Sentiment-Analysis-Project

An NLP-powered sentiment classifier that preprocesses text, extracts TF-IDF features, and uses Logistic Regression to predict Positive, Negative, or Neutral sentiment.

Project Overview

This project demonstrates the complete Natural Language Processing (NLP) pipeline, including:

• Data preprocessing
• Text cleaning
• TF-IDF feature extraction
• Machine Learning model training
• Sentiment prediction
• Model evaluation

The model uses Logistic Regression for sentiment classification and is trained on a labeled dataset.

---

Features

• ✔ Text preprocessing
• ✔ Stopword removal
• ✔ TF-IDF vectorization
• ✔ Logistic Regression classifier
• ✔ Predicts Positive, Negative, and Neutral sentiments
• ✔ Model accuracy evaluation
• ✔ Confusion Matrix visualization
• ✔ User input sentiment prediction

---

Technologies Used

• Python
• Pandas
• NumPy
• Scikit-learn
• NLTK
• Matplotlib
• Jupyter Notebook

---

Installation

Clone the repository:

git clone https://github.com/YourUsername/Sentiment-Analysis-Project.git

Navigate to the project folder:

cd Sentiment-Analysis-Project

Install the required libraries:

pip install -r requirements.txt

Launch Jupyter Notebook:

jupyter notebook

Open:

Sentiment_Analysis.ipynb

Run all the cells.

---

Model Workflow

1. Load the dataset
2. Clean and preprocess text
3. Remove stopwords
4. Convert text into TF-IDF features
5. Split data into training and testing sets
6. Train Logistic Regression model
7. Evaluate model performance
8. Predict sentiment for new user input

---

Results

The model successfully classifies text into:

• 😊 Positive
• 😐 Neutral
• 😞 Negative

Performance is evaluated using:

• Accuracy Score
• Classification Report
• Confusion Matrix

Demonstration

Screenshots

Dataset Preview

Confusion Matrix

Prediction Output

---

Learning Outcomes

Through this project, I learned:

• Fundamentals of Natural Language Processing
• Text preprocessing techniques
• TF-IDF feature extraction
• Machine Learning model training
• Model evaluation using Scikit-learn
• GitHub project organization and documentation

---

Future Improvements

• Deploy as a web application using Streamlit
• Train on larger datasets
• Experiment with Deep Learning models (LSTM/BERT)
• Improve prediction accuracy
• Add real-time sentiment analysis

License

This project is licensed under the MIT License.

---

Author

Abhin R