Earticle

Demand forecasting is a crucial task for an online retail where has to manage daily fresh foods effectively. Failing in forecasting results loss of profitability because of incompetent inventory management. This study investigated the optimal performance of different forecasting models for a very short shelf-life product. Demand data of 13 perishable items with aging of 210 days were used for analysis. Our comparison results of four methods: Trivial Identity, Seasonal Naïve, Feed-Forward and Autoregressive Recurrent Neural Networks (DeepAR) reveals that DeepAR outperforms with the lowest MAPE. This study also suggests the managerial implications by employing coefficient of variation (CV) as demand variation indicators. Three classes: Low, Medium and High variation are introduced for classify 13 products into groups. Our analysis found that DeepAR is suitable for medium and high variations, while the low group can use any methods. With this approach, the case can gain benefit of better fill-rate performance.

This research analyzes the effects of factors on the demands for outbound to the countries such as Japan, China, the United States of America, Thailand, Philippines, Hong Kong, Singapore and Australia, the countries preferred by many Koreans. The factors for this research are (1) economic variables such as Korea Composite Stock Price Index (KOSPI), which could have influences on outbound tourism and exchange rate and (2) unpredictable events such as diseases, financial crisis and terrors. Regression analysis was used to identify relationship based on the monthly data from January 2001 to December 2010. The results of the analysis show that both exchange rate and KOSPI have impacts on the demands for outbound travel. In the case of travels to the United States of America and Philippines, Korean tourists usually have particular purposes such as studying, visiting relatives, playing golf or honeymoon, thus they are less influenced by the exchange rate. Moreover, Korean tourists tend not to visit particular locations for some time when shock reaction happens. As the demands for outbound travels are different from country to country accompanied by economic variables and shock variables, differentiated measure to should be considered to come close to the target numbers of tourists by switching as well as creating the demands. For further study we plan to build outbound tourism forecasting models using Artificial Neural Networks.

Sustainable power systems should include solar energy generation. However, for effective grid management and the integration of renewable energy sources, accurate solar power generation predictions are essential. Therefore, this study compares the prediction of solar power forecasting in Italy and Bulgaria. These are two countries that have alike latitudes but different populations and solar energy production. The historical solar power generation and meteorological data from these countries are preprocessed and then used to apply four different deep learning models including Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU). The results are analyzed to gain insights into how the proximity of geographical locations and the quality and quantity of data impact the precision of prediction algorithms.

With the advancement of deep learning technology, research in time series forecasting is thriving across various fields. In the financial sector, where time series data is complex and volatile, making accurate predictions challenging, the importance of such research is growing as more people invest in financial markets. While deep learning models such as Autoencoders, Recurrent Neural Networks, Long Short-Term Memory networks, and Gated Recurrent Units are actively used in financial forecasting, the Transformer model, known for its efficiency and ability to address long-term dependency issues, has predominantly been applied to stock prediction through market sentiment analysis based on textual information rather than technical price predictions. Moreover, while there is extensive research on stock forecasting, there is a notable lack of studies on Exchange Traded Funds. This study aims to bridge this gap by using Transformer models to forecast future on Exchange Traded Funds prices and performing a comparative analysis of Transformer models of various sizes.

This article examines the volatility forecasting abilities of two approaches: one is GARCH-type model that uses carbon futures prices, and the other is an implied volatility from carbon options prices. Based on the results, we document that GARCH-type models perform better than an implied volatility. This result suggests that carbon options have little information about carbon futures due to their low trading volume. We also investigate whether the volatilities of energy markets, i.e., Brent oil, coal, natural gas, and electricity, forecast following day’s carbon futures volatility. According to the results, we suggest that Brent oil and natural gas may be used to forecast the volatility of carbon futures.

