Dong Wan Shin
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Dong Wan Shin has a total of 24 co-authors. Here are the top co-authors by number of shared publications:
Ji-Eun Choi
Department of Statistics and Data Science, Pukyong National University
18 shared publicationsEunju Hwang
Department of Applied Statistics, Gachon University, SengNamSi, South Korea
11 shared publicationsJi Won Shin
Department of Applied Statistics, Gachon University
4 shared publications- Ji-Eun Choi
Department of Statistics, Pukyong National University, Busan, Korea
4 shared publications - Eunju Hwang
Department of Applied Statistics, Gachon University, Korea
2 shared publications Man-Suk Oh
Department of Statistics , Ewha University , Seoul, 120-750, South Korea
2 shared publications
Latest Publications
- Open Access
- 31 August 2025
- Published by The Korean Statistical Society in Korean Journal of Applied Statistics
- Vol. 38 (4) , 485-495
A comparison of vector SHAP explanations of the S&P 500 realized volatility forecast by LSTM and ADL models
Ji Won Shin, Ji-Eun Choi, Dong Wan Shin. Korean J Appl Stat 2025;38:485-95. https://doi.org/10.5351/KJAS.2025.38.3.485
- 31 August 2025
- Published by The Korean Statistical Society in Korean Journal of Applied Statistics
- Vol. 38 (4) , 485-495
Open AccessA comparison of vector SHAP explanations of the S&P 500 realized volatility forecast by LSTM and ADL models
Ji Won Shin, Ji-Eun Choi, Dong Wan Shin. Korean J Appl Stat 2025;38:485-95. https://doi.org/10.5351/KJAS.2025.38.3.485
- 18 November 2024
- Published by Wiley in Journal of Forecasting
- Vol. 44 (2) , 635-645
Vector SHAP Values for Machine Learning Time Series Forecasting
We propose a new vector SHapley Additive exPlanations (SHAP) to interpret machine learning models for forecasting time series using lags of predictor variables. Unlike the standard SHAP measuring the contribution of each lag of each predictor variable, the proposed vector SHAP measures the contribution of the vector of the lags of each variable. The vector SHAP has an advantage of faster computation over the standard SHAP. Some desirable properties of the vector SHAP (vector local accuracy, vector missingness, and vector consistency) are established. A Monte Carlo simulation shows that the vector SHAP has a much faster computing time than the SHAP; the difference of the standard SHAP and the vector SHAP is small; the sampling SHAP is sensitive to the sampling proportion in a range of practical application; the vector SHAP mitigates the sensitivity issue. The vector SHAP is applied to the realized volatility of world major stock price indices of 16 countries for forecasting the realized volatility of South Korea stock price index, KOSPI. Further vectoring by regions of Europe, North America, and Asia yields vector SHAP value for each region which is very close to the sum of vector SHAP values of the countries of the region, illustrating usefulness of the strategy of vectoring.
- 18 November 2024
- Published by Wiley in Journal of Forecasting
- Vol. 44 (2) , 635-645
Vector SHAP Values for Machine Learning Time Series Forecasting
We propose a new vector SHapley Additive exPlanations (SHAP) to interpret machine learning models for forecasting time series using lags of predictor variables. Unlike the standard SHAP measuring the contribution of each lag of each predictor variable, the proposed vector SHAP measures the contribution of the vector of the lags of each variable. The vector SHAP has an advantage of faster computation over the standard SHAP. Some desirable properties of the vector SHAP (vector local accuracy, vector missingness, and vector consistency) are established. A Monte Carlo simulation shows that the vector SHAP has a much faster computing time than the SHAP; the difference of the standard SHAP and the vector SHAP is small; the sampling SHAP is sensitive to the sampling proportion in a range of practical application; the vector SHAP mitigates the sensitivity issue. The vector SHAP is applied to the realized volatility of world major stock price indices of 16 countries for forecasting the realized volatility of South Korea stock price index, KOSPI. Further vectoring by regions of Europe, North America, and Asia yields vector SHAP value for each region which is very close to the sum of vector SHAP values of the countries of the region, illustrating usefulness of the strategy of vectoring.
