Text multi-label classification method based on semantic unit information

Abstract

The invention discloses a text multi-label classification method based on semantic unit information, which comprises the following steps: establishing a semantic unit multi-label classification modelSU4MLC, taking a recurrent neural network sequence based on an attention mechanism to a sequence model as a baseline model for improvement, and improving the expression of the attention mechanism by improving a source end; Extracting semantic unit related information from the context representation of the source end of the baseline model by using hole convolution in deep learning to obtain semantic unit information; Combining the semantic unit information with the word level information by using a multi-layer mixed attention mechanism, and providing the combined information for a decoder; Anddecoding the tag sequence by using a decoder, thereby realizing text multi-tag classification based on semantic unit information. According to the method, the problems that an existing attention mechanism is easily influenced by noise and contributes to classification insufficiently can be solved, the contribution of the attention mechanism to text classification can be improved, and the text multi-label classification problem can be more efficiently solved.

Description

technical field [0001] The invention relates to natural language processing technology, in particular to a text multi-label classification method based on semantic unit information. Background technique [0002] Text multi-label classification technology is a natural language processing technology, mainly for labeling the input text, which is equivalent to dividing the text into multiple label categories. Work in this field has strong application value, such as labeling and classifying news texts in the news field, or classifying user information to construct user portraits. [0003] In the past, the work of multi-label text was generally regarded as a multi-classification problem. The traditional method used the combination of linguistic knowledge and statistical methods to achieve classification. After the rise of machine learning, many methods based on machine learning algorithms, For example, rank-SVM based on SVM and ML-KNN based on KNN have made great progress compare...

AI Technical Summary

Problems solved by technology

[0005] In order to overcome the deficiencies of the above-mentioned prior art, the present invention provides a text multi-label classification method based on semantic unit information, and establishes a semantic unit multi-label classification model (Semantic Unit for Multi-label Text Classification, the model is referred to as SU4MLC), which will be based on attention The sequence-to-sequence model of the mechanism is improved as a baseline model, and the content that the attention mechanism focuses on is improved, that is, the representation of the source end for the attention mechanism is improved to improve the contribution of the attention mechanism to text classification, so as to solve the current problems based on the attention mechanism. The attention mechanism of the sequence-to-sequence model in the text multi-label classification problem is easily affected by noise and does not contribute enough to the classification. Improve the performance of the sequence-to-sequence model based on the attention mechanism on the text multi-label classification problem

Images