📝 Sentiment Analysis using Traditional Machine Learning Model & Transformers

This project performs sentiment analysis on customer feedback using Machine Learning (ML) models and a Transformer-based model (DistilBERT). The trained models are deployed using Streamlit for easy user interaction.

---

🚀 Project Overview

This project involves:

• Data Preprocessing: Cleaning and tokenizing text data.
• Traditional ML Models: Logistic Regression, Random Forest, XGBoost.
• Transformer Model: Fine-tuned DistilBERT for improved accuracy.
• Model Evaluation: Comparing ML models with transformer-based models.
• Deployment:
  • Streamlit Web App → Live Demo

---

📂 Project Structure

📁 Sentiment-Analysis │── app.py # Streamlit Web App │── requirements.txt # Dependencies │── README.md # Documentation │── 📂 models/ # Saved Models (pkl, transformers)

---

📊 Dataset

• The train dataset contains 3.5 million customer reviews labeled as Positive (1) or Negative (0). It was split to train (90%) and validation (10%)
• The test dataset contains 400 000 customer reviews labeled as Positive (1) or Negative (0)

---

🔧 Installation

1️⃣ Clone the Repository

git clone https://github.com/boiBASH/sentiment-analysis-app.git