Back to search

Old Dominion University

Study of Deep Learning Models to Classify Nasa’s Kepler Light Curves

Abstract

dc:description.abstract

The search for exoplanets has been an ongoing effort since the first discoveries of planets beyond our solar system in the 1990s. Finding a potentially habitable planet outside our solar system could provide key insights on life elsewhere in the universe. NASA Missions such as the Kepler, launched in 2009 and completed in 2018, have provided a massive amount of data in this goal by using the transit method to discover repetitive and periodic dips in visible light around a star. The transit method has been used to measure flux, the brightness of a star over time. These flux time series are referred to as light curves, some of which may belong to that of an exoplanet. Machine learning is a powerful tool that can help identify complex patterns in large datasets and increase the efficiency of the exoplanet vetting process. Utilization of machine learning on confirming light curves that belong to exoplanets has already been conducted in different manners on several NASA Mission datasets, including Kepler. Projects such as Robovetter and Autovetter attempted to use classical methods (decision tree and random forest modeling, respectively) to recreate the manual decision process of classifying exoplanets. Apart from classical methods, deep learning has been used to automatically vet the light curve data. After phase folding the light curve so that the transit event is in the center, a Global View is created with a set amount of equally sized bins. In addition to the Global View, the Local View is created in a similar manner, except a Local View zooms into the transit event. The Global Views and Local Views are then passed into a one dimensional convolutional neural network (1D CNN). Newer solutions include a Transformer-based classification model applied to the TESS light curve dataset to distinguish planetary transits from false positives with success.

Degree

thesis:*
Name thesis:degree_name
Master of Science (MS)
Level thesis:degree_level
Thesis
Discipline thesis:degree_discipline
Computer Science

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Minnich, Heena
Contributors dc:contributor
  • Yaohang Li
  • Lusi Li
  • Sampath Jayarathna

Subjects

dc:subject × 2

Rights

dc:rights
Statement dc:rights
  • <p>In Copyright. URI: <a href="http://rightsstatements.org/vocab/InC/1.0/">http://rightsstatements.org/vocab/InC/1.0/</a> This Item is protected by copyright and/or related rights. You are free to use this Item in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights-holder(s).</p>

Identifiers

dc:identifier.*
Identifier
9798382770727

Chain of custody

source
Harvested from
Old Dominion University
Base URL
digitalcommons.odu.edu/do/oai/
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Minnich, Heena. Study of Deep Learning Models to Classify Nasa’s Kepler Light Curves. Thesis thesis, https://digitalcommons.odu.edu/computerscience_etds/172