• Journal of Intelligence and Information Systems
• /
• v.5 no.1
• /
• pp.103-123
• /
• 1999

Input filtering as a preprocessing method is so much crucial to get good performance in time series forecasting. There are a few preprocessing methods (i.e. ARMA outputs as time domain filters, and Fourier transform or wavelet transform as time-frequency domain filters) for handling time series. Specially, the time-frequency domain filters describe the fractal structure of financial markets better than the time domain filters due to theoretically additional frequency information. Therefore, we, first of all, try to describe and analyze specially some issues on the effectiveness of different filtering methods from viewpoint of the performance of a neural network based forecasting. And then we discuss about neural network model architecture issues, for example, what type of neural network learning architecture is selected for our time series forecasting, and what input size should be applied to a model. In this study an input selection problem is limited to a size selection of the lagged input variables. To solve this problem, we simulate on analyzing and comparing a few neural networks having different model architecture and also use an embedding dimension measure as chaotic time series analysis or nonlinear dynamic analysis to reduce the dimensionality (i.e. the size of time delayed input variables) of the models. Throughout our study, experiments for integration methods of joint time-frequency analysis and neural network techniques are applied to a case study of daily Korean won / U. S dollar exchange returns and finally we suggest an integration framework for future research from our experimental results.

• Proceedings of the KIEE Conference
• /
• 2003.07b
• /
• pp.1072-1074
• /
• 2003

In this paper, a new control strategy of a series active power filter using direct compensating voltage extraction method is proposed. This control algorithm compensates harmonics, reactive power and neutral line currents which are generated by balanced or unbalanced nonlinear loads. The advantage of this method is that the compensating voltage of the series active power filter can be extracted without phase transformation. Therefore, calculation time is shorten and the control method is simple compared with conventional method as the p-q theory In addition, this control strategy was applied for the series active power filter in 3-phase 4-wire system which is widely employed in distributing electric energy to several office building and manufacturing plants. Some results obtained from the experimental model using the proposed method are presented to demonstrate and confirm its validity.

• Korean Journal of Fisheries and Aquatic Sciences
• /
• v.35 no.3
• /
• pp.216-222
• /
• 2002

Autocorrelation in time series data can affect statistical inference in correlation or regression analyses. To improve a regression model from which the residuals are autocorrelated, Yule-Walker method, nonlinear least squares estimation, maximum likelihood method and 'prewhitening' method have been used to estimate the parameters in a regression equation. This study reviewed on the estimation methods of preventing spurious correlation in the presence of autocorrelation and applied the former three methods, Yule-Walker, nonlinear least squares and maximum likelihood method, to a 20-year real data set. Monte carlo simulation was used to compare the three parameter estimation methods. However, the simulation results showed that the mean squared error distributions from the three methods simulated do not differ significantly.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.19 no.5
• /
• pp.689-694
• /
• 2009

In this paper an approach to building fuzzy models for predicting daily and weekly stock prices is presented. Predicting stock prices with traditional time series analysis has proven to be difficult. Fuzzy logic based models have advantage of expressing the input-output relation linguistically, which facilitates the understanding of the system behavior. In building a stock prediction model we bear a burden of selecting most effective indicators for the stock prediction. In this paper information used in traditional candle stick-chart analysis is considered as input variables of our fuzzy models. The fuzzy rules have the premises and the consequents composed of trapezoidal membership functions and nonlinear equations, respectively. DE(Differential Evolution) identifies optimal fuzzy rules through an evolutionary process. The fuzzy models to predict daily and weekly open, high, low, and close prices of KOSPI(KOrea composite Stock Price Index) are built, and their performances are demonstrated.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.30 no.2 s.245
• /
• pp.194-201
• /
• 2006

Gait analysis is essential to identify accurate cause and knee condition from patients who display abnormal walking. Traditional linear tools can, however, mask the true structure of motor variability, since biomechanical data from a few strides during the gait have limitation to understanding the system. Therefore, it is necessary to propose a more precise dynamic method. The chaos analysis, a nonlinear technique, focuses on understand how variations in the gait pattern change over time. Eight healthy eight subjects walked on a treadmill for 100 seconds at 60 Hz. Three dimensional walking kinematic data were obtained using two cameras and KWON3D motion analyzer. The largest Lyapunov exponent from the measured knee angular displacement time series was calculated to quantify local stability. This study quantified the variability present in time series generated from gait parameter via chaos analysis. Knee flexion-extension patterns were found to be chaotic. The proposed Lyapunov exponent can be used in rehabilitation and diagnosis of recoverable patients.

