Domain information classification model generation method, classification method, equipment and storage medium

Abstract

The invention provides a domain information classification method. Steps comprise: preprocessing domain instruction information input by a user, to obtain a domain instruction text; using a domain word segmentation dictionary to segment words of the domain instruction text, to obtain a domain word segmentation set; performing specific word replacement traversal on the domain word segmentation set,to obtain an instruction array, specific words being representative words of a domain; using a word vector model to covert the instruction array to a word vector array; and using a domain classification model to classify the word vector arrays. The invention relates to a domain information classification model generation method. The invention also relates to an electronic device and a readable storage medium, used to execute the domain information classification model generation method and the domain information classification method. A word2vec word vector model is trained through large amount of domain related text corpus data, and more semantic relations can be expanded under the condition of insufficiency of semantic information. A deep learning algorithm training classifier is used,and classification accuracy of the classification model is improved.

Description

technical field [0001] The invention relates to the technical field of text classification, in particular to a method for generating a domain information classification model, a classification method, a device and a storage medium. Background technique [0002] With the development and application of artificial intelligence technology, the research on deep semantic analysis urgently needs to be improved, and the deep semantic analysis of instruction information is particularly urgent. Instructional messages refer to short sentences such as commands, suggestions, and requests, such as "please open the playlist for me", "tell me today's temperature". Since each domain has its own instructions, only by accurately classifying the domains of the instructions, can the key attributes in the instructions be extracted and the intention of the instruction sender be correctly analyzed, so as to execute the instructions accurately. [0003] Limited by the characteristics of instruction...

AI Technical Summary

Problems solved by technology

First of all, the length of instruction information is generally about twelve words, usually does not follow strict grammatical rules, lacks enough information for analysis and inference, and it is difficult for machines to make accurate judgments in a limited context; at the same time, traditional text classification technology , it is difficult to capture the information characteristics of the instruction

Images