From Conditional Random Field (CRF) to Rhetorical Structure Theory(RST): Incorporating Context Information in Sentiment Analysis

2017 ◽  
pp. 283-295 ◽  
Author(s):  
Aggeliki Vlachostergiou ◽  
George Marandianos ◽  
Stefanos Kollias
Keyword(s):  
Random Field ◽  
Sentiment Analysis ◽  
Conditional Random Field ◽  
Structure Theory ◽  
Context Information ◽  
Rhetorical Structure

scholarly journals JU_CSE: A Conditional Random Field (CRF) Based Approach to Aspect Based Sentiment Analysis

2014 ◽  
Author(s):  
Braja Gopal Patra ◽  
Soumik Mandal ◽  
Dipankar Das ◽  
Sivaji Bandyopadhyay
Keyword(s):  
Random Field ◽  
Sentiment Analysis ◽  
Conditional Random Field

Aspect-Based Sentiment Analysis on the Web Using Rhetorical Structure Theory

2016 ◽  
pp. 317-334 ◽  
Author(s):  
Rowan Hoogervorst ◽  
Erik Essink ◽  
Wouter Jansen ◽  
Max van den Helder ◽  
Kim Schouten ◽  
...  
Keyword(s):  
Sentiment Analysis ◽  
Structure Theory ◽  
Rhetorical Structure ◽  
The Web

A BERT Fine-tuning Model for Targeted Sentiment Analysis of Chinese Online Course Reviews

2020 ◽  
Vol 29 (07n08) ◽  
pp. 2040018
Author(s):  
Huibing Zhang ◽  
Junchao Dong ◽  
Liang Min ◽  
Peng Bi
Keyword(s):  
Random Field ◽  
Sentiment Analysis ◽  
Large Scale ◽  
Conditional Random Field ◽  
Semantic Representation ◽  
Fine Tuning ◽  
Online Course ◽  
Analysis Model ◽  
Accurate Analysis ◽  
Course Quality

Accurate analysis of targeted sentiment in online course reviews helps in understanding emotional changes of learners and improving the course quality. In this paper, we propose a fine-tuned bidirectional encoder representation from transformers (BERT) model for targeted sentiment analysis of course reviews. Specifically, it consists of two parts: binding corporate rules — conditional random field (BCR-CRF) target extraction model and a binding corporate rules — double attention (BCR-DA) target sentiment analysis model. Firstly, based on a large-scale Chinese review corpus, intra-domain unsupervised training of a BERT pre-trained model (BCR) is performed. Then, a Conditional Random Field (CRF) layer is introduced to add grammatical constraints to the output sequence of the semantic representation layer in the BCR model. Finally, a BCR-DA model containing double attention layers is constructed to express the sentiment polarity of the course review targets in a classified manner. Experiments are performed on Chinese online course review datasets of China MOOC. The experimental results show that the F1 score of the BCR-CRF model reaches above 92%, and the accuracy of the BCR-DA model reaches above 72%.

Rule-Based Polarity Aggregation Using Rhetorical Structures for Aspect-Based Sentiment Analysis

2019 ◽  
Vol 10 (3) ◽  
pp. 44-60
Author(s):  
Nuttapong Sanglerdsinlapachai ◽  
Anon Plangprasopchok ◽  
Tu Bao Ho ◽  
Ekawit Nantajeewarawat
Keyword(s):  
Sentiment Analysis ◽  
Negative Polarity ◽  
Structure Theory ◽  
Classification Rules ◽  
Rule Based ◽  
Rhetorical Structure ◽  
Confidence Threshold ◽  
Discourse Relations ◽  
Rhetorical Structures ◽  
F Measure

The segments of a document that are relevant to a given aspect can be identified by using discourse relations of the rhetorical structure theory (RST). Different segments may contribute to the overall sentiment differently, and the sentiment of one segment may affect the contribution of another segment. This work exploits the RST structures of relevant segments to infer the sentiment of a given aspect. An input document is first parsed into an RST tree. For each aspect, relevant segments with their relations in the resulting tree are localized and transformed into a set of features. A set of classification rules is subsequently induced and evaluated on data. The proposed framework performs well in several experimental settings, with the accuracy values ranging from 74.0% to 77.1% being achieved. With proper strategies for removing conflicting rules and tuning the confidence threshold, f-measure values for the negative polarity class can be improved.

Sign in / Sign up

Export Citation Format

Share Document