Machine Learning in Magnetospheric Physics Time Series Data

No Thumbnail Available

Files

URL

Journal Title

Journal ISSN

Volume Title

Perustieteiden korkeakoulu | Bachelor's thesis
Electronic archive copy is available locally at the Harald Herlin Learning Centre. The staff of Aalto University has access to the electronic bachelor's theses by logging into Aaltodoc with their personal Aalto user ID. Read more about the availability of the bachelor's theses.
Unless otherwise stated, all rights belong to the author. You may download, display and print this publication for Your own personal use. Commercial use is prohibited.

Authors

Date

2024-04-26

Department

Major/Subject

Mcode

SCI3095

Degree programme

Aalto Bachelor’s Programme in Science and Technology

Language

en

Pages

29

Series

Abstract

Magnetospheric physics today relies heavily on time series data describing the complex dynamics between different magnetic regions and the particle populations within them. Attributed to the availability of large volumes of time series data, the process of manually labelling boundary crossings by hand has shifted to machine learning induced automation. However, few implementations of intricate machine learning based boundary crossing detection methods exist. Two recent studies have applied threshold, deep learning and unsupervised learning models trained on time series data collected by the Cassini probe as it was exploring the magnetosphere of Saturn. In this study a comprehensive comparison between the undertaken approaches is conducted, resulting in an evaluation of 8 total models. Taking into consideration the application context of each model identified by the authors, the best performing model is based on Matrix Profile, scoring a recall value of 63.6% at a precision above 80%. High runtimes, low F1-scores and overfitting models limit current machine learning implementations from being applied locally on-board future space missions. Henceforth, future machine learning model improvements and/or implementations for boundary crossing detection should aim to fully leverage the extensive data available, constructing training sets spanning all years while maintaining computational efficiency.

Description

Supervisor

Korpi-Lagg, Maarit

Thesis advisor

Weigt, Dale

Keywords

time esries analysis, machine learning, deep learning, unsupervised learning, magnetosphere, boundary crossing

Other note

Citation

Permanent link to this item

https://urn.fi/URN:NBN:fi:aalto-202405283908

Collections

[kand] Perustieteiden korkeakoulu / SCI