• Journal of Information Processing Systems
• /
• v.16 no.2
• /
• pp.508-521
• /
• 2020

We compared a novel encoding Lempel-Ziv complexity (ELZC) with three common complexity algorithms i.e., approximate entropy (ApEn), sample entropy (SampEn), and classic Lempel-Ziv complexity (CLZC) so as to determine a satisfied complexity and its corresponding quality indices for assessing quality of multi-lead electrocardiogram (ECG). First, we calculated the aforementioned algorithms on six artificial time series in order to compare their performance in terms of discerning randomness and the inherent irregularity within time series. Then, for analyzing sensitivity of the algorithms to content level of different noises within the ECG, we investigated their change trend in five artificial synthetic noisy ECGs containing different noises at several signal noise ratios. Finally, three quality indices based on the ELZC of the multi-lead ECG were proposed to assess the quality of 862 real 12-lead ECGs from the MIT databases. The results showed the ELZC could discern randomness and the inherent irregularity within six artificial time series, and also reflect content level of different noises within five artificial synthetic ECGs. The results indicated the AUCs of three quality indices of the ELZC had statistical significance (>0.500). The ELZC and its corresponding three indices were more suitable for multi-lead ECG quality assessment than the other three algorithms.

• The Korean Journal of Applied Statistics
• /
• v.22 no.3
• /
• pp.499-513
• /
• 2009

An economic signal in the real world usually reflects complex phenomena. One may have difficulty both extracting and interpreting information embedded in such a signal. A natural way to reduce complexity is to decompose the original signal into several simple components, and then analyze each component. Spectral analysis (Priestley, 1981) provides a tool to analyze such signals under the assumption that the time series is stationary. However when the signal is subject to non-stationary and nonlinear characteristics such as amplitude and frequency modulation along time scale, spectral analysis is not suitable. Huang et al. (1998b, 1999) proposed a data-adaptive decomposition method called empirical mode decomposition and then applied Hilbert spectral analysis to decomposed signals called intrinsic mode function. Huang et al. (1998b, 1999) named this two step procedure the Hilbert-Huang transform(HHT). Because of its robustness in the presence of nonlinearity and non-stationarity, HHT has been used in various fields. In this paper, we discuss the applications of the HHT and demonstrate its promising potential for non-stationary financial time series data provided through a Korean stock price index.

• Smart Structures and Systems
• /
• v.11 no.3
• /
• pp.299-314
• /
• 2013

This paper presents a nonlinear time series analysis technique for evaluating machine defect severity, based on the Recurrence Plot (RP) entropy. The RP entropy is calculated from the probability distribution of the diagonal line length in the recurrence plot, which graphically depicts a system's dynamics and provides a global picture of the autocorrelation in a time series over all available time-scales. Results of experimental studies conducted on a spindle-bearing test bed have demonstrated that, as the working condition of the bearing deteriorates due to the initiation and/or progression of structural damages, the frequency information contained in the vibration signal becomes increasingly complex, leading to the increase of the RP entropy. As a result, RP entropy can serve as an effective indicator for defect severity assessment of rolling bearings.

• Ocean Systems Engineering
• /
• v.7 no.4
• /
• pp.413-433
• /
• 2017

The dynamic and structural responses of a 1000-m long circular submerged floating tunnel (SFT) with both ends fixed under survival irregular-wave excitations are investigated. The floater-mooring nonlinear and elastic coupled dynamics are modeled by a time-domain numerical simulation program, OrcaFlex. Two configurations of mooring lines i.e., vertical mooring (VM) and inclined mooring (IM), and four different buoyancy-weight ratios (BWRs) are selected to compare their global performances. The result of modal analysis is included to investigate the role of the respective natural frequencies and elastic modes. The effects of various submergence depths are also checked. The envelopes of the maximum/minimum horizontal and vertical responses, accelerations, mooring tensions, and shear forces/bending moments of the entire SFT along the longitudinal direction are obtained. In addition, at the mid-section, the time series and the corresponding spectra of those parameters are also presented and analyzed. The pros and cons of the two mooring shapes and high or low BWR values are systematically analyzed and discussed. It is demonstrated that the time-domain numerical simulation of the real system including nonlinear hydro-elastic dynamics coupled with nonlinear mooring dynamics is a good method to determine various design parameters.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.26 no.6
• /
• pp.71-78
• /
• 1989

The imput-output relations for nonlinear systems can be explicitly represented by the Volterra series and they can be characterized by the Volterra kernels. This study is concerned with modeling an elastomer force-displacement relation due to step inputs by utilizing the truncated Volterra series. Since it is practically impossible to apply step inputs that have infinite slope at zero time, the loads due to constant penetration(displacement) rate followed by constant penetration inputs are measured as an alternative approach and estimated for step inputs and then utilized for the truncated Volterra series models. One second order and one third order truncated Volterra series models have been employed to model the force-displacement relation which is one of the prominent properties to characterize the viscoelastic material. The third order truncated Volterra series model has better results, compared with those of the second order truncated Volterra series model.