• Korean Journal of Remote Sensing
• /
• v.18 no.3
• /
• pp.155-162
• /
• 2002

Neural networks have been focused on as a robust classifier for the remotely sensed imagery due to its statistical independency and teaming ability. Also the artificial neural networks have been reported to be more tolerant to noise and missing data. However, unlike the conventional statistical classifiers which use the statistical parameters for the classification, a neural network classifier uses individual training sample in teaming stage. The training performance of a neural network is know to be very sensitive to the discrepancy of the number of the training samples of each class. In this paper, the effect of the population discrepancy of training samples of each class was analyzed with three layered feed forward network. And a method for reducing the effect was proposed and experimented with Landsat TM image. The results showed that the effect of the training sample size discrepancy should be carefully considered for faster and more accurate training of the network. Also, it was found that the proposed method which makes teaming rate as a function of the number of training samples in each class resulted in faster and more accurate training of the network.

• Geomechanics and Engineering
• /
• v.37 no.4
• /
• pp.359-370
• /
• 2024

The primary goal of this study is to evaluate the migration of fine granular materials into overlying layers under cyclic loading using a modified large-scale triaxial system as a physical model test. Samples prepared for the modified large-scale triaxial system comprised a 60 mm thick gravel layer overlying a 120 mm thick subgrade layer, which could be either tailings or railway sand. A quantitative analysis of the migration of fine granular materials was based on the mass percentage and grain size of migrated materials collected in the gravel. In addition, the cyclic characteristics, i.e., accumulated axial strain and excess pore water pressure, were evaluated. As a result, the total migration rate of the railway sand sample was found to be small. However, the total migration rate of the sample containing tailings in the subgrade layer was much higher than that of the railway sand sample. In addition, the migration analysis revealed that finer tailings particles tended to be migrated into the upper gravel layer easier than coarser tailings particles under cyclic loading. This could be involved in significant increases in excess pore water pressure at the last cycles of the physical model test.

• Journal of Nuclear Fuel Cycle and Waste Technology(JNFCWT)
• /
• v.22 no.2
• /
• pp.145-157
• /
• 2024

Undeclared nuclear activities are challenging given the lack of information from the sites involved in such activities. Wide-area environmental sampling (WAES) can be an effective method to detect undeclared nuclear activities. However, it is crucial to address the potential risks during the WAES, including sample tampering or extortions. Therefore, tracking and monitoring of various on-site data is imperative to accurately interpret the status of samples and workers throughout the WAES process. 'Environmental and Geographical Data Transfer (EGDT)' was developed for the real-time monitoring of integrated on-site data. EGDT module is equipped with various sensors and can be attached to a worker's uniform or a sample storage box. This study demonstrated the technical effectiveness of EGDT by exploring three experimental methodologies for feasibility assessment. Compared to the Normal Operation case, the inference of the Sample Extortion case was predominantly based on changes in lux and dose rate. The inference of the Out-of-Work-Area case primarily relied on changes in dose rate and acceleration. Finally, the preliminary evaluation of the performance of the developed prototype was conducted, and a foundation was established for enhancing the application in the WAES process.

• International Journal of Reliability and Applications
• /
• v.1 no.2
• /
• pp.133-143
• /
• 2000

Data from weighted distributions appear, among other situations, when some of the data are missing or are damaged, a case that is important in reliability and life testing. The kernel method for hazard rate estimation is discussed for these data where the basic large sample properties are given. As a by product, the basic properties of the kernel estimate of the distribution function for data from weighted distribution are presented.

• Proceedings of the Korean Society for Quality Management Conference
• /
• 1998.11a
• /
• pp.571-579
• /
• 1998

In this paper we propose smooth nonparametric estimator of Mean Residual Life(MRL) based on a complete sample. This estimator is constructed using estimator of cumulative failure rate which is derived as the maximum likelihood estimate of cumulative failure rate in the class of distributions which have piecewise linear failure rate functions between each pair of observations. We derive the asymptotic properties of the our estimator. The proposed estimator is compared with previously known estimator by Monte Carlo study.

• Journal of Korean Society for Quality Management
• /
• v.25 no.3
• /
• pp.51-61
• /
• 1997

We introduce a method of estimating an unknown failure rate function based on sample data. We estimate failure rate function by a function s from a space of cubic splines constrained to be linear (or constant) in tails using maximum likelihood estimation. The number of knots are determined by Bayesian Information Criterion(BIC). Examples using simulated data are used to illustrate the performance of this method.

• Proceedings of the KSRS Conference
• /
• 2005.10a
• /
• pp.298-300
• /
• 2005

Satellite payload on-board date compression unit are use for saving date storage space and reducing time to transmit payload data to the ground station. The KOMPSAT-3 payload will generate higher data rate than KOMPSAT-2 due to its better ground sample distance capacity. High input data rate and limited output transmission data rate might lead excessive compression and degraded image quality. This paper presents a trade-off study about data storage capacity and compression parameters for estimated KOMPSAT-3 system.

• Communications for Statistical Applications and Methods
• /
• v.17 no.1
• /
• pp.99-106
• /
• 2010

In this paper, we present two methods for obtaining prediction intervals for the times to failure of units censored in multiple stages in a progressively censored sample from proportional hazard rate models. A numerical example and a Monte Carlo simulation study are presented to illustrate the prediction methods.

• Journal of Korean Society for Quality Management
• /
• v.29 no.1
• /
• pp.41-46
• /
• 2001

The problem of trend change in the failure rate is great interest in the reliability and survival analysis. In this paper we develop a test statistic for testing whether or not the failure rate changes its trend based on a complete sample. Monte Carlo simulations are conducted to investigate the speed of convergence of the proposed test statistic.

• Journal of the Korean Society of Tobacco Science
• /
• v.3 no.2
• /
• pp.89-93
• /
• 1981

Some or the parameters affecting the thermal analysis of different types of ground tobacco have been discussed. The sample, Korean flue- cured and burley and Greek orient tobacco leaves were run in air at heating rate of $10^{\circ}$ and 20^{\circ}C$ per minute. The thermograms obtained were divided into four temperature zones. The effect of different heating rates on thermal degradation has been discussed At the lower heating rate the maxima of DTA curves in zone 3 and 4 were 25-$40^{\circ}C$ lower than that of the higher heating rate.