Overview
Disaster Tweet Detection is a machine learning project designed to classify tweets as either related to a disaster or not. The project explores various models to assess their performance in accurately detecting disaster-related tweets.
Features
Dataset Processing: Preprocessing techniques applied to clean and tokenize text data.
Model Evaluation: Multiple machine learning models tested, including Logistic Regression, Naïve Bayes, Random Forest, and deep learning models.
Performance Metrics: Comparison of models using accuracy, precision, recall, and F1-score.
Visualization: Data distribution, word clouds, and model performance visualized using graphs.
Jupyter Notebooks: Interactive notebooks used to run experiments and analyze results.
Installation
- Clone the repository:
git clone https://github.com/amin224/DisasterTweetDetection.git cd DisasterTweetDetection
Install dependencies:
- pip install -r requirements.txt
Launch Jupyter Notebook:
- jupyter notebook
Usage
Open the provided Jupyter Notebooks.
Run preprocessing steps to clean the dataset.
Train and evaluate different machine learning models.
Compare performance metrics and analyze results.
Contributions
Contributions are welcome! Feel free to submit issues or pull requests to enhance the project.