• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.2
• /
• pp.81-86
• /
• 2022

Given that time series are used in various fields, such as finance, IoT, and manufacturing, data analytical methods for accurate time-series forecasting can serve to increase operational efficiency. Among time-series analysis methods, multi-horizon forecasting provides a better understanding of data because it can extract meaningful statistics and other characteristics of the entire time-series. Furthermore, time-series data with exogenous information can be accurately predicted by using multi-horizon forecasting methods. However, traditional deep learning-based models for time-series do not account for the heterogeneity of inputs. We proposed an improved time-series predicting method, called the temporal fusion transformer method, which combines multi-horizon forecasting with interpretable insights into temporal dynamics. Various real-world data such as stock prices, fine dust concentrates and electricity consumption were considered in experiments. Experimental results showed that our temporal fusion transformer method has better time-series forecasting performance than existing models.

• Journal of Distribution Science
• /
• v.10 no.11
• /
• pp.31-37
• /
• 2012

Purpose - This study was conducted under the assumption that brand A, a store of company Z of Pangyo, with a new store at Pangyo station is targeting the Bundang-gu area of the newly developed city of Seongnam. Research design, data, methodology - As a result of demand forecasting using geometric series models, an extrapolation of past trends provided the coefficient estimates, without utilizing regression analysis on a constant increase in children's wear, for which the population size and estimated parameter were required. Results - Demand forecasting on the basis of past trends indicates the likelihood that sales of discount stores in the Bundang area, where brand A currently has a presence, would fetch a higher estimated value than that of the average discount store in the country during 2015. If past trends persist, future sales of operational stores are likely to increase. Conclusions - In evaluating location using the simple weighting model, Seohyun Lotte Mart obtained a high rating amongst new stores in Pangyo, on the basis of accessibility, demand class, and existing stores. Therefore, when opening a new counter at a relevant store, a positive effect can be predicted.

• The Korean Journal of Applied Statistics
• /
• v.18 no.1
• /
• pp.137-146
• /
• 2005

Neural network is one of the most notable technique. It usually provides more powerful forecasting models than the traditional time series techniques. Employing the Ensemble technique in forecasting model, one should provide a initial distribution. Usually the uniform distribution is assumed so that the initialization is noninformative. However, it would be expected a sequential informative initialization based on data rather than the uniform initialization gives further reduction in forecasting error. In this note, a modified Ensemble algorithm using sequential initial probability is developed. The sequential distribution is designed to have much weight on the recent data.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.27 no.3
• /
• pp.406-415
• /
• 2024

As the equipment of the military has become more advanced and expensive, the cost of securing spare parts is also constantly increasing along with the increase in equipment assets. In particular, forecasting demand for spare parts one of the important management tasks in the military, and the accuracy of these predictions is directly related to military operations and cost management. However, because the demand for spare parts is intermittent and irregular, it is often difficult to make accurate predictions using traditional statistical methods or a single statistical or machine learning model. In this paper, we propose a model that can increase the accuracy of demand forecasting for irregular patterns of spare parts demanding by using a combination of statistical and machine learning algorithm, and through experiments on Cheonma spare parts demanding data.

• Journal of Korea Water Resources Association
• /
• v.45 no.11
• /
• pp.1187-1199
• /
• 2012

A hybrid forecasting scheme based on wavelet decomposition coupled to a support vector machine model is presented for water demand series that exhibit nonlinear behavior. The use of wavelet transform followed by the SVM model of each leading component is explored as a model for water demand data. The proposed forecasting model yields better results than a traditional ARIMA time series forecasting model in terms of self-prediction problem as well as reproducing the properties of the observed water demand data by making use of the advantages of wavelet transform and SVM model. The proposed model can be used to substantially and significantly improve the water demand forecasting and utilized in a real operation.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 1997.10a
• /
• pp.105-108
• /
• 1997

Previous studies in stock market predictions using artificial intelligence techniques such as artificial neural networks and case-based reasoning, have focused mainly on spot market prediction. Korea launched trading in index futures market (KOSPI 200) on May 3, 1996, then more people became attracted to this market. Thus, this research intends to predict the daily up/down fluctuant direction of the price for KOSPI 200 index futures to meet this recent surge of interest. The forecasting methodologies employed in this research are the integration of genetic algorithm and artificial neural network (GAANN) and the integration of genetic algorithm and case-based reasoning (GACBR). Genetic algorithm was mainly used to select relevant input variables. This study adopts the categorical data preprocessing based on expert's knowledge as well as traditional data preprocessing. The experimental results of each forecasting method with each data preprocessing method are compared and statistically tested. Artificial neural network and case-based reasoning methods with best performance are integrated. Out-of-the Model Integration and In-Model Integration are presented as the integration methodology. The research outcomes are as follows; First, genetic algorithms are useful and effective method to select input variables for Al techniques. Second, the results of the experiment with categorical data preprocessing significantly outperform that with traditional data preprocessing in forecasting up/down fluctuant direction of index futures price. Third, the integration of genetic algorithm and case-based reasoning (GACBR) outperforms the integration of genetic algorithm and artificial neural network (GAANN). Forth, the integration of genetic algorithm, case-based reasoning and artificial neural network (GAANN-GACBR, GACBRNN and GANNCBR) provide worse results than GACBR.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 2005.05a
• /
• pp.795-797
• /
• 2005

