Emotion classification method and device for short texts in tax field

Abstract

The invention provides an emotion classification method and device for short texts in the tax field, which are used for solving the technical problems that an intelligent response robot and an automatic voice recognition system in the tax field in the prior art cannot understand human language emotion and lack intelligent interaction capacity, so that questions raised by taxpayers are hard and unfriendly to answer, and the user satisfaction is not high. The method comprises : preprocessing tax short texts, training word vectors of the tax short texts through a Word2Vec model, and obtaining atleast one word vector; determining a weight coefficient of each word vector based on the emotional color carried by each word vector in the at least one word vector, and constructing a two-dimensionalmatrix vector of the word vector with the weight coefficient; and processing the two-dimensional matrix vector through a convolutional neural network to obtain an emotion classification result of thetax short text.

Description

technical field [0001] The invention relates to the technical field of information processing, in particular to a method and device for sentiment classification of short texts in the tax field. Background technique [0002] The rapid development of artificial intelligence technology has achieved fruitful results in the fields of finance and education, and the tax field is no exception. Under the clarion call of tax modernization, artificial intelligence is in the ascendant in tax service, tax risk management, tax inspection and other work, and has made rapid progress, effectively improving the work efficiency of tax departments. For example, in the tax service hall and the 12366 tax service hotline, a large number of taxpayers consult tax issues every day. In order to better meet the needs of taxpayers, the tax department has organized and developed an intelligent answering robot for the tax service hall and a 12366 tax service hotline. The automatic speech recognition syst...

AI Technical Summary

Problems solved by technology

[0004] The embodiment of the present invention provides a method and device for sentiment classification of short texts in the tax field, which are used to solve the problem that the intelligent answering robots and automatic speech recognition systems in the tax field of the prior art cannot understand human language emotions and lack intelligent interaction capabilities, resulting in taxpayers The answers to the questions raised are blunt and unfriendly, resulting in technical problems with low user satisfaction

Images