prajwal-sv/Algerian-Forest-Fires-Dataset---Project-

The dataset has 244 records from Algeria’s Bejaia (northeast) and Sidi Bel-abbes (northwest) regions, with 122 instances each collected between June–September 2012. It includes 11 attributes plus one output class, categorizing 138 cases as fire and 106 as not fire # Algerian-Forest-Fires-Dataset---Project- The dataset has 244 records from Algeria’s Bejaia (northeast) and Sidi Bel-abbes (northwest) regions, with 122 instances each collected between June–September 2012. It includes 11 attributes plus one output class, categorizing 138 cases as fire and 106 as not fire # 🌲 Algerian Forest Fires Prediction ## 📌 Project Overview This repository contains a machine learning project focused on predicting the **Fire Weather Index (FWI)**, a crucial component for forest fire risk assessment. The study covers two regions in Algeria: **Bejaia** and **Sidi Bel-abbes**, using data collected between June and September 2012. ## 🗂️ Dataset Information The dataset consists of 244 instances. It includes meteorological observations and calculated FWI system indices. ### Attribute Information: | # | Attribute | Description | Range / Unit | |:---:|:---|:---|:---| | 1 | **Date** | Day, month, and year of observation | DD/MM/YYYY | | 2 | **Temp** | Noon temperature (Max temperature) | 22 to 42 °C | | 3 | **RH** | Relative Humidity | 21 to 90 % | | 4 | **Ws** | Wind speed | 6 to 29 km/h | | 5 | **Rain** | Total daily rainfall | 0 to 16.8 mm | | 6 | **FFMC** | Fine Fuel Moisture Code (Surface litter) | 28.6 to 92.5 | | 7 | **DMC** | Duff Moisture Code (Shallow organic layers) | 1.1 to 65.9 | | 8 | **DC** | Drought Code (Deep organic layers) | 7 to 220.4 | | 9 | **ISI** | Initial Spread Index (Fire spread rate) | 0 to 18.5 | | 10 | **BUI** | Buildup Index (Total fuel available) | 1.1 to 68 | | 11 | **FWI** | Fire Weather Index (**Target Variable**) | 0 to 31.1 | | 12 | **Classes** | Current status: Fire or Not Fire | Categorical | --- ## 🚀 Research & Methodology ### 1. Data Preprocessing * **Cleaning:** Handled missing values and stripped inconsistent string formatting from column names. * **Encoding:** Converted the 'Classes' feature (fire/not fire) into binary labels. * **Regional Split:** Handled the two distinct datasets for Bejaia an …