• Journal of the Korean Operations Research and Management Science Society
• /
• v.38 no.1
• /
• pp.139-152
• /
• 2013

The purpose of this study is to claim the validity of tuning the architecture of neural network models for multi-class classification. A neural network model for multi-class classification is basically constructed by building a series of neural network models for binary classification. Building a neural network model, we are required to set the values of parameters such as number of hidden nodes and weight decay parameter in advance, which draws special attention as the performance of the model can be quite different by the values of the parameters. For better performance of the model, it is absolutely necessary to have a prior process of tuning the parameters every time the neural network model is built. Nonetheless, previous studies have not mentioned the necessity of the tuning process or proved its validity. In this study, we claim that we should tune the parameters every time we build the neural network model for multi-class classification. Through empirical analysis using wine data, we show that the performance of the model with the tuned parameters is superior to those of untuned models.

• Proceedings of the Korean Society for Technology of Plasticity Conference
• /
• 1999.03b
• /
• pp.84-87
• /
• 1999

This paper presents a preform design method that combines the analytic method and inference of known knowledge with neural network. The analytic method is a finite element method that is used to simulate backward extrusion with pre-defined process parameters. The multi-layer network and back-propagation algorithm are utilized to learn the training examples from the simulation results. The design procedures are utilized to learn the training examples from the simulation results. The design procedures are two methods the first the neural network infer the deformed shape from the pre-defined processes parameters. The other the network infer the processes parameters from deformed shape. Especially the latest method is very useful to design the preform From the desired feature it is possible to determine the processes parameters such as friction stroke and tooling geometry. The proposed method is useful for shop floor to decide the processes parameters and preform shapes for producing sound product.

• IE interfaces
• /
• v.11 no.1
• /
• pp.145-153
• /
• 1998

In this paper, two parts of neural networks were proposed for determination of optimal EDM parameters. One is pattern recognition neural network that can be selecting expert neural network suitable to the EDM mode. The other is expert neural network that can be determining optimal EDM parameters such as pulse on time and pulse off time. Prior to determination of EDM parameters, Peak current, which is related to the EDM area closely, determined base on EDM area that is calculated from CAD data, firstly. Then, the other EDM parameters determined by the expert neural network that is selected to the EDM mode.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.18 no.8
• /
• pp.1939-1951
• /
• 1994

This study introduces a new tool breakage detecting technology comprised of an unsupervised neural network combined with adaptive time series autoregressive(AR) model where parameters are estimated recursively at each sampling instant using a parameter adaptation algorithm based on an RLS(Recursive Least Square). Experiment indicates that AR parameters are good features for tool breakage, therefore it can be detected by tracking the evolution of the AR parameters during milling process. an ART 2(Adaptive Resonance Theory 2) neural network is used for clustering of tool states using these parameters and the network is capable of self organizing without supervised learning. This system operates successfully under the wide range of cutting conditions without a priori knowledge of the process, with fast monitoring time.

• Journal of Korea Game Society
• /
• v.2 no.2
• /
• pp.46-51
• /
• 2002

In this study, a real-time simulation method for the phenomena, which are too complex to be simulated during real-time computer games, was proposed based on the neural network. The procedure of proposed method is to 1) obtain correlation data between input parameters and output parameters by mathematical modeling, code analyses, and so on, 2) train the neural network with the correlation data, 3) and insert the trained neural network in a game program as a simulation module. For the case that the number of the input and output parameters is too high to be analyzed, a method was proposed to omit parameters of little importance. The method was successfully applied to severe accidents of nuclear power plants, reflecting that the method was very effective in real time simulation of complex phenomena.

• 제어로봇시스템학회:학술대회논문집
• /
• 1989.10a
• /
• pp.513-519
• /
• 1989

Modeling of the network interface for MAP and its performance analysis is investigated in this study. The parameters for the network interface are selected and a special interest is concentrated on the parameters related to the performance of the network interface itself. A queueing model of the network interface is proposed and simulation is performed to validate the proposed model of the network interface.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2020.11a
• /
• pp.234-236
• /
• 2020

Content-adaptive training and transmission of the model parameters of neural networks can boost up the SR performance with higher restoration fidelity. In this case, efficient transmission of neural network parameters are essentially needed. Thus, we propose a novel method of compressing the network model parameters based on the training of network model parameters in the sense that the residues of filter parameters and content loss are jointly minimized. So, the residues of filter parameters are only transmitted to receiver sides for different temporal portions of video under consideration. This is advantage for image restoration applications with receivers (user terminals) of low complexity. In this case, the user terminals are assumed to have a limited computation and storage resource.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.28 no.7A
• /
• pp.557-562
• /
• 2003

An output multiplexer of manifold type is widely used in a recent satellite transponder for its mass and volume reduction. For correct operation, the filters of such a multiplexer must be singly terminated. In this paper, a simple synthesis method of a 6-pole dual-mode singly-terminated filter is described. From the transfer function of the filter, network parameters such as in/output terminations and coupling coefficients are obtained easily without complicated matrix algebra such as orthogonal projection and similarity transformation. Two different-structure filters are taken into consideration and the network parameters of each filter have been extracted from the same transfer function. It is shown that the responses of two filters are same to each other since their network parameters are obtained from the same transfer function. The method described in this paper can be applied to the other degree singly terminated filter.

• IE interfaces
• /
• v.15 no.3
• /
• pp.256-262
• /
• 2002

In this paper, we develop a framework for identifying the quality of service(QoS) and network performance (NP) in the IMT-2000 services, and analyze the QoS/NP system in the 3GPP IMT-2000 services. Based on the ITU-T E.800, the QoS is classified into customer, technical, contents, telecommunications quality, and internet communication quality, and the NP consists of the service access/transmission, reliability/operating & maintenance, charging performance, and mobility management performance. Under the basic framework, the major parameters in the IMT-2000 services are identified for each QoS/NP criterion. The QoS framework in the IMT-2000 user aspects is also introduced to determine the major QoS parameters. Finally, to define the performance factors of the network elements in the NP system, the various control parameters for the wireless and core networks are presented.

• Journal of the Korean Society of Manufacturing Technology Engineers
• /
• v.8 no.5
• /
• pp.70-75
• /
• 1999

Adaptive control in the robotic GMA(Gas Metal Arc) welding is employed to monitor the information about weld characteristics and process paramters as well as modification of those parameters to hold weld quality within the acceptable limits. Typical characteristics are the bead geometry composition micrrostructure appearance and process parameters which govern the quality of the final weld. The main objectives of this paper are to realize the mapping characteristicso f penetration through the learning. After learning the neural network can predict the pene-traition desired from the learning mapping characteristic. The design parameters of the neural network estimator(the number of hidden layers and the number of nodes in a layer) were chosen from an error analysis. partial-penetration single-pass bead-on-plate welds were fabricated in 12mm mild steel plates in order to verify the performance of the neural network estimator. The experimental results show that the proposed neural network estimator can predict the penetration with reasonable accuracy and gurarantee the uniform weld quality.