HSA: A Hybrid Sentiment Analysis Technique using Machine Learning
Vanshika, Neetu Rani, Ranjan Walia · 2024
Social media provide companies the opportunity to engage in discussions with customers and learn what they are thinking about their products. Thus, businesses are interested in utilizing them to promote their products or services, find innovative prospects, and enhance their brand. This paper provides a novel Hybrid Sentiment Analysis (HSA) approach which uses Machine Learning techniques to enhance sentiment analysis performance on Amazon and Flipkart datasets. The proposed HSA method gathers sentiment features from textual input by combining and using machine learning techniques, namely Random Forest (RF) and Gradient Boost (GB) classifiers. The findings of this study demonstrate that the proposed approach effectively identifies sentiment nuances and offers cutting-edge performance in a range of sentiment analysis tasks. Evaluation metrics including recall, precision, accuracy, and F1-score are used to assess the models after they have been trained on preprocessed data. Using the Flipkart and Amazon datasets, the proposed method is compared to the standard models, achieving the best accuracy rates of ${9 0 . 1 0 \%}$ and ${9 5 . 4 0 \%}$, respectively.