The Application of Deep Learning Techniques in Analyzing Customer Opinions on Social Media Platforms in a Business Context Online
Article Sidebar
Main Article Content
Abstract
Social media has become a primary channel for customers to express their opinions on products and services, providing valuable insights for business strategy development. However, classifying customer opinions from diverse text structures and informal language presents a significant challenge, requiring deep learning techniques to enhance the accuracy of sentiment analysis. This research aims to 1) Analyze customer feedback from social media platforms and identify and classify sentiments as positive, negative and neutral. And 2) Evaluate the performance of deep learning models in analyzing customer feedback from social media platforms. This research is applied research that utilizes deep learning techniques to analyze customer feedback from social media platforms such as Twitter and Facebook. The process begins by collecting customer feedback from both platforms. The data is then cleaned and prepared for analysis. Next, deep learning models are used to classify the feedback into three main categories positive sentiment, negative sentiment and neutral sentiment. The model's performance will be evaluated using four key metrics accuracy, precision, recall and F1-score to assess the effectiveness of the model in classifying customer feedback. The results of the study showed that. 1.) The BERT model achieved the highest accuracy in classifying sentiments with an accuracy of 93.0% for positive sentiment and 92.8% for negative sentiment. The LSTM model had an accuracy of 89.2% but it did not perform as well as BERT in classifying positive sentiment. The CNN model had an accuracy of 86.9% and its performance was lower than both BERT and LSTM in capturing complex meanings. And 2) The performance evaluation of the deep learning models revealed that BERT achieved the highest accuracy at 92.5% followed by LSTM with an accuracy of 88.7% and CNN with an accuracy of 86.3% respectively.
Downloads
Article Details
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
บทความที่ได้รับการตีพิมพ์เป็นลิขสิทธิ์ของวารสารการบัญชีและการจัดการ
ข้อความที่ปรากฏในบทความแต่ละเรื่องในวารสารวิชาการเล่มนี้เป็นความคิดเห็นส่วนตัวของผู้เขียนแต่ละท่านไม่เกี่ยวข้องกับมหาวิทยาลัยมหาสารคาม และคณาจารย์ท่านอื่นๆในมหาวิทยาลัยฯ แต่อย่างใด ความรับผิดชอบองค์ประกอบทั้งหมดของบทความแต่ละเรื่องเป็นของผู้เขียนแต่ละท่าน หากมีความผิดพลาดใดๆ ผู้เขียนแต่ละท่านจะรับผิดชอบบทความของตนเองแต่ผู้เดียว
References
Başarslan, M. S., & Kayaalp, F. (2021). Sentiment analysis on social media reviews datasets with deep learning approach. Sakarya University Journal of Computer and Information Sciences, 4(1), 35–49.
Broklyn, P., Olukemi, A., & Bell, C. (2024). Social media sentiment analysis for brand reputation management. Available at SSRN 4906218, 1-14
Brown, T. B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., Agarwal, S., Herbert-Voss, A., Krueger, G., Henighan, T., Child, R., Ramesh, A., Ziegler, D. M., Wu, J., Winter, C., … Amodei, D. (2020). Language models are few-shot learners. In Advances in Neural Information Processing Systems, 33, 1877-1901.
Chaiwichit C. (2017). Creating and developing model. Silpakorn Education Research Journal, 9(1), 1-11. Retrieved January 18, 2025, from https://so05.tci-thaijo.org/index.php/suedureasearchjournal/issue/view/8124 [in Thai].
Chakraborty, K., Bhattacharyya, K., & Bag, R., (2020). A survey of sentiment analysis from social media data. IEEE Transactions on Computational Social Systems, 7(2), pp. 450-464,
Goodfellow, I., Bengio, Y., & Courville, A. (2020). Deep learning an MIT press book. MIT Press.
Hernández-Romero, I. M., Barajas-Villarruel, L. R., Flores-Tlacuahuac, A., Fuentes-Cortes, L. F., & Rico-Ramirez, V. (2023). Strategic planning for sustainable electric system operations: Integrating renewables and energy storage. Computers & Chemical Engineering, 177, 108312.
Jurafsky, D., & Martin, J. H. (2021). Speech and language processing. (3rded.), Pearson.
Kim, S. M., & Hovy, E. (2020). Determining the sentiment of opinions. Proceedings of the 20th International Conference on Computational Linguistics (COLING), 1367–1373.
Liu, B. (2020). Sentiment analysis: mining opinions, sentiments and emotions (2nd ed.). Cambridge University Press.
Messaoudi, C., Guessoum, Z., & Ben Romdhane, L. (2022). Opinion mining in online social media: a survey. In Social Network Analysis and Mining, 12(1), Springer.
Narejo, K. R., Zan, H., Dharmani, K. P., Zhou, L., Alahmadi, T. J., Sehito, N., Ghadi, Y. Y., & Assam, M. (2024). EEBERT: An emoji-enhanced BERT fine-tuning on amazon product reviews for Text sentiment classification. IEEE Access, 12, 131954- 131967.
Sun, C., Qiu, X., Xu, Y., & Huang, X. (2020). How to fine-tune BERT for text classification. Journal of Computer Science Springer, 194-206.
Venugopalan, M., & Gupta, D. (2022). An enhanced guided LDA model augmented with BERT based semantic strength for aspect term extraction in sentiment analysis. Knowledge-based systems, 246, 108668.
Wang, C., Nulty, P., & Lillis, D. (2020). A comparative study on word embeddings in deep learning for text classification. In Proceedings of the 4th international conference on natural language processing and information retrieval, 37-46.
Win Myint, P. Y., Lo, S. L., & Zhang, Y. (2024). Unveiling the dynamics of crisis events: sentiment and emotion analysis via multi-task learning with attention mechanism and subject-based intent prediction. Information Processing and Management, 61(4), 103695.
Xinjie, L., Gang, X., Gaopeng, G., Zhen, L., Junzheng, S., & Jing, Y. (2022). ET-BERT: A contextualized datagram representation with pre-training transformers for encrypted traffic classification.WWW 2022 - Proceedings of the ACM Web Conference 2022, 633–642.
Zhang, L., Wang, S., & Liu, B. (2020). Deep learning for sentiment analysis: a survey. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, 10(3), 1328.
