• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.13 no.1
• /
• pp.101-107
• /
• 2013

Is generated in real time in the real world, a large amount of time series data from a wide range of business areas. But it is not easy to determine the optimal model for the description and understanding of the time series data is represented as a dynamic feature. In this study, through the HMM suitable for estimating the short and long-term forecasting model of time-series data to estimate a model that can explain the characteristics of these time series data, it was estimated to predict future patterns of movement. The actual stock market through various materials, information criterion and optimal model estimation for the length of the most efficient data was found to accurately estimate the state of the model. Similar movement patterns predictive than the long-term prediction is more similar to the short-term prediction of the experimental result were found to be.

• Korean Institute of Interior Design Journal
• /
• no.3
• /
• pp.91-99
• /
• 1994

The purpose of this study was to determine the long term effect of a pleasantly designed interior on pro-saptial behavior. For pleasantly designed interior, the existing interior was remodeled through the change of finishing materials for major architectural elements such as wall, floor and ceiling, and changes of furniture and it's arrangement . Pro-spatial behavior was operationalized as seat arranging behavior and measured through the arranged condition and observable arranging behavior. Time-series design, one of quasi-experimental design was used. The data in this study were extracted from an existing field experimental research. Five hundred survey video tapes record during 2 years period were used. In conclusion, the pleasantly designed environment has a long term effect on the pro-spatial behavior change . While self-centered pro-spatial was improved continuously and even reinforced , altruistic pro-spatial behavior was improved but diminished as time passed. There were no differences in the effect between male and female children. The result of the research provide scientific background of an answer to why Interior Design.

• Journal of Korean Society for Atmospheric Environment
• /
• v.19 no.5
• /
• pp.561-568
• /
• 2003

Surface ozone concentrations are highly sensitive to meteorological variability. Therefore, in order to reveal the long-term changes in ozone due to the changes in precursor emissions, we need to remove the effects of meteorological fluctuations on the annual distribution of surface ozone. In this paper, the meteorologically adjusted trends of daily maximum surface ozone concentrations in two major Korean cities (Seoul and Busan) are investigated based on ozone data from 11 (Seoul) and 6 (Busan) sites over the period 1992 ∼ 2000. The original time series consisting of the logarithm of daily maximum ozone concentrations are splitted into long-term, seasonal and short-term component using Kolmogorov-Zurbenko (KZ) filter. Meteorological effects are removed from filtered ozone series using multiple linear regression based on meteorologcial variables. The long-term evolution of ozone forming capability due to changes in precursor emission can be obtained applying the KZ filter to the residuals of the regression. The results indicated that meteorologically adjusted long-term daily maximum ozone concentrations had a significant upward trend (Seoul: ＋ 3.02% yr$^{-1}$ , Busan: ＋ 3.45% yr$^{-1}$ ). These changes of meteorologically adjusted ozone concentrations represent the effects of changing background ozone concentrations as well as the more localized changes in emissions.

• The Bulletin of The Korean Astronomical Society
• /
• v.41 no.1
• /
• pp.45.3-46
• /
• 2016

Solar activity shows a self-similarity as it has many periods of activity cycle in the time series of long-term observation, such as 13.5, 51, 150, 300 days, and 11, 88 years and so on. Since fractal dimension is a quantitative parameter for this kind of an irregular time series, we applied this method to long-term observations including sunspot number, total solar irradiance, and 3.75 GHz solar radio flux to predict the start and maximum times as well as expected maximum sunspot number for the next solar cycle. As a result, we found that the radio flux data tend to have lower fractal dimensions than the sunspot number data, which means that the radio emission from the sun is more regular than the solar activity expressed by sunspot number. Based on the relation between radio flux of 3.75 GHz and sunspot number, we could calculate the expected maximum sunspot number of solar cycle 24 as 156, while the observed value is 146. For the maximum time, estimated mean values from 7 different observations are January 2013 and this is quite different to observed value of February 2014. We speculate this is from extraordinary extended properties of solar cycle 24. As the cycle length of solar cycle 24, 10.1 to 12.8 years are expected, and the mean value is 11.0. This implies that the next solar cycle will be started at December 2019.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.29 no.5
• /
• pp.505-511
• /
• 2023

In the present study, a deep learning model was established to predict the motion response of small fishing vessels. Hydrodynamic performances were evaluated for two small fishing vessels for the dataset of deep learning model. The deep learning model of the Long Short-Term Memory (LSTM) which is one of the recurrent neural network was utilized. The input data of LSTM model consisted of time series of six(6) degrees of freedom motions and wave height and the output label was selected as the time series data of six(6) degrees of freedom motions. The hyperparameter and input window length studies were performed to optimize LSTM model. The time series motion response according to different wave direction was predicted by establised LSTM. The predicted time series motion response showed good overall agreement with the analysis results. As the length of the time series increased, differences between the predicted values and analysis results were increased, which is due to the reduced influence of long-term data in the training process. The overall error of the predicted data indicated that more than 85% of the data showed an error within 10%. The established LSTM model is expected to be utilized in monitoring and alarm systems for small fishing vessels.

