Problem classification method and system based on multi-attention mechanism and storage medium

Abstract

The embodiment of the invention provides a problem classification method and system based on a multi-attention mechanism and a storage medium, and belongs to the technical field of problem classification. The problem classification method comprises the steps of obtaining a to-be-classified text; converting the text into a corresponding word vector sequence by adopting a word2vec model; forming a word vector matrix based on a questionnaire attention mechanism according to the word vector sequence; performing part-of-speech tagging and encoding on the text by adopting a preset tagging set to form a part-of-speech vector sequence; respectively calculating the coefficient of each vector in the word vector sequence by adopting a formula (1); standardizing each coefficient by adopting a formula(2); determining a word vector matrix of a part-of-speech attention mechanism by adopting a formula (3); performing convolution operation on the two word vector matrixes to form a combined matrix; andinputting the combined matrix into LSTM to obtain a feature matrix with time sequence features, then obtaining feature vectors by using a self-attention mechanism, and determining the category of thetext according to the feature vectors.

Description

technical field [0001] The present invention relates to the technical field of problem classification, in particular to a problem classification method, system and storage medium based on a multi-attention mechanism. Background technique [0002] In recent years, with the wide-scale popularization of the Internet, more and more people participate in the network information interaction, which promotes the rapid development of the question answering system. As an extension of the field of information retrieval, the question answering system can provide a correct and concise answer to the natural language questions raised by users to meet the information needs of users. Question answering systems generally include three parts: question classification, information retrieval, and answer extraction. Among them, question classification is undoubtedly the basic task of the question answering system. Only by correctly analyzing and classifying questions can the candidate answer space...

AI Technical Summary

Problems solved by technology

[0006] Second: The existing deep learning model does not extract the potential subject information of the question text, but only uses the convolutional neural network or the long-term memory network to extract the text features, and does not further mine the potential subject information of the question text; and some methods are only single The use of convolutional neural networks or long-short-term memory networks does not combine the advantages of the two. Convolutional neural networks can capture deep semantic features extracted from data, and long-short-term memory networks can model the temporal characteristics of text. Preserve contextual semantic information of text

Images