Data mining is the process of searching and analyzing large quantities of data for finding out meaningful patterns and rules. Artificial Neural Network (ANN) is one of the tools of data mining which is becoming very popular in forecasting the future values. Some of the areas where it is used are banking, medicine, retailing and fraud detection. In finance, artificial neural network is used in various disciplines including stock market forecasting. In the stock market time series, due to high volatility, it is very important to choose a model which reads volatility and forecasts the future values considering volatility as one of the major attributes for forecasting. In this paper, an attempt is made to develop two models - one using feed forward back propagation Artificial Neural Network and the other using Autoregressive Conditional Heteroskedasticity (ARCH) technique for forecasting stock market returns. Various parameters which are considered for the design of optimal ANN model development are input and output data normalization, transfer function and neuron/s at input, hidden and output layers, number of hidden layers, values with respect to momentum, learning rate and error tolerance. Simulations have been done using prices of daily close of Sensex. Stock market returns are chosen as input data and output is the forecasted return. Simulations of the Model have been done using MATLAB® 6.1.0.450 and EViews 4.1. Convergence and performance of models have been evaluated on the basis of the simulation results. Performance evaluation is done on the basis of the errors calculated between the actual and predicted values.

본 논문은 고속도로 터널구간을 대상으로 교통사고특성을 다각적으로 분석하여 다양한 독립변수를 선정하고 종속변수를 건, 건/km, 건/백만대km로 다양화하여 다중선형회귀모형을 개발하였다. 그리고 개발된 모형들은 상호 비교 검토하여 최종적으로 교통사고영향요인으로 구성된 신뢰성 있는 교통사고예측모형을 결정하였다. 교통사고예측모형은 모형의 R2, F값 등 검정통계량 수준, 다중공선성, 잔차분석 등 모형검증과정이 수행되었고 터널구간의 교통사고특성 반영여부 등을 검토하여 최종적으로 터널길이에 따라 총 2개의 모형을 선정하였다. 선정된 종속변수는 ln(건/백만대km)이며, 독립변수는 연평균일교통량(AADT), 종단구배, 터널높이로 구성되었다. 추정모형은 RMSE, MAE를 이용하여 예측한 값과 실제 관측값과의 차이를 분석하여 터널구간의 교통사고를 설명하는데 적합한 모형으로 파악되었다.

This paper analyzed the characteristics of traffic accidents in all tunnels on nationwide freeways and selected some various independent variables related to accident occurrence in tunnels. The study aims to develop reliable accident forecasting models using the various dependent variables such as the number of accident (no.), no./km, and no./MVK. Finally, reliable multiple linear regression models were proposed in this paper. This study tested the validity verification of developed models through statistics such as R2, F values, multicollinearity, residual analysis. The paper selected the accident forecasting models considering the characteristics of tunnel accidents and two models were finally proposed according to two groups of tunnel length. In the selected models, natural logarithm of ln(no./MVK) is used for the dependent variable and AADT, vertical slope, and tunnel hight are used for the independent variables. The reliability of two models was proved by the comparison analysis between field data and estimating data using RMSE and MAE. These models may be not only effective in evaluating tunnel safety under design and planning phases of tunnel but also useful to reduce traffic accidents in tunnels and to manage the traffic flow of tunnel.

It is revealed that, as the use of automated manufacturing process is increasing and the process inspection technology is improvement in industry for recently years, the data from mass production system will exhibit some degree of autocorrelation. Therefore, using the EWMA forecast models which has been proposed as a very good forecasting tool when autocorrelated construction contacted with time-series models is explained, I want to analysis sensitivity of quality control charts considering the variation of error or forecast residuals. In this paper, for the AR(1) process of EWMA forecast model, when the constant term ξ are zero and different from zero, the results of analyzed the sensitivity of Χ, CUSUM and EWMA control chart using EWMA forecast residuals are summarized as follows. First, the EWMA statistic is known as a good alternative to the Box-Jenkins forecasts for a variety of time-series models, the violation of the independence assumption can also lead to suboptimal monitoring schemes, particularly for the EWMA and the CUSUM control charts applied to EWMA forecast residuals, and particularly sensitive to the presence of autocorrelation For example, when θ₁ is underestimated, the EWMA and the CUSUM control charts applied to the EWMA forecast residuals provide ARLs which are much larger than anticipated. When θ₁is overestimated, the EWMA and the CUSUM control charts applied to the EWMA forecast residuals provide smaller ARLs than anticipated. Second, for EWMA forecasting model, whether ξ is zero or ξ is different from zero, the performance of control charts is equality. Therefor, the violation of the independence assumption can lead to suboptimal monitoring schemes, particularly for the EWMA and the CUSUM control charts applied to forecast errors. In this cases, whether one could adjust the control limits of the EWMA and the CUSUM control charts applied to forecast errors to obtain the desired in-control ARLs or when either underestimation in θ₁, or overestimation in θ₁, the Shewhart control chart is recommended since it is least sensitive to the violation of the independence assumption.

