project addition - ML Projects/sentiment_prediction_python_package

[vijaytakbhate2002]

I have added sentiment_prediction_python_package which can be installed with pip command, for more information see below readme.md file

---

Twitter Sentiment Prediction Python Package

This project focuses on predicting the sentiment of Twitter posts using Natural Language Processing (NLP) techniques. It is an end-to-end machine learning project that includes data preprocessing, model building, hyperparameter tuning, and deployment.

---

How I built this package

Download code packaging process.pdf

Features

• ETL Pipeline: Extract, transform, and load (ETL) processes implemented using PySpark and SQL.
• Text Processing: Cleaned text data using regex, removed special characters, and vectorized text using TF-IDF.
• Model Building: Implemented Logistic Regression and Multinomial Naive Bayes for sentiment prediction.
• Hyperparameter Tuning: Optimized models using GridSearchCV.
• Experiment Tracking: Logged experiments with MLflow for efficient performance comparison.
• Model Packaging: Prepared reusable model packages using sdist and wheel.
• Deployment Ready: Plan to host the project on Render for public access.

---

Installation

1. Clone the repository:

   git clone https://github.com/vijaytakbhate2002/sentiment_prediction_python_package.git  

2. Install the package using pip:

   pip install git+https://github.com/vijaytakbhate2002/sentiment_prediction_python_package.git  

---

Usage

Prediction Example

To predict the sentiment of a Twitter post, you can use the following Python code:

from sentiment_prediction import predict  
predict.predictor("this is a negative tweet")  

This will return:

['Negative']

Sentiment Classes

The prediction function supports four sentiment classes:

• Positive
• Negative
• Neutral
• Irrelevant

---

Project Structure

sentiment_prediction_python_package/
│
├── sentiment_prediction/                 # Main package directory
│   ├── config/                           
│   │   └── config.py                     # Configuration file
│   │
│   ├── data_manipulation/                # Model training and prediction scripts
│   │   ├── data_handling.py              # Training pipeline
│   │   ├── data_processing.py            # Prediction pipeline
│   │   └── text_filer.py                 # Text filtering utilities
│   │
│   └── trained_models/                   # Trained models and metadata
│       ├── classifier.pkl                # Classifier model
│       ├── vectorizer.pkl                # Vectorizer model
│       └── encoder.pkl                   # Encoder model
│
├── dist/                                 # Distribution packages (generated)
│
├── build/                                # Build files (generated)
│
├── sentiment_prediction.egg-info/        # Egg-info metadata (generated)
│
├── .gitignore                            # Excluded files and folders
├── MANIFEST.in                           # Configuration file with paths and constants
├── README.md                             # Project documentation
├── requirements.txt                      # Project dependencies
└── setup.py                              # Package metadata and configuration

---

Contribution

Feel free to submit issues or pull requests. Contributions are welcome!

---

License

This project is licensed under the MIT License.

---

Author

Vijay Dipak Takbhate

---

GitHub Repository: Twitter Sentiment Prediction Python Package

---