Method and device for identifying industry types of objects

Abstract

The invention discloses a method and device for identifying industry types of objects. The method comprises the steps of: inputting text information of to-be-identified objects into a language model for generating a paragraph vector to carry out learning so as to obtain vector spaces of the to-be-identified objects, which are related to industry types; according to the vector space of each to-be-identified object, selecting a first to-be-identified object from all the to-be-identified objects to as a training sample object, and acquiring annotation data of the training sample data; carrying out training on a constructed industry type identification model by utilizing the vector space and the annotation data of the training sample object so as to obtain a target industry type identificationmodel; and for each second to-be-identified object except for the training sample object, inputting the vector space of the second to-be-identified object into the target industry type identificationmodel to carry out learning so as to obtain an industry type to which the second to-be-identified object belongs. According to the method, accuracy of an identification result of the industry type identification model can be promoted.

Description

technical field [0001] The present invention relates to the technical field of the Internet, in particular to a method and device for identifying an industry type of an object. Background technique [0002] With the continuous development of Internet technology and the popularization of terminal equipment, users and enterprises have more and more information data in various dimensions, and the continuous advancement of technology has made it possible to calculate these information data, and the analysis of users and enterprises And portraits, marketing and recommendations based on this technology are becoming more and more personalized and fine-grained. In this increasingly personalized and fine-grained application scenario, industry information is a crucial link. Users' industry interests are the basis of these personalized data, and enterprises are classified by industry type. , can assist major network platforms to mine potential customers. [0003] In the prior art, by...

AI Technical Summary

Problems solved by technology

[0004] In this way, the recognition result of the industry type recognition model is limited by the degree of coverage of all industry words by the industry word set mined from the text information with industry labels. If the text information of the enterprise to be classified contains more When there are few or even industry words that are not included in the industry word set, the accuracy of the recognition results of the industry type recognition model is low

Images