• Proceedings of the Korea Information Processing Society Conference
• /
• 2001.10a
• /
• pp.363-366
• /
• 2001

It is very difficult to maintain the performance of estimation models for the new breed of projects since the computing environment changes so rapidly in terms of programming languages, development tools, and methodologies. So, we propose to use the relevant cases for a neural network model, whose cost is the decreased number of cases. To balance the relevance and data availability, the qualitative input factors are used as criteria of data classification. With the data sets that have the same value for certain qualitative input factors, we can eliminate the factors from the model making reduced neural network models. So we need to seek the optimally reduced neural network model among them. To find the optimally case-selective neural network, we propose the search techniques and sensitivity analysis between data points and search space.

• Smart Media Journal
• /
• v.10 no.3
• /
• pp.9-22
• /
• 2021

Basically, machine learning models use input data to produce results. Sometimes, the input data is too complicated for the models to learn useful patterns. Therefore, feature engineering is a crucial data preprocessing step for constructing a proper feature set to improve the performance of such models. One of the most efficient methods for automating feature engineering is the autoencoder, which transforms the data from its original space into a latent space. However certain factors, including the datasets, the machine learning models, and the number of dimensions of the latent space (denoted by k), should be carefully considered when using the autoencoder. In this study, we design a framework to compare two data preprocessing approaches: with and without autoencoder and to observe the impact of these factors on autoencoder. We then conduct experiments using autoencoders with classifiers on popular datasets. The empirical results provide a perspective regarding the best suited autoencoder for these factors.

• Journal of Information Technology Services
• /
• v.10 no.2
• /
• pp.177-188
• /
• 2011

This paper proposes a DEA model for the performance evaluation of each brand of coking coals in an integrated steel mill. The performance is defined as the efficiency which is the ratio of two linear combinations of the output factors to the input factors. There is only one input factor considered in the model : unit price of each brand based on CIF. Five output factors are chosen in consideration of their impact to the quality of cokes such as Ash, VM, LMF, TD, and Rm. Some of the output factors are treated as undesirable in DEA model because the quality criteria are given by the range. The CCR and BCC efficiencies are derived by the DEA model, and the scale efficiency is calculated, too. Each brand of coking coal is classified into four categories according to the CCR and BCC efficiencies, and the most inferior brands are identified as a result. The impact of the input and output factors to the efficiency is analyzed using a multiple regression, then the unit price is revealed as the most critical among them. Also, ANOVA results show that there exist efficiency differences among the coal types and the countries imported, respectively. Finally, the quantitative projection for the inefficient brands is performed if they are to be efficient. The result could be utilized in selecting the good or bad brands of coking coal based on the efficiency in an integrated steel mill. Also, this model will be used to assess the relative efficiency of a new brand of coking coal if it is a candidate to be imported.

• Journal of Korea Water Resources Association
• /
• v.36 no.4
• /
• pp.575-586
• /
• 2003

In case of continuous input of a pollutant, dispersion characteristics do not change much with changing dispersion coefficient, but that of an instantaneous input is very sensitive to the changes of dispersion coefficient. The characteristics of behavior of instantaneous input of a pollutant at the downstream of Han river were analyzed in this paper Field measurement of hydraulic and water quality factors at the downstream of Han river were conducted at low flow condition. The hydraulic factors were used to estimate the longitudinal dispersion coefficient, and the reasonable empirical equations for longitudinal dispersion coefficient at the downstream of Han river were suggested. The measured concentrations of BOD were closely matched with the calculated ones from RMA-4 model. In case of instantaneous input, range of dispersion, transport pathway and the traveltimes of the first and maximum concentration with variation of the longitudinal dispersion coefficients and water levels of downstream boundary were evaluated in this paper.

• Proceedings of the Korean Geotechical Society Conference
• /
• 2010.03a
• /
• pp.771-776
• /
• 2010

Since the importance of seismic design is greater, dynamic analysis is more widely using than past. The input motion is one of the most important factors of dynamic analysis. However, in Korea input motions are selected from U.S. and Japan those are captured from large magnitude earthquakes without considering seismic environment or generated in frequency domain. In this research, the methodology for generating input motions those are considered seismic environment and design code is proposed. The seismic environment compatibility is considered by performing deaggregation and the design code compatibility is considered by time-domain artificial time history accelration generation method. The results shows that seismic environment and design code compatible input motions are successfully generated.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.26 no.10
• /
• pp.2067-2072
• /
• 2002

The closed-loop state and input observer is a pole-placement type observer and estimates unknown state and input variables simultaneously. Pole-placement type observers may have poor performances with respect to modeling error and sensing bias error. The effects of these ill-conditioning factors must be minimized for the robust performance in designing observers. In this paper, the steady-state performance of the closed-loop state and input observer is investigated quantitatively and is represented as the estimation error bounds. The performance indices are selected from these error bounds and are related to the robustness with respect to modeling errors and sensing bias. By considering both transient and steady-state performance, the main performance index is determined as the condition number of the eigenvector matrix based on $L_2$-norm.

