OpinionAI: Sentiment Analysis of Consumer Feedback

• Tech Stack: Python, NLTK, spaCy, Hugging Face Transformers, RoBERTa, VADER
• Github URL: Project Link

Developed a sentiment analysis system to predict the sentiment (positive, neutral, or negative) of food product reviews. Implemented data cleaning using NLTK and spaCy to ensure the dataset's quality for analysis.

Employed VADER for initial rule-based sentiment scoring and used RoBERTa from Hugging Face Transformers for a transformer-based analysis capable of capturing nuanced sentiments with high accuracy.

The comparative analysis revealed that while both models provided similar outputs for positive and negative reviews, RoBERTa demonstrated superior precision in neutral sentiment detection.

This project showcased the advantages of using advanced transformer models for handling complex sentiment distinctions, contributing valuable insights for sentiment analysis implementations.