ayushhh026/Algerian-Forest-Fire-Predictor
Domaine:
environment and energyType de record:
software
Créateur:
ayu
Hôte:
End-to-end ML regression pipeline predicting Forest Weather Index (FWI) using Ridge Regression • FastAPI • Deployed on AWS
# Algerian Forest Fire Predictor 🔥
```
███████╗ ██████╗ ██████╗ ███████╗███████╗████████╗ ███████╗██╗██████╗ ███████╗
██╔════╝██╔═══██╗██╔══██╗██╔════╝██╔════╝╚══██╔══╝ ██╔════╝██║██╔══██╗██╔════╝
█████╗ ██║ ██║██████╔╝█████╗ ███████╗ ██║ █████╗ ██║██████╔╝█████╗
██╔══╝ ██║ ██║██╔══██╗██╔══╝ ╚════██║ ██║ ██╔══╝ ██║██╔══██╗██╔══╝
██║ ╚██████╔╝██║ ██║███████╗███████║ ██║ ██║ ██║██║ ██║███████╗
╚═╝ ╚═════╝ ╚═╝ ╚═╝╚══════╝╚══════╝ ╚═╝ ╚═╝ ╚═╝╚═╝ ╚═╝╚══════╝
```
### End-to-End ML Regression Pipeline — From Raw Data to Cloud Deployment
> **EDA · Feature Selection · Regularization · StandardScaler · FastAPI · AWS Deployment**
---
## What is This Project?
A complete machine learning regression pipeline that predicts the **Fire Weather Index (FWI)** — a composite score used by meteorologists to assess wildfire danger — using the **Algerian Forest Fire Dataset**.
The project covers the full ML lifecycle:
- ✅ **Data Cleaning** — handling nulls, fixing dtypes, stripping whitespace, regional encoding
- ✅ **EDA** — correlation heatmaps, pairplots, boxplots, class distribution analysis
- ✅ **Feature Selection** — correlation-threshold-based multicollinearity removal
- ✅ **Regularization Benchmarking** — Linear, Lasso, Ridge, ElasticNet, LassoCV, RidgeCV compared
- ✅ **Deployment** — FastAPI REST API with HTML frontend, deployed on AWS
---
## Pipeline Overview
```
┌──────────────────────────────────────────────────────────────┐
│ RAW DATASET │
│ Algerian Forest Fire (2 Regions · 244 rows) │
└────────────────────────┬─────────────────────────────────────┘
│
▼
┌──────────────────────────────────────────────────────────────┐
│ DATA CLEANING │
│ Null removal · dtype fixes · region encoding · dedup │
└────────────────────────┬─────────────────────────────────────┘
│
▼
┌───── …