• The Korean Journal of Applied Statistics
• /
• v.32 no.5
• /
• pp.671-681
• /
• 2019

Studies on predicting arrhythmia using machine learning have been actively conducted with increasing number of arrhythmia patients. Existing studies have predicted arrhythmia based on multivariate data of feature variables extracted from RR interval data at a specific time point. In this study, we consider that the pattern of the heart state changes with time can be important information for the arrhythmia prediction. Therefore, we investigate the usefulness of predicting the arrhythmia with multivariate time series data obtained by extracting and accumulating the multivariate vectors of the feature variables at various time points. When considering 1-nearest neighbor classification method and its ensemble for comparison, it is confirmed that the multivariate time series data based method can have better classification performance than the multivariate data based method if we select an appropriate time series distance function.

• Journal of the Korean Data and Information Science Society
• /
• v.18 no.4
• /
• pp.915-925
• /
• 2007

Multivariate GARCH has been useful to model dynamic relationships between volatilities arising from each component series of multivariate time series. Methodologies including EWMA(Exponentially weighted moving-average model), DVEC(Diagonal VEC model), BEKK and CCC(Constant conditional correlation model) models are comparatively reviewed for bivariate time series. In addition, these models are applied to evaluate VaR(Value at Risk) and to construct joint prediction region. To illustrate, bivariate stock prices data consisting of Samsung Electronics and LG Electronics are analysed.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2008.05a
• /
• pp.694-698
• /
• 2008

In hydrology, it is a difficult task to deal with multivariate time series such as modeling streamflows of an entire complex river system. Normal distribution based model such as MARMA (Multivariate Autorgressive Moving average) has been a major approach for modeling the multivariate time series. There are some limitations for the normal based models. One of them might be the unfavorable data-transformation forcing that the data follow the normal distribution. Furthermore, the high dimension multivariate model requires the very large parameter matrix. As an alternative, one might be decomposing the multivariate data into independent components and modeling it individually. In 1985, Lins used Principal Component Analysis (PCA). The five scores, the decomposed data from the original data, were taken and were formulated individually. The one of the five scores were modeled with AR-2 while the others are modeled with AR-1 model. From the time series analysis using the scores of the five components, he noted "principal component time series might provide a relatively simple and meaningful alternative to conventional large MARMA models". This study is inspired from the researcher's quote to develop a multivariate simulation model. The multivariate simulation model is suggested here using Principal Component Analysis (PCA) and Independent Component Analysis (ICA). Three modeling step is applied for simulation. (1) PCA is used to decompose the correlated multivariate data into the uncorrelated data while ICA decomposes the data into independent components. Here, the autocorrelation structure of the decomposed data is still dominant, which is inherited from the data of the original domain. (2) Each component is resampled by block bootstrapping or K-nearest neighbor. (3) The resampled components bring back to original domain. From using the suggested approach one might expect that a) the simulated data are different with the historical data, b) no data transformation is required (in case of ICA), c) a complex system can be decomposed into independent component and modeled individually. The model with PCA and ICA are compared with the various statistics such as the basic statistics (mean, standard deviation, skewness, autocorrelation), and reservoir-related statistics, kernel density estimate.

• Journal of the Korean Data and Information Science Society
• /
• v.17 no.1
• /
• pp.115-122
• /
• 2006

A test is suggested for detecting deviations from both multivariate normality and independence. This test can be used for assessing the normality and independence of univariate time series residuals. We derive the limiting distribution of the test statistic and a simulation study is conducted to study the accuracy of the limiting distribution in finite samples. Finally, we apply our method to a real data of time series.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.10 no.11
• /
• pp.5476-5496
• /
• 2016

Energy efficiency of resource-constrained wireless sensor networks is critical in applications such as real-time monitoring/surveillance. To improve the energy efficiency and reduce the energy consumption, the time series data can be compressed before transmission. However, most of the compression algorithms for time series data were developed only for single variate scenarios, while in practice there are often multiple sensor nodes in one application and the collected data is actually multivariate time series. In this paper, we propose to compress the time series data by the Lasso (least absolute shrinkage and selection operator) approximation. We show that, our approach can be naturally extended for compressing the multivariate time series data. Our extension is novel since it constructs an optimal projection of the original multivariates where the best energy efficiency can be realized. The two algorithms are named by ULasso (Univariate Lasso) and MLasso (Multivariate Lasso), for which we also provide practical guidance for parameter selection. Finally, empirically evaluation is implemented with several publicly available real-world data sets from different application domains. We quantify the algorithm performance by measuring the approximation error, compression ratio, and computation complexity. The results show that ULasso and MLasso are superior to or at least equivalent to compression performance of LTC and PLAMlis. Particularly, MLasso can significantly reduce the smooth multivariate time series data, without breaking the major trends and important changes of the sensor network system.

