publications

2024

1. JMSE

   Post-processing maritime wind forecasts from the European Centre for Medium-Range Weather Forecasts around the Korean Peninsula using support vector regression and principal component analysis

   Seung-Hyun Moon ,  Do-Youn Kim ,  and  Yong-Hyuk Kim

   Journal of Marine Science and Engineering, 2024

   Abs

   Accurate wind data are crucial for successful search and rescue (SAR) operations on the sea surface in maritime accidents, as survivors or debris tend to drift with the wind. As maritime accidents frequently occur outside the range of wind stations, SAR operations heavily rely on wind forecasts generated by numerical models. However, numerical models encounter delays in generating results due to spin-up issues, and their predictions can sometimes exhibit inherent biases caused by geographical factors. To overcome these limitations, we reviewed the observations for the first 24 h of the 72-hour forecast from the ECMWF and then post-processed the forecast for the remaining 48 h. By effectively reducing the dimensionality of input variables comprising observation and forecast data using principal component analysis, we improved wind predictions with support vector regression. Our model achieved an average RMSE improvement of 16.01% compared to the original forecast from the ECMWF. Furthermore, it achieved an average RMSE improvement of 5.42% for locations without observation data by employing a model trained on data from the nearest wind station and then applying an adaptive weighting scheme to the output of that model.

2023

1. JKIIS

   Drifter trajectory prediction using stacked ensemble with multiple machine learning algorithms

   Hyeonki Jeong ,  Tae-Hoon Kim ,  Do-Youn Kim ,  Yong-Hyuk Kim ,  and  Seung-Hyun Moon

   Journal of Korean Institute of Intelligent Systems, 2023

2022

1. Math.

   Genetic Mean Reversion Strategy for Online Portfolio Selection with Transaction Costs

   Seung-Hyun Moon ,  and  Yourim Yoon

   Mathematics, 2022

   Abs

   Online portfolio selection (OLPS) is a procedure for allocating portfolio assets using only past information to maximize an expected return. There have been successful mean reversion strategies that have achieved large excess returns on the traditional OLPS benchmark datasets. We propose a genetic mean reversion strategy that evolves a population of portfolio vectors using a hybrid genetic algorithm. Each vector represents the proportion of the portfolio assets, and our strategy chooses the best vector in terms of the expected returns on every trading day. To test our strategy, we used the price information of the S&P 500 constituents from 2000 to 2017 and compared various strategies for online portfolio selection. Our hybrid genetic framework successfully evolved the portfolio vectors; therefore, our strategy outperformed the other strategies when explicit or implicit transaction costs were incurred.

2021

1. Math.

   Genetic feature selection applied to KOSPI and cryptocurrency price prediction

   Dong-Hee Cho ,  Seung-Hyun Moon ,  and  Yong-Hyuk Kim

   Mathematics, 2021

   Abs

   Feature selection reduces the dimension of input variables by eliminating irrelevant features. We propose feature selection techniques based on a genetic algorithm, which is a metaheuristic inspired by a natural selection process. We compare two types of feature selection for predicting a stock market index and cryptocurrency price. The first method is a newly devised genetic filter involving a fitness function designed to increase the relevance between the target and the selected features and decrease the redundancy between the selected features. The second method is a genetic wrapper, whereby we can find the better feature subsets related to KOPSI by exploring the solution space more thoroughly. Both genetic feature selection methods improved the predictive performance of various regression functions. Our best model was applied to predict the KOSPI, cryptocurrency price, and their respective trends after COVID-19.

