This project focuses on extracting opinions about specific product features from user reviews. It leverages Natural Language Processing (NLP) techniques to analyze sentiment at the feature level.
- Source: Publicly available product review datasets from e-commerce platforms.
- Format: Text reviews with metadata such as rating, product name, and timestamps.
- Preprocessing: Tokenization, stopword removal, and lemmatization.
- Feature Extraction: Identifying product features using Named Entity Recognition (NER) and dependency parsing.
- Opinion Mining: Assigning sentiment scores to extracted features using sentiment lexicons and machine learning models.
- Modeling: Implemented machine learning classifiers (Naïve Bayes, SVM) and deep learning models (LSTMs) for sentiment analysis.
- Achieved 85% accuracy in feature-opinion extraction.
- Identified top positive and negative features per product category.
- Developed interactive visualizations to explore sentiment trends over time.
- Programming Languages: Python, SQL
- Libraries: NLTK, SpaCy, Scikit-learn, TensorFlow
- Visualization: Matplotlib, Seaborn
- Expand dataset with multi-domain product reviews.
- Integrate transformer-based models (e.g., BERT) for improved accuracy.
- Develop a real-time API for opinion mining.