• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.57 no.11
• /
• pp.2095-2101
• /
• 2008

In this paper, we consider a performance improvement of neural network for predicting defect size of steam generator tube using early stopping. Usually, neural network is trained until MSE becomes less than a prescribed error goal. The smaller the error goal, the greater the prediction performance for the trained data. However, as the error goal is decreased, an over fitting is likely to start during supervised training of a neural network, which usually deteriorates the generalization performance. We propose that, for the prediction of an axisymmetric defect size, early stopping can be used to avoid the over-fitting. Through various experiments on the axisymmetric defect samples, we found that the difference bet ween the prediction error of neural network based on early stopping and that of ideal neural network is reasonably small. This indicates that the error goal used for neural network training for the prediction of defect size can be efficiently selected by early stopping.

• Computers and Concrete
• /
• v.27 no.1
• /
• pp.63-71
• /
• 2021

In this study, a new stopping criterion, called "backward controlled stopping criterion" (BCSC), was proposed to be used in Genetic Algorithms. In the study, the available stopping citeria; adaptive stopping citerion, evolution time, fitness threshold, fitness convergence, population convergence, gene convergence, and developed stopping criterion were applied to the following four comparison problems; high strength concrete mix design, pre-stressed precast concrete beam, travelling salesman and reinforced concrete deep beam problems. When completed the analysis, the developed stopping criterion was found to be more accomplished than available criteria, and was able to research a much larger area in the space design supplying higher fitness values.

• Journal of Satellite, Information and Communications
• /
• v.6 no.2
• /
• pp.78-83
• /
• 2011

In this paper, we propose a computationally efficient early stopping method to reduce the average number of iterations. The conventional early stopping methods have too much computational complexity to compute the stopping criterion. Thus, only the hard decision based early stopping method is suitable to realize the hardware of LDPC decoder. However, this method also can increase the computational complexity of LDPC decoder. The proposed method can effectively reduce the computational complexity of stopping criterion as we do not compute hard decision, and we combine the stopping criterion with horizontal shuffling scheduling decoding scheme. The simulation results show that a new early stopping method achieves acceptable bit error rate performance also reduces the average number of iterations.

• Journal of The Institute of Information and Telecommunication Facilities Engineering
• /
• v.8 no.4
• /
• pp.207-215
• /
• 2009

In recent digital communication systems, the performance of Turbo Code used as the error correction coding method depends on the interleaver size influencing the free distance determination and iterative decoding algorithms of the turbo decoder. However, some iterations are needed to get a better performance, but these processes require large time delay. Recently, methods of reducing the number of iteration have been studied without degrading original performance. In this paper, the new method combining ME (Mean Estimate) stopping criterion with SDR (sign difference ratio) stopping criterion of previous stopping criteria is proposed, and the fact of compensating each method's missed detection is verified Faster decoding realizes that reducing the number of iterative decoding about 1~2 times by adopting our proposed method into serially concatenation of both decoder. System Environments were assumed DS-CDMA forward link system with intense MAI (multiple access interference).

• Proceedings of the KIPE Conference
• /
• 2011.07a
• /
• pp.296-297
• /
• 2011

Many applications involving induction motors that are controlled using variable frequency drives require the ability to stop quickly. These applications include emergency stops, quick stopping of fans, centrifuges, presses, etc. The technique that is widely accepted in the industry for achieving quick stopping makes use of brake resistors in series with a power semiconductor switch. The switch-resistor combination (brake-unit) is applied across the dc bus. The fastest decelerating time achievable depends on the size of the resistors and the switch employed. In this paper, the authors propose a novel method of achieving quick stopping times without the use of any brake-unit. Experimental test results with and without this method on a large inertia motor-load combination show that the proposed stopping method is able to reduce the stopping time significantly compared to normal decelerated stop without the need for a braking unit.

