Harnessing GPT-2 Sequence Models for Sentiment Analysis on Twitter Data

---

Abstract

Instant messaging platforms like Twitter have become integral parts of modern communication, facilitating rapid exchanges of information and fostering interpersonal connections. However, extracting and understanding sentiments from these conversations pose significant challenges due to their unstructured and dynamic nature. Leveraging AI approaches, particularly natural language processing (NLP) and machine learning algorithms, enables the automated analysis of sentiments within these contexts. This paper includes various Artificial Intelligence (AI) techniques utilized for sentiment analysis, including supervised learning, unsupervised learning, and deep learning methods. Additionally, it discusses the implications of sentiment analysis in enhancing user experience and their views about any things, person or product as positive or negative opinion. This paper used Bernoulli Distribution, Decision Tree Classifier, Logistic Regression as machine learning (ML) algorithms and GPT2 Sequence Classifier (Deep learning) to analyze the attitudes in real-time Twitter data opinion mining. Finally, the performance of these approaches as evaluated using well known parameter of classifier and learning outcome.

---

Keywords

Machine learning, Artificial intelligence, Natural Language Processing, Sentiment Analysis, Deep learning