ridham1102/Algerian-Forest-Fires-Analysis

# Algerian Forest Fires Analysis This repository presents a comprehensive study of the Algerian Forest Fires dataset, with a focus on predicting the Fire Weather Index (FWI). The analysis explores data ingestion, preprocessing techniques, feature selection, and the application of both traditional and regularized linear regression models. ## Project Objective The principal aim of this work is to develop reliable regression models that can estimate the Fire Weather Index (FWI) using meteorological and environmental factors. A comparative assessment is conducted to understand the effects of multicollinearity and the benefits of regularization methods in improving predictive accuracy. ## Dataset Description The dataset encompasses observations from two distinct regions in Algeria, including measurements of temperature, relative humidity, wind speed, rainfall, and other indices related to fire weather conditions. The target variable, FWI, quantifies the potential severity of forest fires. ## Analysis Workflow ### Data Acquisition and Exploration The raw data is loaded and subject to initial inspection to verify structure, data types, and completeness. Descriptive statistics and visualizations are generated to identify patterns, outliers, and missing values. ### Data Cleaning and Preparation Column names are standardized by removing leading or trailing whitespace. Irrelevant temporal attributes, such as day, month, and year, are dropped. Categorical fields, notably the fire classification label, are encoded into numerical representations to facilitate model training. ### Feature Selection A correlation matrix is computed to pinpoint highly correlated predictors. Variables exhibiting excessive multicollinearity are removed to enhance model generalization and interpretability. ### Data Scaling Numerical features are scaled using a standardization technique to ensure that all predictors share a common scale, which is particularly critical for models incorporati …