recommendation services for one or a type of user group

Images