• Journal of applied mathematics & informatics
• /
• v.29 no.1_2
• /
• pp.279-292
• /
• 2011

In this paper, we consider anti-periodic solutions of the following BAM neural networks with multiple delays on time scales: $$\{{x^\Delta_i(t)=-a_i(t)e_i(x_i(t))+{\sum\limits^m_{j=1}}c_{ji}(t)f_j(y_j(t-{\tau}_{ji}))+I_i(t),\atop y^\Delta_j(t)=-b_j(t)h_j(y_j(t))+{\sum\limits^n_{i=1}}d_{ij}(t)g_i(x_i(t-{\delta}_{ij}))+J_j(t),}\$$ where i = 1, 2, ..., n,j = 1, 2, ..., m. Using some analysis skills and Lyapunov method, some sufficient conditions on the existence and exponential stability of the anti-periodic solution to the above system are established.

• Journal of Institute of Control, Robotics and Systems
• /
• v.7 no.7
• /
• pp.582-588
• /
• 2001

The Multiple Model/Controller IMC(MMC-IMC) is a model-based control method which uses a set of model/controller pairs rather than a single model/controller to handle all possible operating conditions in the IMC control structure. During operation, one model/controller pair that best fit, for current plant situation is chosen by the switching algorithm. The major drawback of the switching controller is the bad transient performance due to the model error and the use fo linear controller for nonlinear plants. In this paper, we propose a method that transient response of the MMC-IMC using two recurrent neural networks. Simulation result shows that the proposed method represents better performance than the usual MMC-IMC`s.

• Journal of Intelligence and Information Systems
• /
• v.1 no.2
• /
• pp.103-121
• /
• 1995

This study proposes a artificial neural network method to predict the time to buy and sell the stocks listed on the Korea Composite Stock Price Index(KOSPI). Four types (NN1, NN2, NN3, NN4) of independent networks were developed to predict KOSPIs up/down direction after four weeks. These networks have a difference only in the length of learning period. NN5 - arithmetic average of four networks outputs - shows an higher accuracy than other network types and Multiple Linear Regression (MLR), and buying and selling simulation using systems outputs produces higher reture than buy-and-hold strategy.

• 제어로봇시스템학회:학술대회논문집
• /
• 1993.10b
• /
• pp.109-114
• /
• 1993

It is well known that neural networks can be used to diagnose multiple faults to some limited extent. In this work we present a Multiple Fault Diagnosis Method (MFDM) via neural network which can effectively diagnose multiple faults. To diagnose multiple fault, the proposed method finds the maximum value in the output nodes of the neural network and decreases the node value by changing the hidden node values. This method can find the other faults by computing again with the changed hidden node values. The effectiveness of this method is explored through a neural-network-based fault diagnosis case study of a fluidized catalytic cracking unit (FCCU).

• Asia pacific journal of information systems
• /
• v.9 no.3
• /
• pp.93-110
• /
• 1999

The previous studies regarding the stock returns have advocated that industry effects exist over entire industry. As the industry categories are more rigid, the demand for predicting the industry sectors is rapidly increasing. The advances in Artificial Intelligence and Neural Networks suggest the feasibility of a valuable computational model for stock returns prediction. We propose a sector-factor model for predicting the return on industry stock index using neural networks. As a substitute for the traditional models, neural network model may be more accurate and effective alternative when the dynamics between the underlying industry features are not well known or when the industry specific asset pricing equation cannot be solved analytically. To assess the potential value of neural network model, we simulate the resulting network and show that the proposed model can be used successfully for banks and general construction industry. For comparison, we estimate models using traditional statistical method of multiple regression. To illustrate the practical relevance of neural network model, we apply it to the predictions of two industry stock indexes from 1980 to 1995.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.58 no.3
• /
• pp.621-626
• /
• 2009

This paper presents a multiple branch predictor using perceptrons. The key idea is to apply neural networks to the multiple branch predictor. We describe our design and evaluate it with the SPEC 2000 integer benchmarks. Our predictor achieves increased accuracy than the Bi-Mode and the YAGS multiple branch predictor with the same hardware cost.

• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 2003.07a
• /
• pp.518-521
• /
• 2003

A new measurement system was developed by fault diagnosis system for traction motor using current signal analysis. The motor current signature analysis method is used for traction motor fault diagnosis. The diagnosis system program is constructed by artificial neural networks algorithm, those results from the program are used to train neural networks. The trained neural networks have the ability to compute adaptive results for non-trained inputs, and to calculate very fast due to original parallel structure of neural networks with high accuracy within destined tolerance. This system suggested that available test for checking, the probable extent of aging, and the rate of which aging is taking place.

• The Transactions of the Korea Information Processing Society
• /
• v.7 no.6
• /
• pp.1834-1843
• /
• 2000

In this paper, we propose a modular neural network based classification method for handwritten numerals utilizing rotated and translated images of an input image. The whole numeral pattern space is divided into smaller spaces which overlap each other and form multiple clusters. On these multiple clusters, multiple multilayer perceptrons (MLP) neural networks, specialized in those clusters, are constructed. Thus, each MLP acts as an expert network on the corresponding cluster. An MLP is also used as a gating network functioning as a mediator among the multiple MLPs. In the learning phase, an input numeral image is dithered by tow geometric operations of translation and rotation so that new numeral images similar to original one are generated. In the recognition phase, we utilize not only input numeral image, but also nearly generated images through the rotation and the translation of the original image. Thus, multiple output values for those generated images were combined to make class decision by various combination methods. The experimental results confirm the validity of the proposed method.

• Journal of Information Processing Systems
• /
• v.20 no.4
• /
• pp.501-513
• /
• 2024

In recent years, graph neural networks (GNNs) have been extensively used to analyze graph data across various domains because of their powerful capabilities in learning complex graph-structured data. However, recent research has focused on improving the performance of a single GNN with only two or three layers. This is because stacking layers deeply causes the over-smoothing problem of GNNs, which degrades the performance of GNNs significantly. On the other hand, ensemble methods combine individual weak models to obtain better generalization performance. Among them, gradient boosting is a powerful supervised learning algorithm that adds new weak models in the direction of reducing the errors of the previously created weak models. After repeating this process, gradient boosting combines the weak models to produce a strong model with better performance. Until now, most studies on GNNs have focused on improving the performance of a single GNN. In contrast, improving the performance of GNNs using multiple GNNs has not been studied much yet. In this paper, we propose gradient boosted graph neural networks (GBGNN) that combine multiple shallow GNNs with gradient boosting. We use shallow GNNs as weak models and create new weak models using the proposed gradient boosting-based loss function. Our empirical evaluations on three real-world datasets demonstrate that GBGNN performs much better than a single GNN. Specifically, in our experiments using graph convolutional network (GCN) and graph attention network (GAT) as weak models on the Cora dataset, GBGNN achieves performance improvements of 12.3%p and 6.1%p in node classification accuracy compared to a single GCN and a single GAT, respectively.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2013.10a
• /
• pp.320-324
• /
• 2013

A crack identification method using an equivalent bending stiffness for cracked beam and committee of neural networks is presented. The equivalent bending stiffness is constructed based on an energy method for a straight thin-walled pipe, which has a through-the-thickness crack, subjected to bending. Several numerical analysis for a steel cantilever pipe using the equivalent bending stiffness are carried out to extract the natural frequencies and mode shapes of the cracked beam. The extracted modal properties are used in constructing a training patterns of a neural network. The input to the neural network consists of the modal properties and the output is composed of the crack location and size. Multiple neural networks are constructed and each individual network is trained independently with different initial synaptic weights. Then, the estimated crack locations and sizes from different neural networks are averaged. Experimental crack detection is carried out for 3 damage cases using the proposed method, and the identified crack locations and sizes agree reasonably well with the exact values.