Electrotechnical and Computer Engineering
Vol. 39 No. 11 (2024): Proceedings of Faculty of Technical Sciences
ADVANCED TECHNIQUES FOR SENTIMENT ANALYSIS: A STUDY OF CLASSIFICATION AND GENERATIVE MODELS ON ONLINE COMMENTS
Abstract
This paper explores sentiment analysis of user comments using various natural language processing techniques and deep learning algorithms. Implemented models include Recurrent Neural Networks (RNN), Convolutional Neural Networks (CNN), Word2Vec, GloVe, BERT, and the generative llama-7b-hf model. The dataset was obtained from the Kaggle platform and contains user comments on various products. Model evaluation was performed using the F1 score, enabling a detailed analysis of performance in the context of imbalanced classes. The best results were achieved using transformers and SVM classifiers.
References
[1] https://www.kaggle.com/datasets/nicapotato/womens-ecommerce-clothing-reviews (pristupljeno u martu 2022.)
[2] Abien Fred M. Agarap, „Statistical Analysis on E-Commerce Reviews, with Sentiment Classification using Bidirectional Recurrent Neural Network“, 2018.
[3] Shuangyin Xie, „Sentiment Analysis using machine learning algorithms: online women clothin reviews“, 2019.
[4] Manish Munikar, Sushil Shakya, Aakash Shrestha, “Fine-grained Sentiment Classification using BERT”, 2019.
[5] Jun Xie, Bo Chen, Xinglong Gu, Fengmei Liang, Xinying Xu, “Self-Attention-Based BiLSTM Model for Short Text Fine-Grained Sentiment Classification”, 2018.
[6] https://www.crummy.com/software/BeautifulSoup/bs4/doc/ (pristupljeno u martu 2022.)
[7] https://keras.io/ (pristupljeno u martu 2022.)
[8] https://www.tensorflow.org/ (pristupljeno u martu 2022.)
[9] Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu, “MPNet: Masked and Permuted Pre-training for Language Understanding”, 2020.
[10] https://huggingface.co/meta-llama/Llama-2-7b-hf (pristupljeno u februaru 2024.)
[11] https://spacy.io/ (pristupljeno u martu 2022.)