• Journal of the Korean Data and Information Science Society
• /
• v.8 no.2
• /
• pp.163-171
• /
• 1997

Function approximation from a set of input-output pairs has numerous applications in scientific and engineering areas. Multilayer feedforward neural networks have been proposed as a good approximator of nonlinear function. The back propagation(BP) algorithm allows multilayer feedforward neural networks to learn input-output mappings from training samples. It iteratively adjusts the network parameters(weights) to minimize the sum of squared approximation errors using a gradient descent technique. However, the mapping acquired through the BP algorithm may be corrupt when errorneous training data we employed. When errorneous traning data are employed, the learned mapping can oscillate badly between data points. In this paper we propose a robust BP learning algorithm that is resistant to the errorneous data and is capable of rejecting gross errors during the approximation process, that is stable under small noise perturbation and robust against gross errors.

• Proceedings of the Korean Society for Quality Management Conference
• /
• 1998.11a
• /
• pp.159-173
• /
• 1998

Quality characteristics on the properties of process capability indices (PCIs) are often required to be normally distributed. But, if a characteristic is not normally distributed, serious errors can result from normal-based techniques. In this case, we may well consider the use of new PCIs specially designed to be robust for non-normality. In this paper, a newly proposed measure of process capability is introduced and compared with existing PCIs using the simulated non-normal data.

• Communications for Statistical Applications and Methods
• /
• v.21 no.1
• /
• pp.105-114
• /
• 2014

In this paper we suggest an efficient method to estimate the distribution function using the Bezier curve, and compare it with existing methods by simulation studies. In addition, we suggest a robust version of cross-validation criterion to estimate the number of Bezier points, and showed that the proposed method is better than the existing methods based on simulation studies.

• Proceedings of the IEEK Conference
• /
• 2003.07d
• /
• pp.1279-1282
• /
• 2003

Estimation of human motion states is important enabling technologies for realizing a pervasive computing environment. In this paper, an improved method fur estimating human motion state from accelerometer data is introduced. Our method fur estimating human motion state utilizes various statistics of accelerometer data, such as mean, standard variation, skewness, kurtosis, eccentricity, as features for classification, and therefore is expected to be more robust than other existing methods that rely on only a few simple statistics. A series of experiments fur testing the effectiveness of the proposed method has been performed, and its result is presented.

• Communications for Statistical Applications and Methods
• /
• v.14 no.2
• /
• pp.401-411
• /
• 2007

In this paper we propose an influence measure for detecting potentially influential observations using the infinitesimal perturbation and the local influence in the log-linear regression model. Also, we propose a goodness-of-fit measure for variable selection. A real data set are used for illustration.

• Journal of the Korean Data and Information Science Society
• /
• v.6 no.2
• /
• pp.41-50
• /
• 1995

In this paper, an iterative procedure, which detects the location of the outliers and the joint estimates of the outliers effects and the model parameters in the autoregressive moving average model with two types of outliers, is proposed. The performance of the procedure is compared with the one in Chen and Liu(1993) through the Monte Carlo simulation. The proposed procedure is very robust in the sense that applies the procedures to the stationary time series model with any types of outliers.

• Journal of Korean Society for Quality Management
• /
• v.22 no.2
• /
• pp.166-176
• /
• 1994

Some rank score tests are proposed for testing the equality of all sampling distribution functions against ordered location-scale alternatives in k-sample problem. Under the null hypothesis and a contiguous sequence of ordered location-scale alternatives, the asymptotic properties of the proposed test statistics are investigated. Also, the asymptotic local powers are compared with each others. The results show that the proposed tests based on the Hettmansperger-Norton type statistic are more powerful than others for the general ordered location-scale alternatives. However, the Shiraishi's tests based on the sum of two Bartholomew's rank analogue statistics are robust.

• Journal of the Korean Statistical Society
• /
• v.32 no.4
• /
• pp.385-399
• /
• 2003

We first introduce the quasi-score estimating function and applied the quasi-score estimating function to nonlinear time series models. We proposed the M quasi-score estimating functions bounded functions for the quasi-score estimating functions. Also, we investigated the asymptotic properties of quasi-likelihood estimators and M quasi-likelihood estimators. Simulation results show that the M quasi-likelihood estimators work better than the least squares estimators under the heavy-tailed distributions

• Journal of the Korean Data and Information Science Society
• /
• v.11 no.2
• /
• pp.313-318
• /
• 2000

This article addresses aspects of combining information, with special attention to meta-analysis. In specific, we consider hierarchical Bayesian models for meta-analysis under priors which are scale mixtures of normal, and thus have tail heavier than that of the normal. Numerical methods of finding Bayes estimators under these heavy tailed prior are given, and are illustrated with an actual example.

• Communications for Statistical Applications and Methods
• /
• v.28 no.5
• /
• pp.553-562
• /
• 2021

Directional regression (DR; Li and Wang, 2007) is well-known as an exhaustive sufficient dimension reduction method, and performs well in complex regression models to have linear and nonlinear trends. However, the extension of DR is not well-done upto date, so we will extend DR to accommodate multivariate regression and large p-small n regression. We propose three versions of DR for multivariate regression and discuss how DR is applicable for the latter regression case. Numerical studies confirm that DR is robust to the number of clusters and the choice of hierarchical-clustering or pooled DR.