Employability Of Natural Language Processing (NLP) Algorithmic Techniques In Developing And Enhancing The Efficacy Of Performance Analysis Of Ensemble Methods On Twitter Sentiment Analysis

Astha

Vol. 10, Issue 1, Jan-Dec 2024

Page Number: 50 - 55

Abstract:

Social media platforms such as Twitter have influenced public opinion and sentiment that grows by the day and has made sentiment analysis indispensable. However, traditional machine learning approaches often come up short when they are made to face the noisy, high-dimensional, and context-dependent nature of Twitter data. The paper focuses on using ensemble methods to improve the performance of sentiment analysis tasks. This research integrated advanced ensemble techniques such as Random Forest, Gradient Boosting, and Stacking with state-of-the-art NLP methods. This study focuses on the different pre-processing pipelines of feature extraction using TF-IDF, Word2Vec, BERT embeddings, and hyper parameter optimization that play a crucial role in classifying sentiment for models. The experimental results show that ensemble methods outperform traditional machine learning algorithms with significantly higher accuracy, precision, recall, and F1 scores. Among the methods, Stacking proves to be the most effective because it utilizes the complementary strengths of base models. The results emphasize the capability of ensemble learning combined with advanced NLP techniques to deal with the complexities of Twitter sentiment analysis. This research provides valuable insights for academics and industry professionals looking to improve text classification systems in dynamic and challenging domains.

References

• B. Pang, L. Lee, and S. Vaithyanathan, 'Thumbs up? Sentiment classification using machine learning techniques,' in Proc. ACL-02 Conf. Empirical Methods in Natural Language Processing (EMNLP), pp. 79–86, 2002.
• A. Vaswani et al., 'Attention is all you need,' in Proc. Advances in Neural Information Processing Systems (NeurIPS), 2017.
• L. Breiman, 'Random forests,' Machine Learning, vol. 45, no. 1, pp. 5–32, 2001.
• J. H. Friedman, 'Greedy function approximation: A gradient boosting machine,' Annals of Statistics, vol. 29, no. 5, pp. 1189–1232, 2001.
• X. Zhang et al., 'Text sentiment analysis based on stacking ensemble learning,' IEEE Access, vol. 7, pp. 152401–152412, 2019.
• K. Nigam et al., 'Text classification from labelled and unlabeled documents using EM,' Machine Learning, vol. 39, no. 2-3, pp. 103–134, 2000.
• A. Go, R. Bhayani, and L. Huang, 'Twitter sentiment classification using distant supervision,' CS224N Project Report, Stanford, 2009.

Back Download