• Communications for Statistical Applications and Methods
• /
• v.30 no.6
• /
• pp.577-587
• /
• 2023

Conventional categorical data imputation techniques, such as mode imputation, often encounter issues related to overestimation. If the variable has too many categories, multinomial logistic regression imputation method may be impossible due to computational limitations. To rectify these limitations, we propose a two-stage imputation method. During the first stage, we utilize the Boruta variable selection method on the complete dataset to identify significant variables for the target categorical variable. Then, in the second stage, we use the important variables for the target categorical variable for logistic regression to impute missing data in binary variables, polytomous regression to impute missing data in categorical variables, and predictive mean matching to impute missing data in quantitative variables. Through analysis of both asymmetric and non-normal simulated and real data, we demonstrate that the two-stage imputation method outperforms imputation methods lacking variable selection, as evidenced by accuracy measures. During the analysis of real survey data, we also demonstrate that our suggested two-stage imputation method surpasses the current imputation approach in terms of accuracy.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.18 no.3
• /
• pp.15-19
• /
• 2004

We presented the holographic memory system using a multi-mode optical fiber. In this system, we multiplexed multiple holograms using the lights from multi-mode optical fibers as reference beams. The proposed system has two degrees of multiplexing freedom: speckle pattern of light coming from multi-mode fiber and spatial location of multi-mode fibers. Therefore, the data addressing in the system can be performed by mode and spatial multiplexings.

• Wind and Structures
• /
• v.38 no.5
• /
• pp.381-398
• /
• 2024

This study presents a mode analysis of 3D turbulent velocity data around a square-section building model to identify the dynamic system for Kármán-type vortex shedding. Proper orthogonal decomposition (POD) was first performed to extract the significant 3D modes. Magnitude-squared coherence was then applied to detect the phase consistency between the modes, which were roughly divided into three groups. Group 1 (modes 1-4) depicted the main vortex shedding on the wake of the building, with mode 2 being controlled by the inflow fluctuation. Group 2 exhibited complex wake vortexes and single-sided vortex phenomena, while Group 3 exhibited more complicated phenomena, including flow separation. Subsequently, a third-order polynomial regression model was used to fit the dynamics system of modes 1, 3, and 4, which revealed average trend of the state trajectory. The two limit cycles of the regression model depicted the two rotation directions of Kármán-type vortex. Furthermore, two characteristic periods were identified from the trajectory generated by the regression model, which indicates fast and slow motions of the wake vortex. This study provides valuable insights into 3D mode morphology and dynamics of Kármán-type vortex shedding that helps to improve design and efficiency of structures in turbulent flow.

• Proceedings of the Computational Structural Engineering Institute Conference
• /
• 2006.04a
• /
• pp.778-785
• /
• 2006

In this paper, two possible ways for mode shape expansion are proposed and opened for discussion for future use. The first method minimizes the modal flexibility error between the experimental and analytical mode shapes corresponding to the measured DOFs to find the multiplication matrix which can be treated as the least-squares minimization problem. In the second method, Normalized Modal Difference (NMD) is used to calculate multiplication matrix using the analytical DOFs corresponding to measured DOfs. This matrix is then used to expand the measured mode shape to unmeasured DOFs. A simulated simply supported beam is used to demonstrate the performance of the methods. These methods are then compared with two most promising existing methods namely Kidder dynamic expansion and Modal expansion methods. It is observed that the performance of the modal flexibility method is comparable with existing methods. NMD also have the potential to expand the mode shapes though it is seen more sensitive to the distribution of error between FEM and actual test data.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2023.05a
• /
• pp.179-179
• /
• 2023

