TRAP@NCI

Machine Learning Approach to Classify Transit Signals and Assessing the Exoplanets Probability for Habitability

Baxi, Shreyas Shriram (2018) Machine Learning Approach to Classify Transit Signals and Assessing the Exoplanets Probability for Habitability. Masters thesis, Dublin, National College of Ireland.

[img]
Preview
PDF (Master of Science)
Download (1MB) | Preview

Abstract

There have been thousands of exoplanets which have been confirmed by the scientists. These identification of the exoplanets which were carried out using the transit method involved human intervention for analyzing the signals related to the exoplanets for manually classifying the signals.There have been past work which helped in automating this classification task with the help of machine learning algorithms.This paper utilizes Deep Learning and tries to classify the detected transit signals to be related to exoplanets or non-exoplanets using signal pre-processing techniques in order to study the impact of different pre-processing tasks on the performance parameters.The findings reveal that using Savitzky Golay Filter for the filtering purpose, a higher accuracy of the model is achieved as compared to a case where no pre-processing steps are taken and also better than the case where Gaussian filtering was used. Whereas using Gaussian filtering in the pre-processing stage along with normalization and standardizing steps better recall for the model was obtained as compared to other pre-processing tasks. In addition to that, this paper also utilizes the planetary characteristics related to the exoplanets to assess the probability of an exoplanet being habitable based on the characteristics of an Exoplanet using Naive Bayes algorithm. The probabilities obtained revealed some interesting insights about the habitability which have been discussed in the evaluation section. This study also utilizes Random Forest, KNN and SVM models for the purpose of classification of Exoplanets into Mesoplanets, Psychroplanets and Non habitable planets and all the models seem to fair similar in the performance aspects.

Item Type: Thesis (Masters)
Subjects: Q Science > QA Mathematics > Electronic computers. Computer science
T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science

Q Science > QA Mathematics > Computer software
T Technology > T Technology (General) > Information Technology > Computer software

Q Science > QB Astronomy
Divisions: School of Computing > Master of Science in Data Analytics
Depositing User: CAOIMHE NI MHAICIN
Date Deposited: 06 Nov 2018 10:26
Last Modified: 06 Nov 2018 10:26
URI: http://trap.ncirl.ie/id/eprint/3436

Actions (login required)

View Item View Item