• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.32B no.6
• /
• pp.922-931
• /
• 1995

In this paper, we propose a modified error function to improve the EBP(Error Back-Propagation) algorithm of Multi-Layer Perceptrons. Using the modified error function, the output node of MLP generates a strong error signal in the case that the output node is far from the desired value, and generates a weak error signal in the opposite case. This accelerates the learning speed of EBP algorothm in the initial stage and prevents overspecialization for training patterns in the final stage. The effectiveness of our modification is verified through the simulation of handwritten digit recognition.

• Proceedings of the KIEE Conference
• /
• 1988.07a
• /
• pp.87-90
• /
• 1988

This paper reports a simple posteriori error estimate method for adaptive finite element mesh generation using quadratic shape function especially for the magnetic field problems. The elements of quadratic shape function have more precise solution than those of linear shape function. Therefore, the difference of two solutions gives error quantity. The method uses the magnetic flux density error as a basis for refinement. This estimator is tested on two dimensional problem which has singular points. The estimated error is always under estimated but in same order as exact error, and this method is much simpler and more convenient than other methods. The result shows that the adaptive mesh gives even better rate of convergence in global error than the uniform mesh.

• Structural Engineering and Mechanics
• /
• v.10 no.2
• /
• pp.157-164
• /
• 2000

This paper discusses the error propagation characteristics of the Newmark explicit method, modified Newmark explicit method and ${\alpha}$-function dissipative explicit method in pseudodynamic tests. The Newmark explicit method is non-dissipative while the ${\alpha}$-function dissipative explicit method and the modified Newmark explicit method are dissipative and can eliminate the spurious participation of high frequency responses. In addition, error propagation analysis shows that the modified Newmark explicit method and the ${\alpha}$-function dissipative explicit method possess much better error propagation properties when compared to the Newmark explicit method. The major disadvantages of the modified Newmark explicit method are the positive lower stability limit and undesired numerical dissipation. Thus, the ${\alpha}$-function dissipative explicit method might be the most appropriate explicit pseudodynamic algorithm.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2003.05a
• /
• pp.408-413
• /
• 2003

The frequency response function(FRF) of each substructure is used for the transfer function synthesis method(TFS). The dynamic characteristics of the full system are obtained by synthesizing FRFs of each substructure. The validation of TFS depends on accuracy for FRF of each substructure. Impact hammer testing Is widely used to obtain the modal characteristics of structures However. the FRF obtained from impact hammer testing contains bias errors, such as finite record length error and leakage error of which characteristic depends on data acquisition time which we call record length. In this paper, a method to remove hose errors is proposed so as to enhance results of TFS. Numerical and experimental examples show that the FRF of full structure can be predicted nearly exactly by the method proposed in this paper.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.10 no.1
• /
• pp.59-64
• /
• 2010

To improve system performance, we apply the concept of membership function to Variance Considered Machines (VCMs) which is a modified algorithm of Support Vector Machines (SVMs) proposed in our previous studies. Many classification algorithms separate nonlinear data well. However, existing algorithms have ignored the fact that probabilities of error are very high in the data-mixed area. Therefore, we make our algorithm ignore data which has high error probabilities and consider data importantly which has low error probabilities to generate system output according to the probabilities of error. To get membership function, we calculate sigmoid function from the dataset by considering means and variances. After computation, this membership function is applied to the VCMs.

• Journal of the Korean Data and Information Science Society
• /
• v.15 no.4
• /
• pp.899-910
• /
• 2004

In this paper we have proposed a new loss function, namely, non-linear exponential(NLINEX) loss function, which is quite asymmetric in nature. We obtained the Bayes estimator under exponential(LINEX) and squared error(SE) loss functions. Moreover, a numerical comparison among the Bayes estimators of power function distribution under SE, LINEX, and NLINEX loss function have been made.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.13 no.8
• /
• pp.797-803
• /
• 2002

In this paper we derive the probability density function of the phase error of the received signal over Rician fading channel and verify its propriety as the probability density function using the zeroth moment. In general, for the error probability over fading channel we compute the error probability in the first place when it is only AWGN, and then we get the final result by averaging the first result and the probability density function of the corresponding fading channel. In this paper, however, we compute the error probability by double integration after the probability density function over fading channel is computed.

• International Journal of Contents
• /
• v.8 no.2
• /
• pp.7-12
• /
• 2012

Imbalanced data sets are difficult to be classified since most classifiers are developed based on the assumption that class distributions are well-balanced. In order to improve the error back-propagation algorithm for the classification of imbalanced data sets, a new error function is proposed. The error function controls weight-updating with regards to the classes in which the training samples are. This has the effect that samples in the minority class have a greater chance to be classified but samples in the majority class have a less chance to be classified. The proposed method is compared with the two-phase, threshold-moving, and target node methods through simulations in a mammography data set and the proposed method attains the best results.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 1999.06a
• /
• pp.57.2-62
• /
• 1999

Fractal coding is image compression techniques using one of image characteristics self-transformability. In fractal image coding, the encoding process is to select the domain block similar to a range block. The reconstructed image quality of fractal image coding depends on similitude between a range block and the selected domain block. Domain block similar to a range blocks. In fact, the error of the reconstructed image adds up the generated error in encoding process and the generated error in decoding process. But current domain block estimating function considered only the encoding error. We propose a domain block estimating function to consider not only the encoding error but also the decoding error. By computer simulation, it was verified to obtain the high quality reconstructed image.

• Journal of the Korean Mathematical Society
• /
• v.44 no.4
• /
• pp.835-844
• /
• 2007

This paper considers the problem of estimating the error distribution function in nonparametric regression models. Sufficient conditions are given under which the kernel estimator of the error distribution function based on nonparametric residuals satisfies the law of iterated logarithm.