- 2 October 2024
- Published by Springer Nature in Journal of the Korean Statistical Society
- Vol. 54 (1) , 91-109
A break test for the tail-event correlation matrix based on the self-normalization method
- 2 October 2024
- Published by Springer Nature in Journal of the Korean Statistical Society
- Vol. 54 (1) , 91-109
A break test for the tail-event correlation matrix based on the self-normalization method
- 15 February 2024
- Published by Springer Nature in Journal of the Korean Statistical Society
- Vol. 53 (2) , 495-508
A self-normalization test for structural breaks in a regression model for panel data sets
- 15 February 2024
- Published by Springer Nature in Journal of the Korean Statistical Society
- Vol. 53 (2) , 495-508
A self-normalization test for structural breaks in a regression model for panel data sets
- Open Access
- 30 April 2023
- Published by The Korean Statistical Society in Korean Journal of Applied Statistics
- Vol. 36 (2) , 129-139
Forecasting volatility index by temporal convolutional neural network
Ji Won Shin, Dong Wan Shin. Korean J Appl Stat 2023;36:129-39. https://doi.org/10.5351/KJAS.2023.36.2.129
- 30 April 2023
- Published by The Korean Statistical Society in Korean Journal of Applied Statistics
- Vol. 36 (2) , 129-139
Open AccessForecasting volatility index by temporal convolutional neural network
Ji Won Shin, Dong Wan Shin. Korean J Appl Stat 2023;36:129-39. https://doi.org/10.5351/KJAS.2023.36.2.129
- 23 February 2023
- Published by Taylor & Francis in Journal of Statistical Computation and Simulation
- Vol. 93 (14) , 2357-2369
Bootstrapping tests for breaks in mean or variance based on U-statistics
Bootstrapping tests are implemented for the tests U by Schmidt [Detecting changes in the trend function of heteroscedastic time series; 2021. Preprint: arXiv:2108.09206 [math.ST]] for mean break and Schmidt et al. [An asymptotic test for constancy of the variance under short-range dependence. Ann Stat. 2021;49:3460–3481.] for variance break based on U-statistics. The tests U have good powers against epidemic breaks that are common in practice. The test U for variance break is proved to have the nice property of consistency against a general class of alternatives. However, the tests U have non-ignorable finite sample size distortion under serial correlation and/or conditional heteroscedasticity. Our implementation based on autoregressive residual bootstrapping and moving block bootstrapping are shown to remedy the size distortion problems of U for mean break and for variance break, respectively, in a Monte-Carlo experiment. The experiment also demonstrates the power advantages of bootstrapping tests over the original tests and other standard break tests against epidemic breaks, which, however, are accompanied by disadvantages against simple single breaks. The proposed bootstrapping tests are well illustrated by break analyses of means and variances of the log-return and realized the volatility of the US S&P 500.- 23 February 2023
- Published by Taylor & Francis in Journal of Statistical Computation and Simulation
- Vol. 93 (14) , 2357-2369
Bootstrapping tests for breaks in mean or variance based on U-statistics
Bootstrapping tests are implemented for the tests U by Schmidt [Detecting changes in the trend function of heteroscedastic time series; 2021. Preprint: arXiv:2108.09206 [math.ST]] for mean break and Schmidt et al. [An asymptotic test for constancy of the variance under short-range dependence. Ann Stat. 2021;49:3460–3481.] for variance break based on U-statistics. The tests U have good powers against epidemic breaks that are common in practice. The test U for variance break is proved to have the nice property of consistency against a general class of alternatives. However, the tests U have non-ignorable finite sample size distortion under serial correlation and/or conditional heteroscedasticity. Our implementation based on autoregressive residual bootstrapping and moving block bootstrapping are shown to remedy the size distortion problems of U for mean break and for variance break, respectively, in a Monte-Carlo experiment. The experiment also demonstrates the power advantages of bootstrapping tests over the original tests and other standard break tests against epidemic breaks, which, however, are accompanied by disadvantages against simple single breaks. The proposed bootstrapping tests are well illustrated by break analyses of means and variances of the log-return and realized the volatility of the US S&P 500.- 22 June 2022
- Published by Taylor & Francis in Communications in Statistics - Theory and Methods
- Vol. 53 (2) , 1-14
Subsample scan test for multiple breaks based on self-normalization
The self-normalization of Shao and Zhang and subsample scan method of Yau and Zhao are combined to produce a test SN for testing breaks for time series data. Unlike the original break test based on self-normalization of Shao and Zhang, the proposed test has power against canceling multiple breaks. The proposed test has several advantages over an existing test TN of Zhang and Lavitas designed for the same purpose of detecting canceling multiple breaks: having no computational burden issue; having stabler size; having no need for specification of a trimming parameter. Comparison is also made with the recent test UN by Schmidt for mean break and by Schmidt et al. for variance break based on U-statistics and shows better size of SN than UN for serially correlated samples but worse power of SN than UN. Unlike UN, SN is a general purpose test which can be used for testing break in any parameter, for example, correlation, under some weak conditions. These advantages recommend us the proposed test SN as a practical alternative to TN and UN in spite of some disadvantages of smaller power and requiring specification of a parameter for the number of windows.
