Markovian RNN: an adaptive time series prediction network with HMM-based switching for nonstationary environments

Simple item page
buir.contributor.authorİlhan, Fatih
buir.contributor.authorKaraahmetoğlu, Oğuzhan
buir.contributor.authorKozat , Süleyman Serdar
buir.contributor.orcidİlhan, Fatih|0000-0002-0173-7544
buir.contributor.orcidKaraahmetoğlu, Oğuzhan|0000-0002-0131-6782
buir.contributor.orcidKozat, Süleyman Serdar|0000-0002-6488-3848
dc.citation.epage728en_US
dc.citation.issueNumber2
dc.citation.spage715
dc.citation.volumeNumber34
dc.contributor.authorİlhan, Fatih
dc.contributor.authorKaraahmetoğlu, Oğuzhan
dc.contributor.authorBalaban, İ.
dc.contributor.authorKozat, Süleyman Serdar
dc.date.accessioned2024-03-19T06:17:51Z
dc.date.available2024-03-19T06:17:51Z
dc.date.issued2023-02-01
dc.description.abstractWe investigate nonlinear regression for nonstationary sequential data. In most real-life applications such as business domains including finance, retail, energy, and economy, time series data exhibit nonstationarity due to the temporally varying dynamics of the underlying system. We introduce a novel recurrent neural network (RNN) architecture, which adaptively switches between internal regimes in a Markovian way to model the nonstationary nature of the given data. Our model, Markovian RNN employs a hidden Markov model (HMM) for regime transitions, where each regime controls hidden state transitions of the recurrent cell independently. We jointly optimize the whole network in an end-to-end fashion. We demonstrate the significant performance gains compared to conventional methods such as Markov Switching ARIMA, RNN variants and recent statistical and deep learning-based methods through an extensive set of experiments with synthetic and real-life datasets. We also interpret the inferred parameters and regime belief values to analyze the underlying dynamics of the given sequences.
dc.description.provenanceMade available in DSpace on 2024-03-19T06:17:51Z (GMT). No. of bitstreams: 1 Markovian_RNN_an_adaptive_time_series_prediction_network_with_HMM-based_switching_for_nonstationary_environments.pdf: 3058724 bytes, checksum: 8b62458b1dafba7dceee0b80752c9acf (MD5) Previous issue date: 2023-02-01en
dc.identifier.doi10.1109/TNNLS.2021.3100528
dc.identifier.eissn2162-2388
dc.identifier.issn2162-237X
dc.identifier.urihttps://hdl.handle.net/11693/114924
dc.language.isoenglish
dc.publisherInstitute of Electrical and Electronics Engineers Inc.
dc.relation.isversionofhttps://dx.doi.org/10.1109/TNNLS.2021.3100528
dc.source.titleIEEE Transactions on Neural Networks and Learning Systems
dc.subjectHidden Markov models (HMMs)
dc.subjectNonlinear regression
dc.subjectNonstationarity
dc.subjectRecurrent neural networks (RNNs)
dc.subjectRegime switching
dc.subjectTime series prediction
dc.titleMarkovian RNN: an adaptive time series prediction network with HMM-based switching for nonstationary environments
dc.typeArticle

Files

Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
Markovian_RNN_An_adaptive_time_series_prediction_network_with_HMM-based_switching_for_nonstationary_environments.pdf
Size:
2.88 MB
Format:
Adobe Portable Document Format
Download

Collections

Scholarly Publications - Electrical and Electronics Engineering