• Journal of Korean Society for Quality Management
• /
• v.49 no.4
• /
• pp.539-550
• /
• 2021

Purpose: The purpose of this study is to propose a multivariate CUSUM control chart that can detect the out-of-control state fast while monitoring the cross- and auto- correlated multivariate time series data. Methods: We first build models to estimate the observation data and calculate the corresponding residuals. After then, a multivariate CUSUM chart is applied to monitor the residuals instead of the original raw observation data. Vector Autoregression and Artificial Neural Net are selected for the modelling, and Separated-MCUSUM chart is selected for the monitoring. The suggested methods are tested under a number of experimental settings and the performances are compared with those of other existing methods. Results: We find that Artificial Neural Net is more appropriate than Vector Autoregression for the modelling and show the combination of Separated-MCUSUM with Artificial Neural Net outperforms the other alternatives considered in this paper. Conclusion: The suggested chart has many advantages. It can monitor the complicated multivariate data with cross- and auto- correlation, and detects the out-of-control state fast. Unlike other CUSUM charts finding their control limits by trial and error simulation, the suggested chart saves lots of time and effort by approximating its control limit mathematically. We expect that the suggested chart performs not only effectively but also efficiently for monitoring the process with complicated correlations and frequently-changed parameters.

• Journal of the Korean Data and Information Science Society
• /
• v.18 no.1
• /
• pp.73-80
• /
• 2007

Cointegration(together with VARMA(vector ARMA)) has been proven to be useful for analyzing multivariate non-stationary data in the field of financial time series. It provides a linear combination (which turns out to be stationary series) of non-stationary component series. This linear combination equation is referred to as long term equilibrium between the component series. We consider two sets of Korean bivariate financial time series and then illustrate cointegration analysis. Specifically estimated VAR(vector AR) and VECM(vector error correction model) are obtained and CV(cointegrating vector) is found for each data sets.

• Communications for Statistical Applications and Methods
• /
• v.18 no.1
• /
• pp.13-21
• /
• 2011

This paper studies the analysis of multivariate nonstationary time series with seasonality. Three types of multivariate time series models are considered: seasonal cointegration model, nonseasonal cointegration model with seasonal dummies, and vector autoregressive model in seasonal differences that are compared for forecasting performances using Korean macro-economic time series data. The cointegration models produce smaller forecast errors in short horizons; however, when longer forecasting periods are considered the vector autoregressive model appears preferable.

• Communications for Statistical Applications and Methods
• /
• v.30 no.4
• /
• pp.369-388
• /
• 2023

In this paper, we develop the two-step procedure that detects and estimates the position of structural changes for multivariate nonstationary time series, either on mean parameters or second-order structures. We first investigate the presence of mean structural change by monitoring data through the aggregated cumulative sum (CUSUM) type statistic, a sequential procedure identifying the likely position of the change point on its trend. If no mean change point is detected, the proposed method proceeds to scan the second-order structural change by modeling the multivariate nonstationary time series with a multivariate locally stationary Wavelet process, allowing the time-localized auto-correlation and cross-dependence. Under this framework, the estimated dynamic spectral matrices derived from the local wavelet periodogram capture the time-evolving scale-specific auto- and cross-dependence features of data. We then monitor the change point from the lower-dimensional approximated space of the spectral matrices over time by applying the dynamic principal component analysis. Different from existing methods requiring prior information on the type of changes between mean and covariance structures as an input for the implementation, the proposed algorithm provides the output indicating the type of change and the estimated location of its occurrence. The performance of the proposed method is demonstrated in simulations and the analysis of two real finance datasets.

• Korean Journal of Environmental Biology
• /
• v.34 no.4
• /
• pp.365-373
• /
• 2016

Much of the data used in the analysis of environmental ecological data is being obtained over time. If the number of time points is small, the data will not be given enough information, so repeated measurements or multiple survey points data should be used to perform a comprehensive analysis. The method used for that case is longitudinal data analysis or mixed model analysis. However, if the amount of information is sufficient due to the large number of time points, repetitive data are not needed and these data are analyzed using time series analysis technique. In particular, with a large number of data points in the current situation, when we want to predict how each variable affects each other, or what trends will be expected in the future, we should analyze the data using time series analysis techniques. In this study, we introduce univariate time series analysis, intervention time series model, transfer function model, and multivariate time series model and review research papers studied in Korea. We also introduce an error correction model, which can be used to analyze environmental ecological data.