- 22 June 2022
- Published by Taylor & Francis in Communications in Statistics - Theory and Methods
- Vol. 53 (2) , 1-14
Subsample scan test for multiple breaks based on self-normalization
The self-normalization of Shao and Zhang and subsample scan method of Yau and Zhao are combined to produce a test SN for testing breaks for time series data. Unlike the original break test based on self-normalization of Shao and Zhang, the proposed test has power against canceling multiple breaks. The proposed test has several advantages over an existing test TN of Zhang and Lavitas designed for the same purpose of detecting canceling multiple breaks: having no computational burden issue; having stabler size; having no need for specification of a trimming parameter. Comparison is also made with the recent test UN by Schmidt for mean break and by Schmidt et al. for variance break based on U-statistics and shows better size of SN than UN for serially correlated samples but worse power of SN than UN. Unlike UN, SN is a general purpose test which can be used for testing break in any parameter, for example, correlation, under some weak conditions. These advantages recommend us the proposed test SN as a practical alternative to TN and UN in spite of some disadvantages of smaller power and requiring specification of a parameter for the number of windows.
- 31 January 2022
- Published by The Korean Statistical Society in Communications for Statistical Applications and Methods
- Vol. 29 (1) , 41-51
How to improve oil consumption forecast using google trends from online big data?: the structured regularization methods for large vector autoregressive model
Ji-Eun Choi, Dong Wan Shin. CSAM 2022;29:41-51. https://doi.org/10.29220/CSAM.2022.29.1.041
- 31 January 2022
- Published by The Korean Statistical Society in Communications for Statistical Applications and Methods
- Vol. 29 (1) , 41-51
How to improve oil consumption forecast using google trends from online big data?: the structured regularization methods for large vector autoregressive model
Ji-Eun Choi, Dong Wan Shin. CSAM 2022;29:41-51. https://doi.org/10.29220/CSAM.2022.29.1.041
- 31 January 2022
- Published by The Korean Statistical Society in Communications for Statistical Applications and Methods
- Vol. 29 (1) , 721-731
How to improve oil consumption forecast using google trends from online big data?: the structured regularization methods for large vector autoregressive model
Ji-Eun Choi, Dong Wan Shin. CSAM 2022;29:721-31. https://doi.org/10.29220/CSAM.2022.29.1.721
- 31 January 2022
- Published by The Korean Statistical Society in Communications for Statistical Applications and Methods
- Vol. 29 (1) , 721-731
How to improve oil consumption forecast using google trends from online big data?: the structured regularization methods for large vector autoregressive model
Ji-Eun Choi, Dong Wan Shin. CSAM 2022;29:721-31. https://doi.org/10.29220/CSAM.2022.29.1.721
- 22 December 2021
- Published by Wiley in Journal of Forecasting
- Vol. 41 (6) , 1087-1098
Parallel architecture of CNN‐bidirectional LSTMs for implied volatility forecast
We propose a new forecast method based on artificial neural networks (ANNs), ensemble CNN‐BiLSTM, which is an ensemble of three CNN‐BiLSTMs constructed with the combination of Convolution Neural Network (CNN) and Bidirectional Long Short‐Term Memory (BiLSTM). The new forecast method effectively handles the strong long memory serial dependence feature of the daily VXN by the ensemble CNN‐BiLSTM together with proper normalization and batch size. The long memory features arising from time‐dependent mean and variance are largely reduced by normalizing the data with local mean and local standard deviation (SD). The batch size is determined by the optimal block length of the moving block bootstrap which reflects the long memory. The ensemble CNN‐BiLSTM concentrates on 1‐day, 1‐week, and 2‐week features of the normalized VXN data. An out‐of‐sample forecast comparison reveals that (i) the proposed ensemble CNN‐BiLSTM has better forecast performance than the autoregressive model, DNN, LSTM, BiLSTM, and individual CNN‐BiLSTMs; (ii) the local mean‐SD normalization has superior forecast performance to the standard global mean‐SD normalization; (iii) and the optimal block length improves the forecast performance over a batch size considered in the literature.
- 22 December 2021
- Published by Wiley in Journal of Forecasting
- Vol. 41 (6) , 1087-1098
Parallel architecture of CNN‐bidirectional LSTMs for implied volatility forecast
We propose a new forecast method based on artificial neural networks (ANNs), ensemble CNN‐BiLSTM, which is an ensemble of three CNN‐BiLSTMs constructed with the combination of Convolution Neural Network (CNN) and Bidirectional Long Short‐Term Memory (BiLSTM). The new forecast method effectively handles the strong long memory serial dependence feature of the daily VXN by the ensemble CNN‐BiLSTM together with proper normalization and batch size. The long memory features arising from time‐dependent mean and variance are largely reduced by normalizing the data with local mean and local standard deviation (SD). The batch size is determined by the optimal block length of the moving block bootstrap which reflects the long memory. The ensemble CNN‐BiLSTM concentrates on 1‐day, 1‐week, and 2‐week features of the normalized VXN data. An out‐of‐sample forecast comparison reveals that (i) the proposed ensemble CNN‐BiLSTM has better forecast performance than the autoregressive model, DNN, LSTM, BiLSTM, and individual CNN‐BiLSTMs; (ii) the local mean‐SD normalization has superior forecast performance to the standard global mean‐SD normalization; (iii) and the optimal block length improves the forecast performance over a batch size considered in the literature.