본 연구는 아시아 이머징 주식시장 변동성에 장기기억과 비대칭성이 존재하 는지를 분석하였다. 이를 위해서 AR(1)-FIGARCH모형과 AR(1)-FIAPARCH 모형을 이용하여 실증분석 하였다. 실증분석 결과 아시아 이머징 시장 변동성 에 장기기억과 비대칭성이 존재하는 것으로 분석 되었다. 이는 주식시장에 충 격이 소멸되지 않고 장기간 지속되는 것을 의미한다. 또한, 투자자들은 나쁜 충격에 민감하게 반응하는 것으로 분석 되었다. 마지막으로 장기기억 모형의 예측력을 분석한 결과 장기기억과 비대칭성을 분석할 수 있는 FIAPARCH모 형이 장기기억만을 분석하는 FIGARCH모형보다 우월한 것으로 분석되었다. 이러한 결과는 변동성의 특징은 장기기억 뿐만 아니라 비대칭성을 동시에 포 함하고 있음을 의미하고 있다. 이러한 결과는 변동성을 정확하게 측정하는데 중요한 자료로 이용될 것으로 판단된다. 특히, Value at Risk를 측정하기 위해 서 정교한 변동성 예측 모형이 필요하기 때문에 주식시장 리스크 관리에 도움 이 될 것으로 판단된다.

This study investigates long memory and asymmetry in volatility of Asian emerging markets using the AR(1)-FIGARCH and AR(1)-FIAPARCH models under the Student-t distributions. The empirical results show strong evidence of long memory in the volatility of 8 Asian emerging markets. This evidence indicates that market shocks to volatility slowly disappear over the time. In addition, the FIAPARCH model detect volatility asymmetry in which investors want hedge negative information in Asian emerging markets. Finally, the forecasting error functions suggest that the FIAPARCH model with the Student-t distribution offers a superior forecasting ability to other models. These results provide important implications on assess accurate Value at Risk in the Asian emerging markets.

본 연구는 국내주식시장 자료와 새로운 모수추정법인 2단계반복갱신법을 사용하여 구조형 부도확률모형들의 예측성과를 분석한다. 본 연구에서 채택한 부도확률모형은 Leland(2006)에서 비교한 구조모형들인 Longstaff and Schwartz(1995), Leland and Toft(1996), Merton(1974)과 Brockman and Turtle(2003)의 down and out 콜 옵션모형이다. 본 연구의 2단계반복갱신법은 기존의 반복갱신법으로 추정할 수 없었 던 Leland and Toft 및 down and out 콜옵션모형의 장벽모수를 일별로 추정할 수 있게 한다. 특히 본 연구는 반복갱신법과 역사적변동성법간 구조모형별 부도확률 예 측성과를 비교한다. 각 구조모형에 적합한 2단계 및 기존의 반복갱신법을 적용할 때 해당 구조모형의 부도확률 예측성과가 다른 모수추정법에 비하여 통계적으로 유의하 게 나타난다. 특히 역사적변동성법에서 부도확률의 예측성과가 나타나지 않던 Leland and Toft의 모형은 2단계반복갱신법을 적용할 경우 부도확률의 예측성과가 통계적 으로 유의하게 나타난다. 또한 Bharath and Shumway(2008)에서와는 달리 2단계반 복갱신법으로 추정한 구조모형 부도확률은 충분통계량의 성질을 만족시킨다

