Potentially Hazardous Asteroid Prediction using Adaboost over Linear Regression

Abstract

Asteroids, rocky objects orbiting the sun, have been a key focus of scientific study as they can supply insights into planet formation. With an infinite number of asteroids in space, the possibility of one colliding with our planet and leading to devastating effects constantly looms large. Asteroids that could come in proximity or collide with earth are classified as potentially hazardous asteroids, PHA (NASA, n.d.). However, it becomes cumbersome for humans to manually analyse large datasets for finding all the dangerous asteroids. Thus, machine learning techniques are ideal to study trends and make predictions. Machine learning is a method of data analysis based on computer algorithms that model relationships and improve our ability to analyse asteroid threats. The goal of this study was to train multiple machine learning models on physical and orbital asteroid features and find the model that most accurately classified the asteroids as hazardous or non-hazardous. This project falls under the domain of Supervised Machine Learning. Supervised Learning can be further divided into two parts namely classification and regression. We are going to use classification here since we can find factors that can affect nature of asteroid and will be able to predict it using those factors. Firstly, we are going to clean the dataset by removing some irrelevant columns. All the dataset having different datatype will be converted to a single datatype or can be removed. We will code on the filtered dataset. Lastly, we will try different machine learning models and will print the accuracy and the confusion matrix.