Time Series Forecasting Using Long Short-Term Memory Neural Networks: A Case Study of Seismogram

Hilal Nuha, Mohamed Mohandes*, Bo Liu, Ali Al-Shaikhi

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review

Abstract

This paper presents an investigation of the long short-term memory (LSTM) neural networks for seismogram time series prediction. LSTM has been widely used for time series prediction problems and achieved excellent results therefore it is interesting to utilize this technique for seismology. A seismogram from Albuquerque, New Mexico (Anmo), USA provided by the IRIS website is used for the experiment. The seismogram is recorded to observe passive seismic activities like Earthquake. The data is recorded with a 16 kHz sampling frequency by a station in Albuquerque. The network is with 200 hidden units and trained with 250 maximum iterations. The LSTM is able to achieve 0.00306 of root mean squared error (RMSE).

Original languageEnglish
Title of host publicationAdvances in Science, Technology and Innovation
PublisherSpringer Nature
Pages207-209
Number of pages3
DOIs
StatePublished - 2022

Publication series

NameAdvances in Science, Technology and Innovation
ISSN (Print)2522-8714
ISSN (Electronic)2522-8722

Bibliographical note

Publisher Copyright:
© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Keywords

  • Long short-term memory
  • Seismogram
  • Time series forecasting

ASJC Scopus subject areas

  • Architecture
  • Environmental Chemistry
  • Renewable Energy, Sustainability and the Environment

Fingerprint

Dive into the research topics of 'Time Series Forecasting Using Long Short-Term Memory Neural Networks: A Case Study of Seismogram'. Together they form a unique fingerprint.

Cite this