This study suggests a new approach of water level forecasting for extended lead times using original data preprocessing with variational mode decomposition (VMD). Here, two machine learning algorithms including light gradient boosting machine (LGBM) and random forest (RF) were considered to incorporate extended lead times (i.e., 5, 10, 15, 20, 25, 30, 40, and 50 days) forecasting of water levels. At first, the original data at two water level stations (i.e., SW173 and SW269 in Bangladesh) and their decomposed data from VMD were prepared on antecedent lag times to analyze in the datasets of different lead times. Mean absolute error (MAE), root mean squared error (RMSE), and mean squared error (MSE) were used to evaluate the performance of the machine learning models in water level forecasting. As results, it represents that the errors were minimized when the decomposed datasets were considered to predict water levels, rather than the use of original data standalone. It was also noted that LGBM produced lower MAE, RMSE, and MSE values than RF, indicating better performance. For instance, at the SW173 station, LGBM outperformed RF in both decomposed and original data with MAE values of 0.511 and 1.566, compared to RF's MAE values of 0.719 and 1.644, respectively, in a 30-day lead time. The models' performance decreased with increasing lead time, as per the study findings. In summary, preprocessing original data and utilizing machine learning models with decomposed techniques have shown promising results for water level forecasting in higher lead times. It is expected that the approach of this study can assist water management authorities in taking precautionary measures based on forecasted water levels, which is crucial for sustainable water resource utilization.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.16 no.3
• /
• pp.423-428
• /
• 2021

The performance of wind profilers were evaluated in Bukgangneung and Changwon, where both wind profilers and radiosonde are being operated. From December 2019 to November 2020, the acquisition rate of wind speed data by altitude and month was investigated and compared with the wind speed of radiosonde. The acquisition rate of normal data increased at the minimal observable altitude, and then decreased again from 2 km or more at the two observation sites. The acquisition rate was high in summer (June, July, August), and low in winter (December, January, February). The results compared with radiosonde showed high correlation of 0.9 in the high mode and 0.8 in the low mode.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2002.11a
• /
• pp.319.2-319
• /
• 2002

The main obstacle to high track density in HDD id the structural resonances of the suspension. The most critical mode is sway mode and second torsion mode, when a data is read and written. It is common fact that the effect of two modes is smaller when a thickness is bulky. But the stiffness of suspension is smaller, the slider can follow a disk better. Because these two fact are reciprocal, a compromise is needed. So we investigated another method to improve band width without changing of the thickness of suspension but with changing of the shape. (omitted)

• Journal of the Korea Society for Simulation
• /
• v.21 no.2
• /
• pp.51-58
• /
• 2012

In IEEE 802.16j MMR protocol, two modes about usage of RS are proposed; one is transparent mode to enhance data throughput and the other is non-transparent mode to extend coverage. In this paper, we focus on non-transparent mode and find that the mode can also improve data throughput. Therefore, we analyze data throughput on various RS topology and their extended coverage area by simulation in IEEE 802.16j non-transparent mode. We also compare the simulation results with the single MR-BS system of which coverage is extended by higher transmission power. From the comparisons of simulation results, we see that higher throughput can be obtained in the proposed non-transparent mode.

• ETRI Journal
• /
• v.39 no.4
• /
• pp.582-591
• /
• 2017

To address the wire complexity problem in large-scale globally asynchronous, locally synchronous systems, a current-mode ternary encoding scheme was devised for a two-phase asynchronous protocol. However, for data transmission through a very long wire, few studies have been conducted on reducing the long propagation delay in current-mode circuits. Hence, this paper proposes a current steering logic (CSL) that is able to minimize the long delay for the devised current-mode ternary encoding scheme. The CSL creates pulse signals that charge or discharge the output signal in advance for a short period of time, and as a result, helps prevent a slack in the current signals. The encoder and decoder circuits employing the CSL are implemented using $0.25-{\mu}m$ CMOS technology. The results of an HSPICE simulation show that the normal and optimal mode operations of the CSL achieve a delay reduction of 11.8% and 28.1%, respectively, when compared to the original scheme for a 10-mm wire. They also reduce the power-delay product by 9.6% and 22.5%, respectively, at a data rate of 100 Mb/s for the same wire length.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2006.05a
• /
• pp.1366-1371
• /
• 2006

A structural coupling method is developed for the dynamic analysis of a nonlinear structure with concentrated nonlinear hinge joints or sliding lines. Component mode synthesis method is extended to couple substructures and the nonlinear models. In order to verify the improved coupling method, a numerical plate model consisting of two substructures and torsional springs, is synthesized by using the proposed method and its model parameters are compared with analysis data. Then the coupling method is applied to a three-substructure-model with the nonlinearity of sliding lines between the substructures. The coupled structural model is verified from its dynamic analysis. The analysis results show that the improved coupling method is adequate for the structural nonlinear analyses with the nonlinear hinge and sliding mode condition.