Logo Lanfrica

sury53038/algerian_forest_fire

Domaine:

environment and energy

Type de record:

datasetproject
Créateur:
sur
HĂ´te:
Algerian forest fire prediction # 🔥 Algerian Forest Fire Prediction A machine learning project to predict the occurrence of forest fires in the Algerian region based on meteorological and environmental data. ## 📌 Overview Forest fires pose serious threats to biodiversity, air quality, and human life. This project utilizes the **Algerian Forest Fires Dataset** to build a classification model that can predict whether a fire will occur on a given day. The end goal is to aid in early fire detection and prevention efforts using data-driven techniques. --- ## 🧠 Technologies Used - Python - Pandas, NumPy - Matplotlib, Seaborn - Scikit-learn - Streamlit (for web app) - Jupyter Notebook --- ## 📊 Dataset Description The dataset includes observations from two Algerian regions: **Bejaia** and **Sidi Bel-Abbes**. Each data point contains: - Meteorological indices: FFMC, DMC, DC, ISI - Weather features: Temperature, RH (Relative Humidity), Wind, Rain - Target variable: `Fire` (Yes/No or 1/0) Source: UCI Machine Learning Repository --- ## 🚀 Workflow 1. **Data Preprocessing** - Merging regional data - Handling null/missing values - Encoding categorical features 2. **Exploratory Data Analysis (EDA)** - Feature correlation - Visualizations for trends and patterns 3. **Model Training** - Algorithms: Logistic Regression, Random Forest, SVM, etc. - Model tuning & cross-validation 4. **Evaluation** - Accuracy, Precision, Recall, F1-score - Confusion Matrix and ROC Curve 5. **Deployment** - Built a simple UI using Streamlit for predictions based on user input --- ## 🖥️ How to Run Locally 1. Clone the repo: ```bash git clone github.com cd algerian-forest-fire-prediction