• Journal of Korea Water Resources Association
• /
• v.36 no.4
• /
• pp.587-596
• /
• 2003

Application of a catchment modelling system requires recorded information to ascertain the reliability and robustness of the predicted flow conditions. Where this recorded information is not available, the necessary information for reliable and robust predictions must be obtained from other available information sources. The alternative approach presented in this paper used inference models for getting this necessary information that is required to calibrate and validate the catchment modelling system for both an ungauged and a gauged catchments. In this study, inference models were developed for determination of control parameters of the Storm Water Management Model (SWMM), mainly based on landuse component of the catchment, which is a major factor to impact on quantity and quality of catchment runoff. Results from the study show that the new approach for determination of the spatially variable control parameters produced more accurate estimates than a traditional approach. Also, the number of control parameters estimated can be reduced significantly as the proposed method only requires determination of control parameters associated with each land use of the catchment while a traditional approach needs to assign a number of control parameters for a number of subcatchment.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.18 no.4
• /
• pp.287-297
• /
• 2016

This study aims to identify the relationship between climatic factors and stand mean Diameter at Breast Height (DBH) for two major tree species; Pinus densiflora and Quercus mongolica in Andong-dam basin. Forest variables such as age, diameter distribution and number of trees per hectare from the $5^{th}$ and $6^{th}$ National Forest Inventory data were used to develop a DBH estimation model. Climate data were collected from six meteorological observatory station and twelve Automatic Weather System provided by Korea Meteorological Administration to produce interpolated daily average temperature map with Inverse Distance Weighting (IDW) method. Andong-dam basin reflects rugged mountainous terrain, so temperature were adjusted by lapse rate based correction. As a result, predictions of model were consistent with the previous studies; that the rising temperature is negatively related to the growth of Pinus densiflora whereas opposing trend is observed for Quercus mongolica.

• Journal of the Korea Concrete Institute
• /
• v.20 no.2
• /
• pp.221-229
• /
• 2008

An improved unbonded-type column strengthening procedure using wire rope and T-shaped steel plate units was proposed. Eight strengthened columns and an unstrengthened control column were tested under concentric axial load. The main variables considered were the volume ratio of wire rope and the flange width and configuration of T-shaped steel plates. Axial load capacity and ductility ratio of columns tested were compared with predictions obtained from the equation specified in ACI 318-05 and those of conventionally tied columns tested by Chung et al., respectively. In addition, a mathematical model was proposed to evaluate the complete stress-strain relationship of concrete confined by the wire rope and T-plate units. Test results showed that the axial load capacity and ductility of columns increased with the increase of the volume ratio of wire rope and the flange width of T-plates. In particular, at the same lateral reinforcement index, a much higher ductility ratio was observed in the strengthened columns having the volume ratio of wire rope above 0.0039 than in the tied columns. A mathematical model for the stress-strain relationship of confined concrete using the proposed strengthening procedure is developed. The predicted stress-strain curves were in good agreement with test results.

• Journal of Korean Society of Environmental Engineers
• /
• v.32 no.10
• /
• pp.936-941
• /
• 2010

This study was conducted to investigate formaldehyde removals by silver nano-particles attached on the surface of granular activated carbon (Ag-AC) and to compare the results to those obtained with ordinary activated carbon (AC). The BET analysis showed that the overall surface area and the fraction of micropores (less than $20{\AA}$ diameter) of the Ag-AC were significantly decreased because the silver particles blocked the small pores on the surface of the Ag-AC. The formaldehyde removal capacity of the Ag-AC determined using the Freundlich isotherm was higher than that of AC. Despite the decreased BET surface area and micropore volume, the Ag-AC had the increased removal capacity for formaldehyde, presumably due to catalytic oxidation by silver nano-particles. In contrast, the adsorption intensity of the Ag-AC, estimated by 1/n in the Freundlich isotherm equation, was similar to that of the ordinary AC, indicating that the surface modification using silver nano-particles did not affect the adsorption characteristics of AC. In a column experiment, the Ag-AC also showed a longer breakthrough time than that of the AC. Simulation results using the homogeneous surface diffusion model (HSDM) were well fitted to the breakthrough curve of formaldehyde for the ordinary AC, but the predictions showed substantial deviations from the experimental data for the Ag-AC. The discrepancy was due to the catalytic oxidation of silver nano-particles that was not incorporated in the HSDM. Consequently, a new numerical model that takes the catalytic oxidation into accounts needs to be developed to predict the combined oxidation and adsorption process more accurately.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.6
• /
• pp.35-47
• /
• 2020

Box-office prediction is important to movie stakeholders. It is necessary to accurately predict box-office and select important variables. In this paper, we propose a multivariate time series classification and important variable selection method to improve accuracy of predicting the box-office. As a research method, we collected daily data from KOBIS and NAVER for South Korean movies, selected important variables using Random Forest and predicted multivariate time series using Deep Learning. Based on the Korean screen quota system, Deep Learning was used to compare the accuracy of box-office predictions on the 73^rd day from movie release with the important variables and entire variables, and the results was tested whether they are statistically significant. As a Deep Learning model, Multi-Layer Perceptron, Fully Convolutional Neural Networks, and Residual Network were used. Among the Deep Learning models, the model using important variables and Residual Network had the highest prediction accuracy at 93%.

