• Journal of the Korean Society for Precision Engineering
• /
• v.1 no.1
• /
• pp.34-49
• /
• 1984

The stability of machine tool systems is analyzed by considering the machining process as a stochastic process without decomposing into machine tool structural dynamics and cutting processes. In doing so the time series analysis technique developed by Wu and Pandit is applied systematically to the relative vibration between cutting tool and work- piece measured under actual working conditions. Various characteristic properties derived from the fitted ARMA(Autoregressive Moving Average) Models and those from raw data directly are investigated in relation with the system stability. Both damping ratio and absolute value of the characteristic roots of the AR part of the most significant dynamic mode are preferred as stability indicating factors to the other pro-perties such as theoretical variance .gamma. (o) or absolute power of the most dominant dynamic mode. Maximum aplitude during a certain interval and variance estimated from raw data are shown to be very sensi- tive to the type of the signal and the location of measurement point although they can be obtained rather easily. The relative vibration signal is also analyzed by FFT(Fast Fourier Transform) Analyzer for the purpose of comparison with the spectrums derived from the fitted ARMA models.

• 제어로봇시스템학회:학술대회논문집
• /
• 2003.10a
• /
• pp.402-407
• /
• 2003

• Communications for Statistical Applications and Methods
• /
• v.7 no.3
• /
• pp.837-843
• /
• 2000

Kreiss and Franke(192) and Allen and Datta(1999) proposed bootstrapping the M-estimators in ARMA models. In this paper, we introduce the robust estimating function and investigate the bootstrap approximations of the M-estimators which are solutions of the estimating equations in TAR models. A number of simulation results are presented to estimate the sampling distribution of the M-estimators, and asymptotic validity of the bootstrap for the M-estimators is established.

• The Korean Journal of Applied Statistics
• /
• v.20 no.3
• /
• pp.541-550
• /
• 2007

One of main aspects of time series analysis is to forecast future values of series based on values up to a given time. The prediction interval for future values is usually obtained under the normality assumption. When the assumption is seriously violated, a transformation of data may permit the valid use of the normal theory. We investigate the prediction problem for future values in the original scale when transformations are applied in ARMA models. In this paper, we introduce the methodology based on Yeo-Johnson transformation to solve the problem of skewed data whose modelling is relatively difficult in the analysis of time series. Simulation studies show that the coverage probabilities of proposed intervals are closer to the nominal level than those of usual intervals.

• The Korean Journal of Applied Statistics
• /
• v.26 no.6
• /
• pp.1053-1061
• /
• 2013

In this paper, we propose the time series forecasting models for internet traffic with long memory and heteroscedasticity. To control and forecast traffic volume, we first introduce the traffic forecasting models which are determined by the volatility and heteroscedasticity of the traffic. We then analyze and predict the heteroscedasticity and the long memory properties for forecasting traffic volume. Depending on the characteristics of the traffic, Fractional ARIMA model, Fractional ARIMA-GARCH model are applied and compared with the MAPE(Mean Absolute Percentage Error) Criterion.

• Journal of Biomedical Engineering Research
• /
• v.24 no.3
• /
• pp.151-158
• /
• 2003

Numerous studies of short-term, beat-to-beat variability in cardiovascular signals have used linear analysis techniques. However, no study has been done about the appropriateness of linear techniques or the comparison between linearities and nonlinearities in short-term, beat-to-beat variability. This paper aims to verify the appropriateness of linear techniques by investigating nonlinearities in short-term, beat-to-beat variability. We compared linear autoregressive moving average(ARMA) with nonlinear neural network(NN) models for predicting current instantaneous heart rate(HR) and mean arterial blood pressure(BP) from past HRs and BPs. To evaluate these models. we used HR and BP time series from the MIMIC database. Experimental results indicate that NN-based nonlinearities do not play a significant role and suggest that 10 technique provides adequate characterization of the system dynamics responsible for generating short-term, beat-to-beat variability.

• Journal of the Korean Statistical Society
• /
• v.35 no.1
• /
• pp.79-90
• /
• 2006

We consider the subdiagonal bilinear model and ARMA model with subdiagonal bilinear errors. Sufficient conditions for geometric ergodicity of associated Markov chains are derived by using results on generalized random coefficient autoregressive models and then strict stationarity and ,a-mixing property with exponential decay rates for given processes are obtained.

• Journal of applied mathematics & informatics
• /
• v.12 no.1_2
• /
• pp.67-79
• /
• 2003

An approximation for the bias in lag window estimators of the degree of differencing in fractionally integrated time series models is derived. The expression obtained is compared with the observed bias from simulations for various windows.

• The Korean Journal of Applied Statistics
• /
• v.21 no.3
• /
• pp.537-546
• /
• 2008

One of main goals of time series analysis is to estimate prediction of future values. In this paper, we investigate the bias problem when the transformation and back- transformation approach is applied in ARMA models and introduce a modified smearing estimation to reduce the bias. An empirical study on the returns of KOSDAQ index via Yeo-Johnson transformation was executed to compare the performance of existing methods and proposed methods and showed that proposed approaches provide a bias-reduced estimation of the prediction value.

• International Journal of Quality Innovation
• /
• v.3 no.2
• /
• pp.174-182
• /
• 2002

System identification and controller formulation are essential in dynamic process control. In system identification, data for system identification are obtained, and then they are analyzed so that the system model of the process is built, identified, and diagnosed. In controller formulation, the control equation is derived based on the result of the system identification. There has been much theoretical research on system identification and controller formulation. These theories are very useful when they are appropriately applied. To our regret, however, these theories are not always effectively applied in practice because the engineers and the operators who manage the process often do not have the necessary understanding of required time series analysis methods. On the other hand, because of widespread use of statistical packages, system identification such as estimating ARMA models can be done with little understanding of time series analysis methods. Therefore, it might be said that the most theoretically difficult part in practice is the controller formulation. In this paper, lists of control equations are proposed as a useful tool for practitioners to use. The tool supports bridging the gap between theory and practice in dynamic process control. Also, for some models, the generalized control equations are obtained.