SUBHRANILGHOSH07/ALGERIAN-FOREST-FIRES

# 🌲 Algerian Forest Fires – Data Cleaning, EDA & Machine Learning Prediction ### **📍 Dataset Overview** The **Algerian Forest Fires Dataset** contains meteorological and fire-weather measurements collected from two regions in Algeria: * **Bejaia** (Northeast Algeria) * **Sidi Bel-Abbès** (Northwest Algeria) Each region contributes **122 instances**, giving a total of **244 samples**. ### **📅 Time Period** Data was collected **from June 2012 to September 2012**, capturing the peak of the fire season. --- ### **📊 Dataset Composition** * **Total instances:** 244 * **Regions:** * 122 Bejaia * 122 Sidi Bel-Abbès * **Input attributes:** 11 * **Output attribute:** 1 (target class) * **Class distribution:** * **Fire:** 138 instances * **Not fire:** 106 instances --- ### **📑 Attributes Included** The 11 input features consist of: * Meteorological variables (Temperature, RH, Wind speed, Rain) * Fire Weather Index (FWI) system components: * FFMC, DMC, DC, ISI, BUI, FWI * Date-related fields (day, month, year) The **output attribute** is: * `Classes` → *fire* / *not fire* --- If you'd like, I can integrate this directly into the previously generated README or format it as a standalone documentation section. --- ## 📚 **Project Overview** Forest fires are a major environmental concern in Northern Algeria, particularly in the Bejaia and Sidi Bel-Abbes regions. This project aims to: * Understand climate and environmental factors influencing fires * Build models to predict **fire occurrence** or **FWI (Fire Weather Index)** * Provide reproducible data cleaning and modeling pipelines * Save trained models for future integration into apps or APIs --- ## 📁 **Repository Structure** ``` │ ├── Algerian_forest_fires_dataset_cleaned_dataset.csv # Cleaned and structured dataset ├── Algerian_forest_fires_dataset_UPDATE.csv # Raw, original dataset (unprocessed) │ ├── cleaningdatasetandeda.ipynb # Data cleaning + EDA notebook ├── …