Enhancement of support vector machine utilizing Roberta applied to sentiment analysis of facebook data.

ABSTRACT: Many people are using social media sites lie Facebook to express their opinions, experiences. Or whatever they want to post online. Understanding user sentiment has become crucial for various applications, ranging from marketing to public opinion analysis. Researchers use natural language processing (NLP) and machine learning algorithms to evaluate textual information from Facebook posts and classify sentiments as positive, negative, or neutral. This study delves into sentiment analysis of Facebook data to better understand how users express their emotions. Additionally, the method addresses the limitations of sentiment analysis on social media due to informal language, slang, and context-dependent phrases. The study aims to develop an enhanced Support Vector Machine algorithm for sentiment analysis of Facebook data by utilizing the RoBERTa ( A Robustly optimized BERT) model. To enhanced sentiment accuracy, thus the performance of the traditional SVM algorithm, the proposed approach uses VADER to predict initial sentiment labels, loads a pre-trained RoBERTa model as preprocessing techniques, fine-tunes the RoBERTa model and extracts RoBERTa embeddings to optimize the SVM algorithm. This improves the model’s capacity to handle imbalanced datasets and efficiently manage larger datasets while filtering out noisy or irrelevant characteristics. To analyse the performance of the proposed technique, results are compared with the result of existing algorithms. The enhanced SVM algorithm significantly outperforms the existing approach in terms of accuracy, precision, recall, and F1-score, with a 4% to 8% improvement in accuracy over the previous algorithm. This research highlights the potential of integrating RoBERTa techniques with SVM for enhanced sentiment analysis.

5

ABSTRACT: Many people are using social media sites lie Facebook to express their opinions, experiences. Or whatever they want to post online. Understanding user sentiment has become crucial for various applications, ranging from marketing to public opinion analysis. Researchers use natural language processing (NLP) and machine learning algorithms to evaluate textual information from Facebook posts and classify sentiments as positive, negative, or neutral. This study delves into sentiment analysis of Facebook data to better understand how users express their emotions. Additionally, the method addresses the limitations of sentiment analysis on social media due to informal language, slang, and context-dependent phrases. The study aims to develop an enhanced Support Vector Machine algorithm for sentiment analysis of Facebook data by utilizing the RoBERTa ( A Robustly optimized BERT) model. To enhanced sentiment accuracy, thus the performance of the traditional SVM algorithm, the proposed approach uses VADER to predict initial sentiment labels, loads a pre-trained RoBERTa model as preprocessing techniques, fine-tunes the RoBERTa model and extracts RoBERTa embeddings to optimize the SVM algorithm. This improves the model's capacity to handle imbalanced datasets and efficiently manage larger datasets while filtering out noisy or irrelevant characteristics. To analyse the performance of the proposed technique, results are compared with the result of existing algorithms. The enhanced SVM algorithm significantly outperforms the existing approach in terms of accuracy, precision, recall, and F1-score, with a 4% to 8% improvement in accuracy over the previous algorithm. This research highlights the potential of integrating RoBERTa techniques with SVM for enhanced sentiment analysis.

Filipiniana

5