Conditional restricted Boltzmann machine as a generative model for body-worn sensor signals

Creative Commons License

Karakus E., Köse H.

IET SIGNAL PROCESSING, vol.14, no.10, pp.725-736, 2020 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 14 Issue: 10
  • Publication Date: 2020
  • Doi Number: 10.1049/iet-spr.2020.0154
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, PASCAL, Aerospace Database, Applied Science & Technology Source, Communication Abstracts, Compendex, Computer & Applied Sciences, INSPEC, Metadex, Civil Engineering Abstracts
  • Page Numbers: pp.725-736
  • Keywords: statistical distributions, statistical analysis, feature extraction, Boltzmann machines, medical signal processing, signal classification, time-domain analysis, frequency-domain analysis, deep learning (artificial intelligence), graph theory, random processes, time series, classification models, frequency domain features, classification accuracy, deep learning-based classification techniques, energy-based probabilistic graphical model, conditional restricted Boltzmann machine, time-series signals, signal features, generative model training, Wasserstein generative adversarial network, CRBM model, body-worn sensor signals, sensor-based human activity recognition problem, frequency domain feature extraction, signal processing techniques, time domain features, random variable, binary probability distribution, RBM, gradient penalty, performance criterion, statistical analysis
  • Istanbul Technical University Affiliated: Yes


Sensor-based human activity classification requires time and frequency domain feature extraction techniques. The set of choice in time and frequency domain features may have a significant impact on the overall classification accuracy. Another problem is to train deep learning models with sufficient dataset. The use of generative models eliminates the requirement of choosing certain features of the signal. As a generative model, restricted Boltzmann machine (RBM) is an energy-based probabilistic graphical model which factorises the probability distribution of a random variable over a binary probability distribution. Conditional restricted Boltzmann machines (CRBMs) is an extension to RBM, which can capture temporal information in time-series signals and can be deployed as a generative model in classification. In this study, the authors show how CRBMs can be trained to learn signal features. They present four generative model training results, RBM, CRBM, generative adversarial network, Wasserstein generative adversarial network - gradient penalty and compare the models' performances with a performance criterion. They show that the CRBM model can generate signals closest to true signals with a significantly higher success rate as compared to other presented generative models. They present a statistical analysis of the findings and show that the findings significantly hold.