This repository documents a project focused on predicting forest cover types using the Kaggle dataset: Forest Cover Type Prediction Dataset. The project is part of the Big Data module of ENIT's 3rd year MIndS 2024-2025 program and is undertaken by Group 4: Chaima Balti, Roukaya Lakhzouri, and Salsabil Rouahi. We are working under the supervision of our professor, Moez Ben Haj Hmida.

The primary goal of this project is to explore and apply various machine learning techniques to accurately classify forest cover types based on features related to soil, climate, and topography.

The repository currently includes the following files and notebooks:

• README.md: This file provides an overview of the project, its objectives, and the file structure.
• SVM_all_data.ipynb: A Jupyter notebook that documents the application of a Support Vector Machine (SVM) model using the entire dataset.
• SVM_sampled_features.ipynb: A Jupyter notebook that documents the application of a Support Vector Machine (SVM) model using a selected subset of features for optimized performance.
• train.csv: The training dataset containing features and target labels for building the model.
• test.csv: The test dataset used for evaluating the model's performance.
• test_target.csv: The target labels corresponding to the test dataset for validation purposes.

More models and analysis files will be added in the future as the project progresses.

The dataset used for this project contains various attributes related to environmental factors such as:

• Soil characteristics.
• Climate conditions.
• Topographical information.

These features are used to classify different types of forest cover, making the dataset a valuable resource for testing various classification algorithms.

1. Data Exploration: Perform an initial analysis of the dataset to understand the distributions, correlations, and relevance of features.
2. Feature Engineering: Identify the most impactful features and engineer new features if necessary.
3. Model Development: Implement a range of machine learning models, such as:
   • Support Vector Machine (SVM)
   • Random Forests
4. Model Evaluation: Evaluate the models using performance metrics like accuracy, precision, recall, and F1 score.
5. Model Optimization: Fine-tune the models for better performance and compare the results.
6. Future Work: Integrate more advanced models and explore ensemble methods for improved accuracy.

To run the notebooks and replicate the results, you need the following software and libraries installed:

• Python 3.8 or higher
• Jupyter Notebook
• pandas
• numpy
• scikit-learn
• matplotlib
• seaborn

Clone the repository to your local machine:

git clone https://github.com/ChaimaBalti/forest-cover-type-prediction

Install the required Python packages:

pip install -r requirements.txt

• Chaima Balti
• Roukaya Lakhzouri
• Salsabil Rouahi

Supervised by Professor Moez Ben Haj Hmida.

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

• ENIT and Moez Ben Haj Hmida for providing the resources and guidance.
• Kaggle for the Forest Cover Type Prediction Dataset.