• Journal of Environmental Science International
• /
• v.26 no.2
• /
• pp.221-230
• /
• 2017

To determine the effect of air pollution reduction policies, the long-term trend of air pollutants should be analyzed. Kolmogorov-Zurbenko (KZ) filter is a low-pass filter, produced through repeated iterations of a moving average to separate each variable into its temporal components. The moving average for a KZ(m, p) filter is calculated by a filter with window length m and p iterations. The output of the first pass subsequently becomes the input for the next pass. Adjusting the window length and the number of iterations makes it possible to control the filtering of different scales of motion. To break down the daily mean $PM_{10}$ into individual time components, we assume that the original time series comprises of a long-term trend, seasonal variation, and a short-term component. The short-term component is attributable to weather and short-term fluctuations in precursor emissions, while the seasonal component is a result of changes in the solar angle. The long-term trend results from changes in overall emissions, pollutant transport, climate, policy and/or economics. The long-term trend of the daily mean $PM_{10}$ decreased sharply from $59.6ug/m^3$ in 2002 to $44.6ug/m^3$ in 2015. This suggests that there was a long-term downward trend since 2005. The difference between the unadjusted and meteorologically adjusted long-term $PM_{10}$ is small. Therefore, we can conclude that $PM_{10}$ is unaffected by the meteorological variables (total insolation, daily mean temperature, daily mean relative humidity, daily mean wind speed, and daily mean local atmospheric pressure) in Busan.

• The Journal of Bigdata
• /
• v.7 no.2
• /
• pp.217-224
• /
• 2022

Since online shopping has become common, people can easily buy fashion goods anytime, anywhere. Therefore, consumers quickly respond to various environmental variables such as weather and sales prices. Therefore, utilizing big data for efficient inventory management has become very important in the fashion industry. In this paper, the changes in sales volume of fashion goods due to changes in temperature is analyzed via the proposed big data analysis algorithm by utilizing actual big data from Korean fashion company 'A'. According to the simulation results, it was confirmed that Bidirectional-LSTM(Bi-LSTM) compared to LSTM(Long Short-Term Memory) takes more simulation time about more than 50%, but the prediction accuracy of non-periodic time series data such as clothing product sales data is the same.

• Communications for Statistical Applications and Methods
• /
• v.31 no.2
• /
• pp.213-234
• /
• 2024

In recent decades, increasing research attention has been directed toward predicting the price of stocks in financial markets using deep learning methods. For instance, recurrent neural network (RNN) is known to be competitive for datasets with time-series data. Long short term memory (LSTM) further improves RNN by providing an alternative approach to the gradient loss problem. LSTM has its own advantage in predictive accuracy by retaining memory for a longer time. In this paper, we combine both supervised and unsupervised dimension reduction methods with LSTM to enhance the forecasting performance and refer to this as a dimension reduction based LSTM (DR-LSTM) approach. For a supervised dimension reduction method, we use methods such as sliced inverse regression (SIR), sparse SIR, and kernel SIR. Furthermore, principal component analysis (PCA), sparse PCA, and kernel PCA are used as unsupervised dimension reduction methods. Using datasets of real stock market index (S&P 500, STOXX Europe 600, and KOSPI), we present a comparative study on predictive accuracy between six DR-LSTM methods and time series modeling.

• Health Policy and Management
• /
• v.32 no.4
• /
• pp.389-399
• /
• 2022

Background: This study aimed to analyze changes in medical utilization and cost before and after long-term care (LTC) implementation. Methods: We used the National Health Information Database from National Health Insurance Service. The participants were selected who had a new LTC grade (grade 1-5) for 2015. Medical utilization was analyzed before and after LTC implementation. Segmented regression analysis of interrupted time series was conducted to evaluate the overall effect of the LTC implementation on medical costs. Results: The total number of participants was 41,726. A major reason for hospitalization in grade 1 was cerebrovascular diseases, and dementia was the top priority in grade 5. The proportion of hospitalization in grade 1 increased sharply before LTC implementation and then decreased. In grade 5, it increased before LTC implementation, but there was no significant difference after LTC implementation. As for medical cost, in grades 1 to 4, the total cost increased sharply before the LTC implementation, but thereafter, changes in level and trend tended to decrease statistically, and for grade 5, immediately after LTC implementation, the level change was decreasing, but thereafter, the trend change was increasing. Conclusion: Long-term care grades showed different medical utilization and cost changes. Long-term care beneficiaries would improve their quality of life by adequately resolving their medical needs by their grades.

• Journal of Information Processing Systems
• /
• v.14 no.6
• /
• pp.1508-1520
• /
• 2018

In this paper, we propose an improved model to provide users with a better long-term prediction of waterworks operation data. The existing prediction models have been studied in various types of models such as multiple linear regression model while considering time, days and seasonal characteristics. But the existing model shows the rate of prediction for demand fluctuation and long-term prediction is insufficient. Particularly in the deep running model, the long-short-term memory (LSTM) model has been applied to predict data of water purification plant because its time series prediction is highly reliable. However, it is necessary to reflect the correlation among various related factors, and a supplementary model is needed to improve the long-term predictability. In this paper, convolutional neural network (CNN) model is introduced to select various input variables that have a necessary correlation and to improve long term prediction rate, thus increasing the prediction rate through the LSTM predictive value and the combined structure. In addition, a multiple linear regression model is applied to compile the predicted data of CNN and LSTM, which then confirms the data as the final predicted outcome.