• Journal of Korean Physical Therapy Science
• /
• v.3 no.2
• /
• pp.1065-1077
• /
• 1996

Fees for medical insurance services in Korea has not being been set based on service costing. Recently in the USA, fees for physican services are determined by developing and applying Resoure Based Relative Value Scales (RBRVS). This study attempts to develop relative value scales for personal factors of rehabilitative and physical therapeutic services. The personal factors were classified into four categories as having been done in the USA;service time, treatment technology and physical efforts, mental efforts and judgement, and stress. Input factors were measured using Magnitude Estimation Method (MEM), and relative value units were calculated for each of twenty eight rehabilitative and physical therapeutic services. Results show that service time surveyed differs from that provided in the public fee schedules in 24 services; the three personal factors but the service time are highly correlated; the physical therapists hold treatment technology and physical efforts to be the most important factor in setting the for services; and that relative values developed for noninsurance services such as Silver Spike Electrode (SSP) and Proprioceptive Neuromuscular Facilitation (PNF) are higher than those of similar insurance services. The policy implications and measures for improvement for the above findings were suggested respectively.

• Proceedings of the KIEE Conference
• /
• 1999.07c
• /
• pp.1063-1066
• /
• 1999

This paper presents an extended integral control with the PID controller by introducing the delay and decaying factors. The convolution integral control scheme is developed by substituting proportional convolution integral controls for the proportional-integral control. So far, the integral part of the PI controller produces a signal that is proportional to the time integral of the input of the controller. The steady-state operation points are affected forever by the errors in the past due to the input signal containing the information of the errors in the past. These phenomina may cause some disturbances for other control purposes related to the given PI control. Introduction of forgetting factors of the error in the past can resolve the disturbance problems. Various forgetting factors are developed using the delay, the decaying factors, and the combination of the delay and the decaying factors. The proposed various extended integral control schemes can be applicable to corresponding PI control designs in which the error in the past may badly affect to the current steady-state operation points and may cause some disturbances for other control purposes.

• The Transactions of the Korean Institute of Electrical Engineers A
• /
• v.49 no.7
• /
• pp.345-349
• /
• 2000

This paper presents an extended integral control with the PI controller by introducing the delay and decaying factors. The extended integral control scheme is developed by substituting the proportional convolution integral control for the PI(Proportional Integral) control. So far, the integral part of PI controller produces a signal that is proportional to the time integral of the input signal to the controller. The steady-state operation points are affected forever by errors in the past due to the input signal containing the information of the error in the past. These phenomena may cause some disturbances for other control purposes related to the given PI control. Introduction of forgetting factors to the error in the past can resolve the disturbance problems. Various forgetting factors are developed using the delay elements, the decaying factors, and the combination of the delay and decaying factors. The proposed various extended integral control schemes can be applicable to the corresponding PI control designs in which the error in the past may badly affect the current steady-state operation points and may cause some disturbances for other control purposes.

• Proceedings of the KWS Conference
• /
• 2002.10a
• /
• pp.714-719
• /
• 2002

In The Great Hanshin-Awaji Earthquake, the general yield brittle fractures were observed in beam-column connections of steel building frames. Among many influencing factors which affect the general yield brittle fracture, it can be considered that fracture toughness has substantial effects. Some studies are making clear the required toughness for the base metal and the weld metal, but general values are not proposed. Moreover, it seems that it is also important to pay attention to the toughness decrease in the weld heat affected zone (weld HAZ), because the toughness decrease occurs in the HAZs of mild steel. In this paper, the relationship between toughness of simulated HAZs of "the rolled steels for building structures (SN)" and the weld heat-input limit of the SN steel are investigated, in an attempt to provide the required toughness for HAZs. The relationships between the increase of the hardness value and toughness, and changes of microstructure after weld heat-input are also discussed. The main results are summarized as follows. 1) The SN400B can keep its toughness at higher heat-inputs compare to the SN490Bs. 2) The steel grade, which becomes harder than other steel grades at the same heat-input, has smaller absorbed energy and smaller limit of heat-input. 3) The weld heat-input limit of the SN400B and the SN490B are proposed separately for some required toughness values.