We construct a new iterative method to estimate parameters in structural default probability models and compare their forecasting performances using stock return and accounting data of Korea. To adopt the new iterative method, we select four structural default probability models: Longstaff and Schwartz (1995: LS), Leland and Toft (1996: LT), Merton (1974: DD), and the down and out call option model adopted in Brockman and Turtle (2003: DOC). The new method makes it possible to daily estimate the barrier parameter in LT and DOC, which is not possible under the existing methods [see Vassalou and Xing (2004) and Bharath and Shumway (2008)]. When we adopt the new iterative method to LT and DOC and the existing iterative method to LS and DD in order to analyze out-of-sample and accuracy ratio tests, forecasting performances of above calculated default probabilities are more statistically sufficient than those of default probabilities under other estimation methods. Especially, default probabilities in LT using the new iterative method show statistically supported forcasting performances, though those using the other estimation method have no forecasting performance. Moreover, unlike the results in Bharath and Shumway (2008), default probabilities under the new iterative method satisfy properites of sufficient statistics. The new iterative method is significant in three aspects. First, the new iterative method can offer a new approach to solving two puzzles in asset pricing, the equity premium puzzle and the credit spread puzzle, since the new method provides relatively effective default probabilities among other methods in structural default probability models. Vassalou and Xing (2004) construct rank portfolios by DD’s default probabilities with the existing iterative method used by KMV and study whether equity returns reflect performances for those portfolios. Goldstein (2009) and Chen et al. (2009b) study the relationship between the credit spread puzzle and the equity premium puzzle with structural default probability models. Second, the new iterative method is a simple alternative to MLE(maximum likelihood estimation) which requires heavy and complex calculation [see Ercsson and Reneby (2005), Chen et al. (2009a), Forte and Lovreta (2009), Wong and Choi (2009), and Chen et al. (2010)]. The new iterative method is, in fact, the only method that can daily provide barriers as well as total firm values. Both methods let the first passage time stochastic process of a firm follow a geometric Brownian motion with a drift rate and a volatility rate in, for instance, DOC and LT. Last, we test whether default probabilities of DOC, LT, LS, and DD are sufficient statistics using the forward induction iterative method. Bharath and Shumway (2008) test sufficient statistics only with DD’s default probabilities through the existing forward iterative method of KMV. In order to daily estimate implied barriers and total firm values for DOC using the new iterative method, we follow four steps. First, having fixed a moving window, ordinarily 1 year with 250 business days, we calculate standard deviations of moving averages of equity returns as initial volatility values of a firm under the forward induction. We use the total firm value equal to the book value of the total debt plus the market capitalization as an initial value for a firm on a given day. Second, on that day, we use the Newton-Rahpson method and calculate a implied barrier by taking the derivative of the option value with respect to the barrier. Then using the Newton-Rahpson method again, we calculate a new total firm value by taking the derivative of the option value with respect to the total firm value. Third, we repeat these procedures for the period within the window to get implied barriers and new total firm values. With daily returns of these new total firm values, we compute a new standard deviation and a new mean. Then we update the initial volatility with the new standard deviation for the next iteration, and use the mean as the drift term of the geometric Brownian motion for the total firm value. Finally, we obtain a new volatility following the second and third steps using the new variables generated in the third step. For the period within the window, we repeat the second step until the difference between the existing volatility and the updated volatility converges below the critical level, 10E-4. After we get the converged volatility for the period within the window, we move the window forward by one business day in the forward induction. Next, we follow the first step and repeat the second and third steps until we get a new converged volatility. Hence, we estimate daily total firm values, daily implied barriers, and other parameters to construct the geometric Brownian motion through the full sample period.

