• 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.

• Journal of Korea Water Resources Association
• /
• v.31 no.6
• /
• pp.769-777
• /
• 1998

In this study, various time series are analyzed to check nonlinearities of the data. The nonlinearity of a system can be investigated by testing the randomness of the time series data. To test the randomness, four nonparametric test statistics and a new test statistic, called the BDS statistic are used and the results and the results are compared. The Brock, Dechert, and Scheinkman (BDS) statistic is originated from the statistical properties of the correlation integral which is used for searching for chaos and has been shown very effective in distinguishing nonlinear structures in dynamic systems from random structures. As a result of application to linear and nonlinear models which are well known, the BDS statistic is found to be more effective than nonparametric test statistics in identifying nonlinear structure in the time series. Hydrologic time series data are fitted to ARMA type models and the statistics are applied to the residuals. The results show that the BDS statistic can distinguish chaotic nonlinearity from randomness and that the BDS statistic can also be used for verifying the validity of the fitted model.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.29 no.2B
• /
• pp.163-171
• /
• 2009

Classical linear models have been generally used to analyze and forecast hydrologic time series. However, there is growing evidence of nonlinear structure in natural phenomena and hydrologic time series associated with their patterns and fluctuations. Therefore, the classical linear techniques for time series analysis and forecasting may not be appropriate for nonlinear processes. In recent, the BDS (Brock-Dechert-Scheinkman) statistic instead of conventional techniques has been used for detecting nonlinearity of time series. The BDS statistic was derived from the statistical properties of the correlation integral which is used to analyze chaotic system and has been effectively used for distinguishing nonlinear structure in dynamic system from random structures. DVS (Deterministic Versus Stochastic) algorithm has been used for detecting chaos and stochastic systems and for forecasting of chaotic system. This study showed the DVS algorithm can be also used for detecting nonlinearity of the time series. In this study, the stochastic and hydrologic time series are analyzed to detect their nonlinearity. The linear and nonlinear stochastic time series generated from ARMA and TAR (Threshold Auto Regressive) models, a daily streamflow at St. Johns river near Cocoa, Florida, USA and Great Salt Lake Volume (GSL) data, Utah, USA are analyzed, daily inflow series of Soyang dam and the results are compared. The results showed the BDS statistic is a powerful tool for distinguishing between linearity and nonlinearity of the time series and DVS plot can be also effectively used for distinguishing the nonlinearity of the time series.

• Water for future
• /
• v.29 no.2
• /
• pp.179-190
• /
• 1996

The characteristics of water quality in the downstream of the Han River were analyzed by statistical techniques. Basic characteristics, areal and temporal variations, and correlations of water quality data were investigated. Monthly water quality data have been investigated systematically by exploring data analysis, including time series plot, summary statistics, distribution test, time dependence test, seasonality test and flow relatedness test. Results show that water quality data in this river have seasonality. And applicability of stochastic models such as Thomas-Fiering model and ARMA(1,1) model was identified. From the examination of water quality data related to discharge, it was found that DO and SS are sensitive to water temperature rather than discharge, while BOD and COD are sensitive to discharge at dry seasons. Seasonal periodicities were identified in all water quality variables from the cross correlation analysis.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.9 no.2
• /
• pp.88-99
• /
• 2007

Soil moisture is one of the essential components in determining surface hydrological processes such as infiltration, surface runoff as well as meteorological, ecological and water quality responses at watershed scale. This paper discusses soil moisture transfer processes measured at hillslope scale in the Gwangneung forest catchment to understand and provide the basis of stochastic structures of soil moisture variation. Measured soil moisture series were modelled based upon the developed univariate model platform. The modeling consists of a series of procedures: pre-treatment of data, model structure investigation, selection of candidate models, parameter estimation and diagnostic checking. The spatial distribution of model is associated with topographic characteristics of the hillslope. The upslope area computed by the multiple flow direction algorithm and the local slope are found to be effective parameters to explain the distribution of the model structure. This study enables us to identify the key factors affecting the soil moisture distribution and to ultimately construct a realistic soil moisture map in a complex landscape such as the Gwangneung Supersite.

• Journal of the Korean Data and Information Science Society
• /
• v.22 no.6
• /
• pp.1065-1074
• /
• 2011

During several decades, many researchers in the field of finance have studied Value at Risk (VaR) to measure the market risk. VaR indicates the worst loss over a target horizon such that there is a low, pre-specified probability that the actual loss will be larger (Jorion, 2006, p.106). In this paper, we compare conditional copula method with two conventional VaR forecasting methods based on simple moving average and exponentially weighted moving average for measuring the risk of the portfolio, consisting of two domestic stock indices. Through real data analysis, we conclude that the conditional copula method can improve the accuracy of portfolio VaR forecasting in the presence of high kurtosis and strong correlation in the data.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.25 no.12
• /
• pp.1756-1761
• /
• 2021

Although the COVID-19 outbreak that occurred in Wuhan, Hubei around December 2019, seemed to be gradually decreasing, it was gradually increasing as of November 2020 and June 2021, and estimated confirmed cases were 192 million worldwide and approximately 184 thousand in South Korea. The Central Disaster and Safety Countermeasures Headquarters have been taking strong countermeasures by implementing level 4 social distancing. However, as the highly infectious COVID-19 variants, such as Delta mutation, have been on the rise, the number of daily confirmed cases in Korea has increased to 1,800. Therefore, the number of cumulative confirmed COVID-19 cases is predicted using ARIMA algorithms to emphasize the severity of COVID-19. In the process, differences are used to remove trends and seasonality, and p, d, and q values are determined and forecasted in ARIMA using MA, AR, autocorrelation functions, and partial autocorrelation functions. Finally, forecast and actual values are compared to evaluate how well it was forecasted.