Javascript must be enabled for the correct page display

Discovering exoplanets using Convolutional Neural Networks

van Amerongen, Philippe (2018) Discovering exoplanets using Convolutional Neural Networks. Bachelor's Thesis, Astronomy.

[img]
Preview
Text
thesis.pdf

Download (1MB) | Preview
[img] Text
toestemming.pdf
Restricted to Registered users only

Download (94kB)

Abstract

The NASA Kepler Space mission had as main objective to discover ex- trasolar planets by observing over 300,000 stars (Barentsen, 2018). This results in light curves of those 300,000 stars, which need to be analyzed individually for presence of exoplanets. With the upcoming TESS mis- sion more than 2,000,000 stars will be observed (Ricker et al., 2010). In order to efficiently analyze all that data this thesis puts forth an ap- proach to automate this analysis using a machine learning algorithm, a convolutional neural network. The objective of this thesis is to ex- plore the possibilities and accuracy of such an algorithm in the field of exoplanet detection. There will be made use of a high leven neu- ral network framework called Keras, which is built upon the Tensor- Flow library. The code made for this thesis can be found at https: //github.com/phicoder/exoplanet-detection.

Item Type: Thesis (Bachelor's Thesis)
Supervisor name: Vogelaar, M.G.R.
Degree programme: Astronomy
Thesis type: Bachelor's Thesis
Language: English
Date Deposited: 12 Jul 2018
Last Modified: 18 Jul 2018 12:49
URI: https://fse.studenttheses.ub.rug.nl/id/eprint/17826

Actions (login required)

View Item View Item