2. GECCO

   An improved predictor of daily stock index based on a genetic filter

   Dong-Hee Cho ,  Seung-Hyun Moon ,  and  Yong-Hyuk Kim

   In Proceedings of the Genetic and Evolutionary Computation Conference (GECCO) Companion , 2021

   Abs

   In this study, a genetic algorithm-based filter feature selection was applied to the data on the rates of change in various economic indices used worldwide to predict the rates of change of the Korea Composite Stock Price Index (KOSPI). The fitness function is composed of a combination of the results of information gains, F-test, and correlation coefficients. Data for the past 12 years (from 2007 to 2018) were divided into sections according to the yearly time intervals, and feature selection was applied to the data in each section. It was found that the amount of calculation time required for a genetic filter was approximately 75% lower compared to that required for a previous genetic wrapper. It was also found that the mean absolute error of the genetic filter was approximately 26% lower compared to that of the genetic wrapper. The analytic results verified that the genetic filter exhibited better calculation performance and required less time to calculate the optimal solution. Furthermore, a new combination of features related to the KOSPI was obtained.

2020

1. Appl. Sci.

   An improvement on estimated drifter tracking through machine learning and evolutionary search

   Yong-Wook Nam ,  Hwi-Yeon Cho ,  Do-Youn Kim ,  Seung-Hyun Moon ,  and  Yong-Hyuk Kim

   Applied Sciences, 2020

   Abs

   In this study, we estimated drifter tracking over seawater using machine learning and evolutionary search techniques. The parameters used for the prediction are the hourly position of the drifter, the wind velocity, and the flow velocity of each drifter position. Our prediction model was constructed through cross-validation. Trajectories were affected by wind velocity and flow velocity from the starting points of drifters. Mean absolute error (MAE) and normalized cumulative Lagrangian separation (NCLS) were used to evaluate various prediction models. Radial basis function network showed the lowest MAE of 0.0556, an improvement of 35.20% over the numerical model MOHID. Long short-term memory showed the highest NCLS of 0.8762, an improvement of 6.24% over MOHID.

2. Atoms. Res.

   Forecasting lightning around the Korean Peninsula by postprocessing ECMWF data using SVMs and undersampling

   Seung-Hyun Moon ,  and  Yong-Hyuk Kim

   Atmospheric Research, 2020

   Abs

   We use machine learning to generate binary forecasts of the occurrence of lightning within a particular location and time interval. The training data is weather variables found in the forecasts from the European Centre for Medium-range Weather Forecasts, correlated against subsequent lightning reports, for a region containing the Korean Peninsula. Lightning is uncommon, so the amount of data which does not involve lightning tends to swamp the training process. Thus we only consider spatial locations at which lightning frequently occurs, and we also undersample the subset of the remaining data-points which are not associated with lightning. Results from support vector machines and random forests had equitable threat scores of 0.0885 and 0.0828, respectively. The ETS of results from SVMs can be increased to 0.1241 if temporal resolution is reduced by a factor of 2, and 0.1499 if spatial resolution is reduced by a factor of 3.

3. Appl. Sci.

   Detection of precipitation and fog using machine learning on backscatter data from lidar ceilometer

   Yong-Hyuk Kim ,  Seung-Hyun Moon ,  and  Yourim Yoon

   Applied Sciences, 2020

   Abs

   The lidar ceilometer estimates cloud height by analyzing backscatter data. This study examines weather detectability using a lidar ceilometer by making an unprecedented attempt at detecting weather phenomena through the application of machine learning techniques to the backscatter data obtained from a lidar ceilometer. This study investigates the weather phenomena of precipitation and fog, which are expected to greatly affect backscatter data. In this experiment, the backscatter data obtained from the lidar ceilometer, CL51, installed in Boseong, South Korea, were used. For validation, the data from the automatic weather station for precipitation and visibility sensor PWD20 for fog, installed at the same location, were used. The experimental results showed potential for precipitation detection, which yielded an F1 score of 0.34. However, fog detection was found to be very difficult and yielded an F1 score of 0.10.

