Commodity label generation method and device

Abstract

The invention provides a commodity tag generation method and apparatus. The method comprises the steps that a server obtains text features of commodities according to text description information corresponding to the commodities in a commodity set, wherein the commodity set comprises at least two commodities; the server obtains adjacent objects of the commodities according to a behavior sequence of a user, wherein the adjacent objects include commodities, meeting a condition that a time interval of commodity occurrence is smaller than a preset time interval threshold, in the behavior sequenceof the user; the server determines an initial tag set of part of the commodities in the commodity set; and the server generates tags for the commodities in the commodity set by adopting a heterogenoustag topic model according to the text features, the adjacent objects and the initial tag set. According to the tag generation method provided by an embodiment of the invention, the user does not needto participate in commodity tag allocation, so that the cost is relatively low and the applicability is wider; and in addition, the text description information of the commodities and the adjacent objects of the commodities extracted from the sequence of the user are utilized at the same time, so that the accuracy is relatively high.

Description

technical field [0001] The present invention relates to the technical field of the Internet, in particular to a method and device for generating commodity labels. Background technique [0002] With the continuous development of Internet technology, more and more users conduct related activities through the Internet, such as browsing news, shopping and so on. In order to help users classify and organize products in more detail, and to facilitate users to quickly locate the products they are interested in, many Internet applications assign tags to products to refine the description and classification of products. Recommend corresponding products to users. Among them, the label refers to the phrase used to describe the relevant attributes of the product. [0003] At present, the method of extracting tags generally adopts the following two methods: manually assigning product tags, or the server extracts appropriate tags by analyzing the relevant description information of the ...

AI Technical Summary

Problems solved by technology

[0004] The above-mentioned first method, due to the endless emergence of new products on the Internet, only relies on a small amount of manpower to manually assign product labels, and there is a problem of low efficiency; the second method, only based on the relevant description information of the product, has the problem of low accuracy and incompleteness

Images