Home // MMEDIA 2017, The Ninth International Conferences on Advances in Multimedia // View article


Analysis of Spatial and Temporal Features to Classify the Deep Moonquake Sources Using Balanced Random Forest

Authors:
Kodai Kato
Ryuhei Yamada
Yukio Yamamoto
Masaharu Hirota
Shohei Yokoyama
Hiroshi Ishikawa

Keywords: Planetary Science; Machine learning; Geophysical

Abstract:
In this paper, we evaluate other features different from the waveforms to classify seismic sources. Classification of sources of the deep moonquakes is an important issue for analyzing the focal mechanisms and the lunar deep structures. It was found that deep moonquakes that occur from the same source have similar waveforms. Some studies have been conducted to identify the deep moonquake sources using the waveform similarities. However, classifying some deep moonquakes using only the waveforms is difficult due to large noise and the small amplitude. If we could show that other features different from the waveforms are effective for classification of deep moonquakes, we can increase the number of classifiable moonquakes even if moonquakes include noise and small amplitude of the waveforms. Therefore, we use other features to classify deep moonquakes (position and velocity relative to the Earth, Sun, Jupiter, and Venus,as seen from the Moon.We apply these features to classify deep moonquakes that are not classified based on only waveforms, and it is useful to analyze the deep moonquake occurrence mechanisms. Our experiments showed that the position and velocity relation between the Moon and the Earth or Jupiter are effective for classification.

Pages: 51 to 56

Copyright: Copyright (c) IARIA, 2017

Publication date: April 23, 2017

Published in: conference

ISSN: 2308-4448

ISBN: 978-1-61208-548-7

Location: Venice, Italy

Dates: from April 23, 2017 to April 27, 2017