During the past few years, various traffic-flow forecasting models, i.e. an ARIMA, an ANN, and so on, have been developed to predict more accurate traffic flow. However, these models analyze historical data in an attempt to predict future value of a variable of interest. They make use of the following basic strategy. Past data are analyzed in order to identify a pattern that can be used to describe them. Then this pattern is extrapolated, or extended, into the future in order to make forecasts. This strategy rests on the assumption that the pattern that has been identified will continue into the future. So ARIMA or ANN models with its traditional architecture cannot be expected to give good predictions unless this assumption is valid; The statistical models in particular, the time series models are deficient in the sense that they merely extrapolate past patterns in the data without reflecting the expected irregular and infrequent future events Also forecasting power of a single model is limited to its accurate. In this paper, we compared with an ANN model and ARIMA model and tried to combine an ARIMA model and ANN model for obtaining a better forecasting performance. In addition to combining two models, we also introduced judgmental adjustment technique. Our approach can improve the forecasting power in traffic flow. To validate our model, we have compared the performance with other models. Finally we prove that the proposed model, i.e. ARIMA + ANN + Judgmental Adjustment, is superior to the other model.

• Journal of IKEEE
• /
• v.23 no.1
• /
• pp.127-133
• /
• 2019

Electric power demand forecasting is one of the important research areas for future smart grid introduction. However, It is difficult to predict because it is affected by many external factors. Traditional methods of forecasting power demand have been limited in making accurate prediction because they use raw power data. In this paper, a probability-based CRBM is proposed to solve the problem of electric power demand prediction using raw power data. The stochastic model is suitable to capture the probabilistic characteristics of electric power data. In order to compare the mid-term power demand forecasting performance of the proposed model, we compared the performance with Recurrent Neural Network(RNN). Performance comparison using electric power data provided by the University of Massachusetts showed that the proposed algorithm results in better performance in mid-term energy demand forecasting.

• Journal of The Korean Society of Agricultural Engineers
• /
• v.63 no.1
• /
• pp.11-25
• /
• 2021

The objective of this study was to analyze the impact of activation functions on flood forecasting model based on Artificial neural networks (ANNs). The traditional activation functions, the sigmoid and tanh functions, were compared with the functions which have been recently recommended for deep neural networks; the ReLU, leaky ReLU, and ELU functions. The flood forecasting model based on ANNs was designed to predict real-time runoff for 1 to 6-h lead time using the rainfall and runoff data of the past nine hours. The statistical measures such as R2, Nash-Sutcliffe Efficiency (NSE), Root Mean Squared Error (RMSE), the error of peak time (ETp), and the error of peak discharge (EQp) were used to evaluate the model accuracy. The tanh and ELU functions were most accurate with R2=0.97 and RMSE=30.1 (㎥/s) for 1-h lead time and R2=0.56 and RMSE=124.6~124.8 (㎥/s) for 6-h lead time. We also evaluated the learning speed by using the number of epochs that minimizes errors. The sigmoid function had the slowest learning speed due to the 'vanishing gradient problem' and the limited direction of weight update. The learning speed of the ELU function was 1.2 times faster than the tanh function. As a result, the ELU function most effectively improved the accuracy and speed of the ANNs model, so it was determined to be the best activation function for ANNs-based flood forecasting.

• The Journal of the Korea Contents Association
• /
• v.10 no.1
• /
• pp.372-380
• /
• 2010

In this study, we consider a mid-and-long term technological forecasting method based on simulation technique. We, first, gather information about a point of appearance time of new technologies which will be developed in the future and influence relationship among those technologies by Delphi survey. And then we propose a simulation-based heuristic approach searching for the key technology among new technologies which will be developed to attain a normative objective using the Delphi data. We also provide the range of occurrence time for individual technology and define key technologies in this study in contrast that a expert's estimate to occurrence time is only one point in traditional Delphi survey. The information for key technologies which are detected by this procedure gives priorities of R&D planning and aids the R&D planner or project manager in resource allocation.