Chinese aspect level sentiment classification method based on pre-training sentiment embedding

A sentiment classification and pre-training technology, applied in neural learning methods, semantic analysis, character and pattern recognition, etc., can solve the problems of few level sentiment classification methods, research, no patient review methods, etc., to improve medical quality and doctor services. horizontal effect

Pending Publication Date: 2022-02-18
DALIAN UNIVERSITY
View PDF0 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, there are few facet-level sentiment classification methods for Chinese, and there is no related method research for patient reviews.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Chinese aspect level sentiment classification method based on pre-training sentiment embedding
  • Chinese aspect level sentiment classification method based on pre-training sentiment embedding
  • Chinese aspect level sentiment classification method based on pre-training sentiment embedding

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0052] The present invention extracts patients' comments on doctors from the Haodofu online website, and builds a Chinese patient comment data set according to the screening; furthermore, it provides a deep learning model based on pre-trained emotion embedding to deal with aspect-level emotion classification tasks, according to Chinese Two embedding methods are proposed for the characteristics of comments: 1) At the level of Chinese words, the embedding of emotional words in comment sentences is trained using an adversarial method, and the embedding vectors are sent to a two-layer bidirectional LSTM encoder for pre-training, and the obtained 2) At the level of Chinese characters, the comment dataset is extended, and the BERT model is used to obtain the semantic embedding of comment sentences. Then, the extracted emotional features and the knowledge of the semantic embedding layer are integrated by using linear weighting and multi-head self-attention mechanism, and then the aspe...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a Chinese aspect level sentiment classification method based on pre-training sentiment embedding, and provides a deep learning model combining pre-training sentiment embedding and semantic embedding. The model converts an aspect-level sentiment classification problem into a sentence pair classification problem, fine-tunes the BERT model on the word level of Chinese comments, and trains sentiment feature vectors on the word level by using an adversarial learning method and a double-layer bidirectional LSTM encoder module. And the obtained emotion information is dynamically combined with the last several layers of semantic information of BERT through linear weighting and a multi-head self-attention mechanism, so that the semantic and emotion information of a specific aspect category is obtained, and accurate emotion classification is carried out. The characteristics of Chinese patient comments are combined, rich decision information is provided for the patient, some doctors who pay more attention to performance and weaken medical science and ability are made to improve own work, and important guarantee is provided for improvement of online medical quality and doctor service level.

Description

technical field [0001] The invention relates to the technical field of natural language processing, in particular to an aspect-level emotion classification method for Chinese patient comments based on pre-trained emotion embedding. Background technique [0002] With the development of online medical service platforms, more and more patients choose to seek medical treatment online and are willing to share their feelings and experiences during the medical treatment process on the medical platform. Previous studies categorized the sentiment of patient reviews, which could determine whether patients were satisfied or dissatisfied with their physicians. However, one patient review may involve multiple aspects, and this classification is too general to distinguish the emotional polarity of multiple aspects in a Chinese patient review. For example, given the comment sentence "Dr. Lu's attitude is very cold, but after he performed the operation on me, my symptoms were significantly...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06K9/62G06F40/30G06F40/284G06N3/04G06N3/08
CPCG06N3/08G06F40/284G06F40/30G06N3/044G06F18/2415
Inventor 车超单咏雪魏小鹏
Owner DALIAN UNIVERSITY
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap