Aspect-Based Sentiment Analysis Using Local Context Focus Mechanism with DeBERTa
Authors:
- 22 September 2023
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Abstract
Text sentiment analysis, often termed as opinion mining, delves into quantifying individuals' opinions, evaluatio ns, attitudes, and emotions conveyed about entities. Sentiment a nalysis of text can be categorized into text-level, sentence-level, and aspect-level analyses. Aspect-Based Sentiment Analysis (A BSA) represents a detailed sub-discipline within sentiment anal ysis, with its primary goal being to ascertain the sentiment pola rity of specific aspects. The research of pre-training neural mod el has significantly improved the performance of many natural language processing tasks. In recent years, pre training model (PTM) has been applied in ABSA. Therefore, there has been a q uestion, which is whether PTMs contain sufficient syntactic info rmation for ABSA. In this paper, we explored the recent DeBE RTa model (Decoding-enhanced BERT with disentangled attention) to solve Aspect-Based Sentiment Analysis problem. DeBE RTa is a kind of neural language model based on transformer, which uses self-supervised learning to pre-train on a large num ber of original text corpora. Based on the Local Context Focus (LCF) mechanism, by integrating DeBERTa model, we purpos e a multi-task learning model for aspect-based sentiment analys is. The experiments result on the most commonly used the lapto p and restaurant datasets of SemEval-2014 and the ACL twitte r dataset show that LCF mechanism with DeBERTa has signifi cant improvement.Keywords
All Related Versions
Funding Information
- National Natural Science Foundation of China (62192784,U22B2038,62172056)
This publication has 39 references indexed in Scilit:
- SemEval-2016 Task 5: Aspect Based Sentiment Analysis2016
- Survey on Aspect-Level Sentiment AnalysisIEEE Transactions on Knowledge and Data Engineering, 2015
- High‐precision formation control of nonlinear multi‐agent systems with switching topologies: A learning approachInternational Journal of Robust and Nonlinear Control, 2014
- Adaptive Recursive Neural Network for Target-dependent Twitter Sentiment Classification2014
- SemEval-2014 Task 4: Aspect Based Sentiment Analysis2014
- XRCE: Hybrid Classification for Aspect-based Sentiment Analysis2014
- Glove: Global Vectors for Word Representation2014
- A review of object representation based on local featuresJournal of Zhejiang University SCIENCE C, 2013
- Tracking Algorithms for Multiagent SystemsIEEE Transactions on Neural Networks and Learning Systems, 2013
- Robust iterative learning protocols for finite-time consensus of multi-agent systems with interval uncertain topologiesInternational Journal of Systems Science, 2013