SentimentSense

SentimentSense is a sentiment analysis application built with a fine-tuned BERT model. The app predicts whether a movie review is positive or negative and provides a confidence score for the prediction. It includes a user-friendly GUI built with Streamlit, making it easy to analyze custom or example reviews.

---

Features

• Fine-tuned BERT model trained on the IMDB Dataset of 50,000 reviews.
• Supports real-time sentiment analysis with a simple GUI.
• Confidence scores displayed alongside predictions for better insights.
• Option to enter custom reviews or select from predefined examples.
• Saves the trained model for reuse, avoiding repeated training.

---

Dataset

The model is trained on the IMDB Dataset containing 50,000 movie reviews with binary sentiment labels:

• Columns Used:
  • review: Text of the movie review.
  • sentiment: Label for sentiment (positive = 1, negative = 0).

---

Setup

1. Install Dependencies

Make sure you have Python 3.8+ installed. Use the requirements.txt file to install dependencies:

pip install -r requirements.txt

2. Download Dataset

Download the IMDB Dataset.csv file and place it in the project directory. Ensure the file contains the required columns (review, sentiment).

3. Run the Application

Start the Streamlit app:

streamlit run app.py

4. Open in Browser

After running, the app will open in your default browser, or you can access it at the local URL provided (e.g., http://localhost:8501).

---

Usage

GUI Features:

1. Pre-Trained Model Check:

   • If a trained model exists, it will load it automatically.
   • If no model is found, the app will train the model and save it for future use.

2. Analyze Sentiment:

   • Enter your own movie review in the provided text box.
   • Or select a predefined review from the dropdown menu.

3. See Results:

   • The app will display:
     • Sentiment Prediction: Positive or Negative.
     • Confidence Score: A value between 0 and 1, indicating the model's certainty.

---

Example

Input:

Review: "The movie was absolutely fantastic! A must-watch."

Output:

Prediction: Positive
Confidence: 0.982

---

Project Structure

• app.py: Main application file for Streamlit.
• IMDB Dataset.csv: The dataset used for training and testing.
• bert-imdb-output/: Directory where trained models are saved.
• requirements.txt: List of dependencies.

---

Requirements

Python Version:

• Python 3.8+

Dependencies:

Install all dependencies using requirements.txt:

streamlit==1.15.2
transformers==4.33.3
torch==2.0.1
scikit-learn==1.3.0
pandas==1.5.3
numpy==1.23.5

---

Notes

• The application will only train the model if a pre-trained model is not found.
• Training requires a GPU for faster performance.
• The trained model is saved in the bert-imdb-output/best_model directory.

---

This project is licensed under the MIT License. See the LICENSE file for details.