Abstract
Aspect-Based Sentiment Classification (ABSC) is the core technology for e-commerce websites, offering business entities in-depth customer preference information through online review sentiment extraction. In this paper, a Hybrid Bi-LSTM Framework is proposed to integrate weak supervision with deep learning models to increase the accuracy of sentiment classification. The approach is based on aspect term extraction, weak labeling with Snorkel-based methods, and a multi-model sentiment analysis method known as HABSC (Hybrid Aspect-Based Sentiment Classification). It uses Amazon product review datasets to evaluate the performance difference between a Hybrid Bi-LSTM model, the HABSC algorithm, and an ABSC DeBERT model in sentiment classification for five important aspects: price, quality, usability, size, and service. Experimental results indicate that Hybrid Bi-LSTM performs better than ABSC DeBERT, achieving 93.4% accuracy, lower Hamming loss, and improved precision, recall, and F1-score. Comparative analysis also reflects the performance of an ensemble-based sentiment analysis approach employing VADER, SentiWordNet, and BERT scoring. This article contributes to enhanced automated aspect-based sentiment analysis, presenting an efficient and scalable solution for recommendation systems, business intelligence, and e-commerce websites.
References
K. Subbaiah and B. K. Bolla, “Aspect category learning and sentimental analysis using weakly supervised learning,” Procedia Computer Science, vol. 235, (2024): 1246–1257.
Kannikaklang, N.Thamviset, Wongthanavasu. S (2024). BiLSTCAN: A novel SRS-based bidirectional long short-term capsule attention network for dynamic user preference and next-item recommendation. IEEE Access, 12, 6879–6899.
Ozyurt, Baris, and M. Ali Akcayol. "A new topic modeling based approach for aspect extraction in aspect based sentiment analysis: SS-LDA." Expert Systems with Applications 168 (2021): 114231.
Zhao, Anping, and Yu Yu. "Knowledge-enabled BERT for aspect-based sentiment analysis." Knowledge-Based Systems 227 (2021): 107220.Ozyurt and M. A. Akcayol, "A New Topic Modeling-Based Approach for Aspect Extraction in Aspect-Based Sentiment Analysis: SS-LDA," International Journal of Computational Linguistics, 2022.
Liu, Haoyue, et al. "Aspect-based sentiment analysis: A survey of deep learning methods." IEEE Transactions on Computational Social Systems 7.6 (2020): 1358-1375.
Zhang, Wenxuan, et al. "A survey on aspect-based sentiment analysis: Tasks, methods, and challenges." IEEE Transactions on Knowledge and Data Engineering 35.11 (2022): 11019-11038.
Nazir, Ambreen, et al. "Issues and challenges of aspect-based sentiment analysis: A comprehensive survey." IEEE Transactions on Affective Computing 13.2 (2020): 845-863.
K. Lin, "Sentiment Analysis Algorithm Based on BERT and Convolutional Neural Network," International Journal of Machine Learning Applications, vol. 18, no. 3, (2021): 79–95.
R. E. Fikes and N. J. Nilsson, "STRIPS: A New Approach to the Application of Theorem Proving to Problem Solving," in Proceedings of the 2nd International Joint Conference on Artificial Intelligence (IJCAI’71), (1971): 608–620.