충청북도는 웰니스, 생태, 축제 관광을 중심으로 외래관광객이 증가하고 있으나, 내륙 광역지자체를 대상으로 한 관광수요 예측 연구는제한적이다. 이 연구는 충청북도 월별 외래관광객 수요를 대상으로 Prophet, XGBoost, Hybrid 모형의 예측 정확도를 비교하였다. 분석자료는 2020년 1월부터 2025년 8월까지의 68개월 월별 자료이며, 환율, 기상, 축제 외생변수 포함 여부에 따라 다섯 가지 시나리오를구성하였다. 자료는 훈련 48개월, 검증 8개월, 예측 12개월로 구분하였고, 예측 성능은 MAPE와 MASE로 평가하였다. 분석 결과, Hybrid 모형이 가장 우수한 예측 성능을 보였다. 외생변수 미포함 조건에서는 MAPE 14.26%, MASE 0.25를 기록하였고, 외생변수 포함 조건에서는 MAPE 13.71%, MASE 0.26을 기록하였다. Prophet은 성수기 계절 패턴을 비교적 잘 포착하였으나, 외생변수 추가 시 예측 성능이불안정해졌다. XGBoost는 전반적으로 안정적인 예측을 보였지만, 급격한 성수기 수요 증가를 과소예측하는 경향을 나타냈다. 이 결과는충청북도와 같은 내륙형 지역관광 수요예측에서 Hybrid 모형의 활용 가능성을 보여주며, 외생변수는 모형 구조와 자료 특성을 고려해선별적으로 투입할 필요가 있음을 시사한다.

Foreign tourist arrivals in Chungcheongbuk-do have increased through wellness, eco, and festival tourism, but empirical demand forecasting research on this inland province remains limited. This study compares the forecasting accuracy of Prophet, XGBoost, and a Hybrid Prophet–XGBoost model for monthly foreign tourist arrivals in Chungcheongbuk-do. Monthly data from January 2020 to August 2025 were used. Five scenarios were constructed according to the inclusion of exchange rate, weather, and festival variables. The dataset was divided into 48 months for training, 8 months for validation, and 12 months for testing. Forecasting accuracy was evaluated using MAPE and MASE. The results show that the Hybrid model achieved the best overall performance. Without exogenous variables, the Hybrid model recorded a MAPE of 14.26% and a MASE of 0.25. With exogenous variables, it recorded a MAPE of 13.71% and a MASE of 0.26. Prophet captured seasonal peak patterns but showed unstable performance when multiple exogenous variables were added. XGBoost produced relatively stable forecasts but tended to underestimate abrupt peak-season demand. These findings suggest that Hybrid forecasting can improve regional tourism demand prediction, while exogenous variables should be selectively included according to model structure and data characteristics.

본 연구에서는 ARIMA, ARIMA-GARCH, ARIMA-RS 모형들의 토지가격 예측력을 비교하였다. 분석에 사용된 자료는 1987년 1분기부터 2022년 1분기까지 토지가격지수 이며, 예측시계별로 표본내(in-sample)와 표본외(out-of-sample) 예측력을 추정하였 다. 예측력은 평균제곱오차제곱근(RMSFE)와 평균절대오차(MAFE)를 적용하였고 각 모 형들의 예측력이 통계적 유의성이 있는지 여부에 대해 DM 검정 수행하였다. 분석 결과 는 다음과 같다. 첫째, 표본내외의 모든 시계에서 ARIMA(1,1,0)-GARCH 모형의 예측 력 가장 우수하게 나타났다. 둘째, 표본내에서는 ARIMA(1,1,0) 모형의 예측력이 ARIMA(2,1,0) 모형보다 상대적으로 더 우수하게 나타났지만, 표본외에서는 반대로 나 타났다. 또한 표본내에서는 ARIMA(3,1,0)-RS 모형의 예측력이 ARIMA(1,1,0)-RS 모 형의 예측력보다 상대적으로 더 우수하게 나타났지만, 표본외에는 반대로 나타났다. 셋 째, 표본내외의 모든 예측시계에서 국면전환모형(RSM)예측력이 현저히 떨어지는 것으 로 나타났다. 부동산시장의 안정화를 위해 이분산성의 특징을 잘 표착할 수 있는 ARIMA(1,1,0)-GARCH 모형을 예측 모형의 대안으로 고려할 필요성이 있다.

