Society of Streams

Taipei, Taiwan

Society of Streams

Taipei, Taiwan

Time filter

Source Type

Pan Y.-H.,National Taiwan Ocean University | Wang C.,National Taiwan University | Lin W.-Y.,National Taiwan University | Wang Y.-H.,Society of Streams | And 2 more authors.
Proceedings of the Institution of Mechanical Engineers, Part B: Journal of Engineering Manufacture | Year: 2011

The conventional approach for online monitoring of a machine's operating condition is based on linear time-frequency analysis and is therefore limited by the point that the vibrations are non-linear and non-stationary in nature. This problem has been addressed by the proposal of the multiscale entropy (MSE) approach to non-linear time series analysis. This paper proposes an online feature extractor to allow the vibration signal of shafts experiencing different problems to be differentiated using the MSE approach. © 2011 Authors.


Pan Y.-H.,National Taiwan Ocean University | Lin W.-Y.,National Taiwan University | Wang Y.-H.,Society of Streams | Lee K.-T.,National Taiwan Ocean University
Journal of Marine Science and Technology | Year: 2011

Multi-scale entropy (MSE) is a measurement of a system's complexity. It has received a great deal of attention in recent years, and its effectiveness has been verified, and applied in a number of different fields. However, the algorithms proposed in past studies required O(N2), which represented a degree of execution time insufficient for on-line applications, or for applications with long-term correlations. In this study, we showed that the probability function in the entropy term could be transformed into an orthogonal range search in the field of computational geometry. We then developed an efficient new algorithm for computing multi-scale entropy. The execution time in the results of our experiments with electrocardiogram (ECG), electroencephalography (EEG), interbeat interval (RR), and mechanical and ecological signals showed a significant improvement from 10 to 70 times over that of conventional methods for N = 80,000. Because the execution time has been significant reduced, the new algorithm could be applied to online diagnosis, in the computation of MSE for long-term correlation of signal.


Liang S.-F.,National Cheng Kung University | Kuo C.-E.,NCKU | Hu Y.-H.,NCKU | Pan Y.-H.,Ledder Technologies Holdings Ltd. | Wang Y.-H.,Society of Streams
IEEE Transactions on Instrumentation and Measurement | Year: 2012

In this paper, we propose an automatic sleep-scoring method combining multiscale entropy (MSE) and autoregressive (AR) models for single-channel EEG and to assess the performance of the method comparatively with manual scoring based on full polysomnograms. This is the first time that MSE has ever been applied to sleep scoring. All-night polysomnograms from 20 healthy individuals were scored using the Rechtschaffen and Kales rules. The developed method analyzed the EEG signals of C3-A2 for sleep staging. The results of automatic and manual scorings were compared on an epoch-by-epoch basis. A total of 8480 30-s sleep EEG epochs were measured and used for performance evaluation. The epoch-by-epoch comparison was made by classifying the EEG epochs into five states (Wake/REM/S1/S2/SWS) by the proposed method and manual scoring. The overall sensitivity and kappa coefficient of MSE alone are 76.9% and 0.65, respectively. Moreover, the overall sensitivity and kappa coefficient of our proposed method of integrating MSE, AR models, and a smoothing process can reach the sensitivity level of 88.1% and 0.81, respectively. Our results show that MSE is a useful and representative feature for sleep staging. It has high accuracy and good home-care applicability because a single EEG channel is used for sleep staging. © 2012 IEEE.


Pan Y.-H.,National Taiwan Ocean University | Wang Y.-H.,Society of Streams | Liang S.-F.,National Cheng Kung University | Lee K.-T.,National Taiwan Ocean University
Computer Methods and Programs in Biomedicine | Year: 2011

Both sample entropy and approximate entropy are measurements of complexity. The two methods have received a great deal of attention in the last few years, and have been successfully verified and applied to biomedical applications and many others. However, the algorithms proposed in the literature require O(N 2) execution time, which is not fast enough for online applications and for applications with long data sets. To accelerate computation, the authors of the present paper have developed a new algorithm that reduces the computational time to O(N 3/2)) using O(N) storage. As biomedical data are often measured with integer-type data, the computation time can be further reduced to O(N) using O(N) storage. The execution times of the experimental results with ECG, EEG, RR, and DNA signals show a significant improvement of more than 100 times when compared with the conventional O(N 2) method for N=80,000 (N=length of the signal). Furthermore, an adaptive version of the new algorithm has been developed to speed up the computation for short data length. Experimental results show an improvement of more than 10 times when compared with the conventional method for N>4000. © 2010 Elsevier Ireland Ltd.


Pan Y.-H.,National Taiwan Ocean University | Lee K.-T.,National Taiwan Ocean University | Wang Y.-H.,Society of Streams
3CA 2010 - 2010 International Symposium on Computer, Communication, Control and Automation | Year: 2010

It is well known that an optimal algorithm for logarithm query time and linear storage for two dimensional orthogonal range searches is nonexistent except for very few cases in which certain computational models have been assumed. A new algorithm called sliding kd tree is presented in this article where the length of the query box is fixed in any direction. The new algorithm is optimal in handling two-dimensional problems, and can be applied to VLSI Design Automations presented to show the performance of the SKD algorithm. © 2010 IEEE.

Loading Society of Streams collaborators
Loading Society of Streams collaborators