Forest Fire Prediction using Machine Learning

Forest fires have been some of the oldest threats to our nature. A forest fire destroys not only the flora and fauna, but also disrupts the ecological cycle of the area.Wildfires can be awfully threatening,as they cause of loss of not only the trees and the green cover of the area but also result in loss of wildlife, charred, damaged soil, loss of houses and other structures and smoke and respiratory diseases. Forest fires are being a common occurrence these days, like the (2019) Amazon Rainforest wildfires, Brazil; (2020) Australia Bushfires, (resulted in loss of 46.03 million acres) etc have been ravaging the lands of our planet causing huge losses.

Problem Statement & Objectives

Forest Fires have been a major contributing factor in the loss of our forests ecosystem.

• Thus, early prediction of forest fires is of paramount importance to prevent huge losses to our eco-system , human and wild life and habitat.

• Also, the available methods which give us predictive results for the same are not available or outdated.

Algorithms and techniques used:

1. Regression: It is the set of processes for estimating the relationships between different variables used in the analysis. It focuses on the relationships between one dependent variable and one or more independent variables.

   • Linear Regression(Supervised Learning/Regression): The simplest form of regression, linear regression is used to understand the relationship between two continuous variables. It involves finding the line, that most closely fits the data according to a specific mathematical criterion.

   • Logistic Regression (Supervised Learning/Regression) : Logistic Regression is a machine learning method used for modeling a binary dependent variable. It is a form of binomial regression.The dependent variable takes a binary form – 1 or 0, yes or no. The relationship between the dependent variable and the independent variable helps it to predict the target variable. It uses sigmoid function to determine their probability and map them to some discrete values.

2. Classification : Classification is a task that requires the use of machine learning algorithms that learn how to assign a class label to examples from the problem domain. It specifies the class to which data elements belong to and is best used when the output has finite and discrete values. It predicts a class for an input variable as well.

   • Support Vector Machine Algorithm (SVM) (Supervised Learning): The objective of the support vector machine algorithm is to find a hyperplane in an N-dimensionalspace(N — the number of features) that distinctly classifies the data points.It can be used for both regression and classification purposes.

   • K-Means(Unsupervised Learning/Clustering): that aims to partition n observations into k clusters in which each observation belongs to the cluster with the nearest mean (cluster centers or cluster centroid), serving as a prototype of the cluster.