ashishrana1501 / forest-fire-prediction

Algerian Forest Fire Prediction

Forest Fire Prediction

Heroku App (https://forestfire-predictions.herokuapp.com)

Demonstration

Classification

Forestfire.Prediction.mp4

Regression

ForestFire.Regression.mp4

A brief description of what this project is all about.

Forest Fire Prediction is a Supervised Machine learning problem statements. Using Regression and Classification Algorithm, Regression and Classification Model is build that detected future fires based on certain Weather report.

A framework is created using Flask and deployed on Heroku

Library Used in this Project

Data Pre-Processing

• Numpy, Pandas, Matplotlib, Seaborn

Model Building

• Sklearn, statsmodels

Hyperparameter Tuning

• RandomizedSearchCV, GridSearchCV

Introduction

Algerian Forest Fires

Data set Available at: link text

Data Set Information:

• The dataset includes 244 instances that regroup a data of two regions of Algeria,namely the

• Bejaia region located in the northeast of Algeria and the Sidi Bel-abbes region located in the northwest of Algeria.

• 122 instances for each region.

• The period from June 2012 to September 2012.

• The dataset includes 11 attribues and 1 output attribue (class)

• The 244 instances have been classified into fire (138 classes) and not fire (106 classes) classes.

Attribute Information:

1. Date : (DD/MM/YYYY) Day, month ('june' to 'september'), year (2012)

Weather data observations

2. Temp : temperature noon (temperature max) in Celsius degrees: 22 to 42

3. RH : Relative Humidity in %: 21 to 90

4. Ws : Wind speed in km/h: 6 to 29

5. Rain: total day in mm: 0 to 16.8

FWI Components

6. Fine Fuel Moisture Code (FFMC) index from the FWI system: 28.6 to 92.5

7. Duff Moisture Code (DMC) index from the FWI system: 1.1 to 65.9

8. Drought Code (DC) index from the FWI system: 7 to 220.4

9. Initial Spread Index (ISI) index from the FWI system: 0 to 18.5

10. Buildup Index (BUI) index from the FWI system: 1.1 to 68

11. Fire Weather Index (FWI) Index: 0 to 31.1

12. Classes: two classes, namely Fire and not Fire

Steps

1. Data Collection
2. Data Pre-Processing
3. Exploratory Data Analysis
4. Feature Engineering
5. Feature Selection
6. Model Building
7. Model Selection
8. Hyperparameter Tuning
9. Flask framework
10. Model deployment

Model Building

Regression

• For regression analysis FWI(Fire weather Index) considered as dependent feature because it highly correlated with classes variable with more than 90% correlation.

Model Used:

1. Linear regression
2. Lasso Regression
3. Ridge Regression
4. Decision tree
5. Random forest
6. K-Nearest Neighbour regressor
7. Support Vector Regressor

Classification

• For Classification Classes is dependent feature which is a Binary Classification(fire, not fire)

Model Used:

1. Logistic Regression
2. Decision Tree
3. Random Forest
4. K-Nearest Neighbour
5. XGboost.

Model Selection

HyperParameter Tuning performed using RandomizedsearchCV for the model which performed best for both Regression and Classification.

• For Regression r2_score metrics is used to select best model.

• For Classification Stratified Kfold Cross-Validation metrics is used.

• The best Mean CV Accuracy Model is used for Model Deployment.

Model Deployment

Flask

• framework is created using Flask.

Heroku

• Model deployed on Heroku.

🔗 Links

forest-fire-prediction's Issues

Classification model produces regression output due to model-saving error

Description:
Expected Behavior:
The classification model should output class probabilities or labels.

Actual Behavior:
The model is producing continuous values, indicating a regression output.