Generating and Analyzing Sentiment using LLM and Machine Learning Techniques

This paper explores the generation and sentiment analysis of song lyrics using a fine-tuned LLM Long Language Model and machine learning techniques. Our study focuses on generating lyrics inspired by Taylor Swift’s songwriting style, followed by a detailed sentiment analysis to evaluate the emotional tone of the generated content. Previous situation text generation was still done by hand, with a number of drawbacks that appear, mostly from the point of time, error, insufficient creation, lack of anti- plagiarism skills from a few writers which may lead to a reduction in both the quality and variety of the sentences encased. This research was carried out by creating an application that allows even a passive user to automatically generate text using GPT-2 AI models. Sentiment analysis is the process to understand the feelings like emotions, attitudes, opinions, thoughts. We analyze the sentiment of the generated lyrics using machine learning algorithms such as Support Vector Machines (SVM). This dual approach is aimed not just at reproducing the style of the lyrics but at understanding how the emotions embedded in the text can be changed in the text to be generated. The results demonstrate the capability of AI-driven language models to generate creative content and offer meaningful sentiment analysis for application in the music industry.