The Purpose of this paper is to compare different models’ forecasting performance of ARIMA, ARIMA-GARCH, ARIMA-RS models in Korean land market covering the period from the first quarter of 1987 to the first quarter of 2022. I carry out in-sample and out-of-sample forecasting power to Korean land price index. For the comparison of forecast powers, I calculate root mean squared forecast errors and mean absoute forecast errors for each forecasting horizon of the second quarter of 2021 to the first quarter of 2022. I test the statistical signification of the predictive comparison results using DM test. The empirical results are as follows: First, the forecasting power of the ARIMA(1,1,0)-GARCH model is the best in all horzions in- and out-of-sample. Second, within the sample, the forecasting power of the ARIMA(1,0) model is relatively better than that of the ARIMA(2,1,0) model, but outside the sample, it is the opposite. In addition, within the sample, the forecasting power of the ARIMA(3,1,0)-RS model is relatively better than that of the ARIMA(1,0)-RS model, but it is contrary to the sample. Third, it is found that the predictive power of the regime-switching model(RSM) is significantly decreased in all predictive horzions in- and out-of-sample. In order to stabilize the real estate market, it is necessary to consider the ARIMA(1,0)-GARCH model as an alternative to the prediction model.

인천공항은 대한민국으로 들어오거나 나가는 관문으로 나라의 이미지에 큰 영향을 미치므로 공항의 서비스 질을 유지하기 위해선 장기적인 공항 이용객 수 예측이 필요하다. 본 연구에서는 인천공항의 이용객 수요를 예측하기 위한 다양한 시계열 모형의 예측성능을 비교하였다. 인천공항 이용객 자료를 2002년 1월부터 2019년 12월까지 월 단위로 수집하여 살펴보면 일반적인 시계열자료에서 보이는 추세성과 계절성을 지니고 있다. 본 연구에서는 추세성과 계절성이 고려된 나이브 기법, 분해법, 지수 평활법, SARIMA, 그리고 PROPHET을 이용하여 단기, 중기, 장기예측 시계열모형을 비교하였다. 분석결과 단기예측은 최근 자료에 가중치를 준 지수 평활법이 우수했고 예상 2020년 연간 이용객 수는 약 7,350만명이다. 3년 후 인 2022년 중기예측은 정상성이 고려된 SARIMA모형이 우수하였고 예상 연간 이용객 수는 약 7,980만명이다. 4단계 인천공항 건설사업이 완료되는 2024년 예상 연간 여객수용 인원은 9,910만명이고 PROPHET모형이 가장 우수하였다.

The Incheon airport is a gateway to and from the Republic of Korea and has a great influence on the image of the country. Therefore, it is necessary to predict the number of airport passengers in the long term in order to maintain the quality of service at the airport. In this study, we compared the predictive performance of various time series models to predict the air passenger demand at Incheon Airport. From 2002 to 2019, passenger data include trend and seasonality. We considered the naive method, decomposition method, exponential smoothing method, SARIMA, PROPHET. In order to compare the capacity and number of passengers at Incheon Airport in the future, the short-term, mid-term, and long-term was forecasted by time series models. For the short-term forecast, the exponential smoothing model, which weighted the recent data, was excellent, and the number of annual users in 2020 will be about 73.5 million. For the medium-term forecast, the SARIMA model considering stationarity was excellent, and the annual number of air passengers in 2022 will be around 79.8 million. The PROPHET model was excellent for long-term prediction and the annual number of passengers is expected to be about 99.0 million in 2024.