vamsikrishnapvkr/algerian-forest-fires

# Algerian Forest Fires Prediction using Ridge Regression ## Project Overview The objective of this project is to analyze the Algerian Forest Fires Dataset and build a Machine Learning model to predict the **Fire Weather Index (FWI)**. The dataset contains meteorological and fire-related observations collected from two Algerian regions: - **Bejaia Region** - **Sidi Bel-Abbes Region** This project includes: - Data Cleaning and Preprocessing - Exploratory Data Analysis (EDA) - Feature Engineering (FE) - Correlation Analysis and Feature Selection - Model Training using Ridge Regression - Model Optimization using RidgeCV - Saving trained artifacts for deployment --- # Dataset Information The dataset contains weather and fire-related attributes such as: - Temperature - Relative Humidity (RH) - Wind Speed (Ws) - Rain - FFMC - DMC - DC - ISI - BUI - FWI - Classes (Fire / Not Fire) ## Target Variable - **FWI (Fire Weather Index)** --- # Project Workflow ## 1. Data Cleaning and Preprocessing - Removed unnecessary columns - Handled missing values - Converted categorical variables into numerical format - Changed data types where required - Applied Standard Scaling before model training --- ## 2. Exploratory Data Analysis (EDA) ### Fire Distribution - Dataset contains both: - Fire - Not Fire classes - Fire occurrences are slightly higher in the **Sidi Bel-Abbes** region. ### Seasonal Trend Most forest fires occurred during: - June - July - August August recorded the highest number of fires. ### Correlation Analysis Important positively correlated features with **FWI**: - Temperature - FFMC - DMC - ISI Features with correlation higher than **0.85** were removed to reduce multicollinearity. ### Outlier Analysis - Boxplot analysis showed some outliers in the FWI feature. - Standard Scaling helped normalize feature distributions before training. --- # Model Training ## Ridge Regression Ridge Regression was selected to handle multicollinearity and i …