4. Atoms. Res.

   An improved forecast of precipitation type using correlation-based feature selection and multinomial logistic regression

   Seung-Hyun Moon ,  and  Yong-Hyuk Kim

   Atmospheric Research, 2020

   Abs

   Accurate prediction of precipitation type is an important part of weather forecasting. But using meteorological insight to make such predictions from a small set of weather variables achieves only limited success. We use correlation-based feature selection to assemble an effective subset of the large number of weather variables available in short-range weather forecasts, and from these we obtain the coefficients for multinomial regression, which can then be used to predict precipitation type. We applied this approach to data for significant locations in South Korea, obtained from the European Centre for Medium-Range Weather Forecasts and from the Regional Data Assimilation and Prediction System, and achieved predictions which are respectively 15% and 13% more accurate than those contained in the original forecasts.

5. GECCO

   A daily stock index predictor using feature selection based on a genetic wrapper

   Dong-Hee Cho ,  Seung-Hyun Moon ,  and  Yong-Hyuk Kim

   In Proceedings of the Genetic and Evolutionary Computation Conference (GECCO) Companion , 2020

   Abs

   In this experiment, we applied a genetic feature selection to global macroeconomics indicators, and used support vector regression to create a model for predicting the daily increase/decrease rate of Korea Composite Stock Price Index (KOSPI). Experiments were conducted by dividing the time series of 2007 to 2018 into various yearly time intervals. Overall, the mean absolute error was approximately 15% lower when the genetic algorithm-applied feature selection method was used. The feature selection improved the predictive performance by assembling discriminatory subsets of macroeconomic indicators that effectively represent the stock market trends in a given time interval.

6. Spatiotemporal approaches for quality control and error correction of atmospheric data through machine learning

   Hye-Jin Kim ,  Sung Min Park ,  Byung Jin Choi ,  Seung-Hyun Moon ,  and  Yong-Hyuk Kim

   Computational Intelligence and Neuroscience, 2020

2019

1. J. Hydrol.

   Application of machine learning to an early warning system for very short-term heavy rainfall

   Seung-Hyun Moon ,  Yong-Hyuk Kim ,  Yong Hee Lee ,  and  Byung-Ro Moon

   Journal of Hydrology, 2019

   Abs

   The purpose of an early warning system (EWS) is to issue warning signals prior to extreme events. Extreme weather events, however, are hard to predict due to their chaotic behavior. This paper suggests a method for an effective EWS for very short-term heavy rainfall with machine learning techniques. The EWS produces a warning signal when it is expected to reach the criterion for a heavy rain advisory within the next 3 h. We devised a selective discretization method that converts a subset of continuous input variables to nominal ones. Meteorological data obtained from automatic weather stations are preprocessed by the selective discretization and principal component analysis. As a classifier, logistic regression is used to predict whether or not a warning is required. A comparative evaluation was performed on the EWS models generated by various classifiers. The tests were run for 652 locations in South Korea from 2007 to 2012. The empirical results showed that the preprocessing methods improved the prediction quality and logistic regression works well on heavy rainfall nowcasting in terms of F-measure and equitable threat score.

2018

1. Adv. Meteorol.

   Detecting anomalies in meteorological data using support vector regression

   Min-Ki Lee ,  Seung-Hyun Moon ,  Yourim Yoon ,  Yong-Hyuk Kim ,  and  Byung-Ro Moon

   Advances in Meteorology, 2018

2014

1. SMC

   Correcting abnormalities in meteorological data by machine learning

   Min-Ki Lee ,  Seung-Hyun Moon ,  Yong-Hyuk Kim ,  and  Byung-Ro Moon

   In IEEE International Conference on Systems, Man, and Cybernetics (SMC) , 2014

2006

1. GECCO

   A hybrid genetic search for multiple sequence alignment

   Seung-Hyun Moon ,  Sung-Soon Choi ,  and  Byung-Ro Moon

   In Proceedings of the 8th Annual Conference on Genetic and Evolutionary Computation (GECCO) , 2006

   Abs

   This paper proposes a hybrid genetic algorithm for multiple sequence alignment. The algorithm evolves guide sequences and aligns input sequences based on the guide sequences. It also embeds a local search heuristic to search the problem space effectively. In the experiments for various data sets, the proposed algorithm showed the performance comparable to existing algorithms.