Comment target emotion analysis based on BERT fine adjustment model

Abstract

The invention discloses comment target emotion analysis based on a BERT fine adjustment model, which comprises a BCRCRF target extraction model and a BCCDA target emotion classification model, and ischaracterized in that the BCCDA target emotion classification model is divided into experimental results on online course comment emotion analysis, the BCRCRF target extraction model, the BCCDA targetemotion analysis model and a real Chinese online course comment data set; the BCRCRF target extraction model comprises the following steps: 1, performing intra-domain unsupervised training on a BERTpre-training model BCR based on a large-scale Chinese comment corpus; 2, introducing a CRF layer, adding grammatical constraints to an output sequence of a semantic representation layer in the BCR model, ensuring the rationality of a dependency relationship between prediction tags, and accurately extracting a comment target in a course comment text; and 3, constructing a BCCDA model containing double attention layers to express the emotion polarity of the course comment target in a classified manner. According to the method, the target emotion contained in the online course comments can be accurately analyzed, and the method has important significance in understanding the emotion change of learners and improving the course quality.

Description

technical field [0001] The invention relates to the technical field of online education, in particular to a comment target sentiment analysis based on a BERT fine-tuning model. Background technique [0002] In recent years, online courses have rapidly increased the number of online learners because of their openness, convenience and high quality. A large number of learners with different backgrounds and levels pose new challenges to the management of online courses: real-time grasp of learning situations or emotional changes online, and understanding learning motivations to better provide personalized education services. In order to adapt to this trend, online learning platforms will provide community functions such as comments and forums to facilitate the communication and interaction between learners, teachers and platform managers. Sentiment analysis can effectively obtain learners' emotional attitudes, learning experience, etc., so as to understand the changes in learne...

AI Technical Summary

Problems solved by technology

However, most of the existing online course review sentiment analysis based on neural network is to classify the emotional polarity of the entire sentence review, which cannot effectively identify fine-grained review targets and calculate their corresponding emotional information.

In addition, machine learning and deep learning methods usually require a large amount of manually labeled data for model training, which is difficult to directly apply to the sentiment analysis of Chinese online course comment targets, which is extremely expensive for manual labeling.

Images