• Korean Journal of Remote Sensing
• /
• v.37 no.6_2
• /
• pp.1881-1890
• /
• 2021

Particulate matter (PM) affects the human, ecosystems, and weather. Motorized vehicles and combustion generate fine particulate matter (PM_2.5), which can contain toxic substances and, therefore, requires systematic management. Consequently, it is important to monitor and predict PM_2.5 concentrations, especially in large cities with dense populations and infrastructures. This study aimed to predict PM_2.5 concentrations in large cities using meteorological and chemical variables as well as satellite-based aerosol optical depth. For PM_2.5 concentrations prediction, a random forest (RF) model showing excellent performance in PM concentrations prediction among machine learning models was selected. Based on the performance indicators R², RMSE, MAE, and MAPE with training accuracies of 0.97, 3.09, 2.18, and 13.31 and testing accuracies of 0.82, 6.03, 4.36, and 25.79 for R², RMSE, MAE, and MAPE, respectively. The variables used in this study showed high correlation to PM_2.5 concentrations. Therefore, we conclude that these variables can be used in a random forest model to generate reliable PM_2.5 concentrations predictions, which can then be used to assess the vulnerability of schools to PM_2.5.

• International Area Studies Review
• /
• v.18 no.3
• /
• pp.47-66
• /
• 2014

A noticeable aspect of Korean firms' outward sequential FDIs to China is that they occur sequentially, which means that they implement the outward FDIs to China with a long-term perspective. To analyze the strategic advantages of sequential investment, we introduce Cournot type quantity competition model. According to the model, three important implications are derived. First, sequential FDIs enhances the Korean parents' production capabilities. Second, the parents are more likely to establish new Chinese subsidiaries as they stay longer in China. Third, the production effect of sequential investments incurs more sequential investments. Some regression models are tested for verifying the predictions. According to empirical results, three important results are found. First, initial entry mode affects the size expansion of the Korean parents. Second, the longer the duration of intial subsidiary in China, the more the sequential investment will be. Third, sequential investments are positively associated with the productivity of the Korean parents.

• Smart Media Journal
• /
• v.9 no.4
• /
• pp.26-35
• /
• 2020

Attempts to predict future stock prices have been studied steadily since the past. However, unlike general time-series data, financial time-series data has various obstacles to making predictions such as non-stationarity, long-term dependence, and non-linearity. In addition, variables of a wide range of data have limitations in the selection by humans, and the model should be able to automatically extract variables well. In this paper, we propose a 'sliding time step normalization' method that can normalize non-stationary data and LSTM autoencoder to compress variables from all variables. and 'moving transfer learning', which divides periods and performs transfer learning. In addition, the experiment shows that the performance is superior when using as many variables as possible through the neural network rather than using only 100 major financial variables and by using 'sliding time step normalization' to normalize the non-stationarity of data in all sections, it is shown to be effective in improving performance. 'moving transfer learning' shows that it is effective in improving the performance in long test intervals by evaluating the performance of the model and performing transfer learning in the test interval for each step.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.42 no.5
• /
• pp.639-650
• /
• 2022

In areas where large-scale sediment transport occurs, it is important to apply appropriate countermeasure method because the phenomenon tends to accelerate by time duration. Among the various countermeasure methods applied so far, beach nourishment needs to be reviewed as an erosion prevention measure because the erosion pattern is mitigated and environmentally friendly depending on the particle size. In the case of beach nourishment. a detailed review is required to determine the size, range, etc., of an appropriate particle diameter. In this study, we investigated the characteristics of the related topographic change using the change in the particle size of nourishment materials, the application of partial area, and the condition under the coexistence of waves and wind as variables because those factors are hard to be analyzed and interpreted within results and limitation of that the existing numerical models are not able to calculate and result out so that it is required that phenomenon or efforts are reviewed at the same time through physical model experiments, field monitoring and etc. So we attempt to reproduce the tendency of beach erosion and deposition and predict possible phenomena in the future using machine learning techniques for phenomena that it is not able to be interpreted by numerical models. we used the hydraulic experiment results for the training data, and the accuracy of the prediction results according to the change in the training method was simultaneously analyzed. As a result of the study it was found that topographic changes using machine learning tended to be similar to those of previous studies in short-term predictions, but we also found differences in the formation of scour and sandbars.

• Journal of Intelligence and Information Systems
• /
• v.28 no.3
• /
• pp.161-183
• /
• 2022

Opinion polls have become a powerful means for election campaigns and one of the most important subjects in the media in that they predict the actual election results and influence people's voting behavior. However, the more active the polls, the more often they fail to properly reflect the voters' minds in measuring the effectiveness of election campaigns, such as repeatedly conducting polls on the likelihood of winning or support rather than verifying the pledges and policies of candidates. Even if the poor predictions of the election results of the polls have undermined the authority of the press, people cannot easily let go of their interest in polls because there is no clear alternative to answer the instinctive question of which candidate will ultimately win. In this regard, we attempt to retrospectively grasp public opinion on the 20th presidential election by applying the 'YouTube Analysis' function of Sometrend, which provides an environment for discovering insights through online big data. Through this study, it is confirmed that a result close to the actual public opinion (or opinion poll results) can be easily derived with simple YouTube data results, and a high-performance public opinion prediction model can be built.