• Journal of information and communication convergence engineering
• /
• v.21 no.3
• /
• pp.198-207
• /
• 2023

Responding to changes in artificial intelligence models and the data environment is crucial for increasing data-learning accuracy and inference stability of industrial applications. A learning model that is overfitted to specific training data leads to poor learning performance and a deterioration in flexibility. Therefore, an early stopping technique is used to stop learning at an appropriate time. However, this technique does not consider the homogeneity and independence of the data collected by heterogeneous nodes in a differential network environment, thus resulting in low learning accuracy and degradation of system performance. In this study, the generalization performance of neural networks is maximized, whereas the effect of the homogeneity of datasets is minimized by achieving an accuracy of 99.7%. This corresponds to a decrease in delay time by a factor of 2.33 and improvement in performance by a factor of 2.5 compared with the conventional method.

• Journal of the Korean Statistical Society
• /
• v.29 no.3
• /
• pp.337-350
• /
• 2000

This paper suggests a new diagnostic measure and a stopping rule for detecting influential observations in multiple discriminant analysis (MDA). It is developed from a Bayesian point of view using a default Bayes factor obtained from the fractional Bayes factor methodology. The Bayes factor is taken as a discriminatory information in MDA. It is shown that the effect of an observation over the discriminatory information is fully explained by the diagnostic measure. Based on the measure, we suggest a stopping rule for detecting influential observations in a given training sample. As a tool for interpreting the measure a graphical method is sued. Performance of the method is used. Performance of the method is examined through two illustrative examples.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.27 no.1
• /
• pp.38-46
• /
• 2021

In the era of the 4th Industrial Revolution, the interest in autonomous ship technology is increasing as high-tech technologies are being increasingly utilized throughout the industry. Therefore, we conducted a basic study on autonomous ships. In particular, a passenger ship model was produced and an autonomous navigation system was established by applying the ardupilot used for drones. The possibility of automatic control of the autonomous ship operations was confirmed by executing various voyage plans using the built model ship. In the performance test for maintaining the course the model ship could not follow the designated course straight and sailed up to 5.4 m away from the course while navigating in a zigzag (S-shape); however, after the parameters were modified, the deviation distance was reduced to a maximum of 1.8 m. In the turning performance test, the maximum diameter of the turning sphere was found to be approximately 9.3 m, but no significant change could be confirmed even after the parameters were modified. However, the results of our tests on slowing down the ship before arriving at the WP confirmed that the diameter of the turning sphere was reduced to a maximum of approximately 3.2 m. In order to evaluate the stopping performance, the last scheduled stopping position of all experiments was compared with the actual stopping position of the model ship and it was confirmed that the model ship stopped at a point at least 0.4 m and a maximum of 6.2 m away from the stopping position. In the future, improvement of course stability, turning performance, and stopping performance is required through modification and supplementation of various parameters. Moreover, a study on automatic berthing of the model ship through automatic control is planned.

• Journal of the Korean Data and Information Science Society
• /
• v.8 no.1
• /
• pp.9-20
• /
• 1997

Several methods are proposed for optimizing Fourier series estimators with respect to Mean Integrated Square Error metrics. Traditionally, such method have followed. one of two basic strategies; A stopping rules or the rules of determine multipliers. A central hypothesis of this study is that better estimates can be obtained by combining the two strategies. A new multiplier sequence is proposed, which used in conjunction with any of the stopping rules, is shown to improve the performance of estimator which relies solely on a stopping rule.

• Proceedings of the IEEK Conference
• /
• 2003.11c
• /
• pp.265-268
• /
• 2003

In this paper, we propose new simple and efficient stopping criteria for W-CDMA turbo decoding. The new criteria are based on the sum of logarithm of absolute extrinsic information values and the count of one-estimations calculated after each component decoder. The calculation of the proposed criteria can be made simpler by summing one of ten values without any penalty. This Proposed method can be implemented with negligible overhead and no extra memory. Simulation results on W-CDMA standard shows that the number of iterations required is reduced compared to the existing stopping criteria without degrading the BER performance. Reduced iterations and simple operations make the proposed criteria desirable for low-power turbo decoder.