• Journal of the Korean Data and Information Science Society
• /
• v.12 no.2
• /
• pp.1-10
• /
• 2001

Neural networks we increasingly being seen as an addition to the statistics toolkit which should be considered alongside both classical and modern statistical methods. Neural networks are usually useful for classification and function estimation. In this paper we concentrate on function estimation using neural networks with weight decay factor The use of weight decay seems both to help the optimization process and to avoid overfitting. In this type of neural networks, the problem to decide the number of hidden nodes, weight decay parameter and iteration number of learning is very important. It is called the optimization of weight decay neural networks. In this paper we propose a automatic optimization based on genetic algorithms. Moreover, we compare the weight decay neural network automatically learned according to automatic optimization with ordinary neural network, projection pursuit regression and support vector machines.

• Journal of Korea Water Resources Association
• /
• v.56 no.11
• /
• pp.697-703
• /
• 2023

In this study, a modified indirect evaluation method was developed to predict the deterioration level of water pipes. The accuracy of the modified method was verified by comparing it with the direct method. The weights of index were adjusted by reducing the weight of water quality corrosion, soil corrosion, lay depth and road type according to the importance of the existing evaluation factors and adding the weight of pipe thickness. In the results, the weight of pipe thickness was determined to be 0.1530. Comparing with the direct evaluation method, the accuracy of the modified indirect evaluation method increased by 31.03% compared to the indirect evaluation method. The modified indirect evaluation method will be able to select relatively old pipes more accurately and efficiently than the existing indirect evaluation method when prioritizing the improvement of old water pipes.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1998.10a
• /
• pp.251-254
• /
• 1998

본 논문에서는 신경회로망의 수렴속도를 높이기 위한 알고리즘을 제안한다. 전형적인 역전파 학습방식은 느린 수렴속도가 단점으로 제기되는데 이는 비용함수의 계곡부근에서 가중치의 궤적이 심한 진동현상을 보이기 때문이다. 이 문제를 해결하기 위해서 본 논문에서는 경사법에서 사용되는 갱신방향을 계곡의 진행방향을 이용하여 변경한다. 모의실험을 통하여 제안된 방법으로 가중치의 궤적에 나타나는 진동을 줄이고 수렴속도를 향상시킬 수 있음을 보인다.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2010.10a
• /
• pp.209-212
• /
• 2010

본 논문에서는 영상을 확대할 경우에 발생하는 영상의 품질 저하를 최소화하기 위하여 원본 영상 픽셀과 확대된 결과 영상 픽셀 간의 명암도 차이와 보간 수행시 적용되는 가중치 값을 퍼지 기법에 적용하여 영상을 확대하는 방법을 제안한다. 제안된 방법은 기존의 양선형 보간법으로 도출된 결과 영상 픽셀과 원본 영상 픽셀 간의 명암도 차이와 보간 수행시 네 개의 픽셀 값에 곱하게 되는 가중치 값을 퍼지 소속 함수에 적용하여 원본 영상의 픽셀 정보와 가장 근접한 특징을 가진 확대된 결과 영상의 픽셀 정보를 최종적으로 도출한다. 제안된 방법을 실험한 결과, 기존의 양선형 보간법에 비해 영상 확대시, 발생하는 문제점인 흐림 현상이 상대적으로 감소하여 영상의 품질이 개선되는 것을 확인하였다.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.13 no.4
• /
• pp.67-77
• /
• 2010

Growth of oil and chemical industries has been remarkable during recent years. Hazardous materials (Hazmat) make frequent use in the wide range of industries. It increases the frequency of Hazmat transport and it leads to increase the number of accidents. Optimal Hazmat routes can reduce damage. Thus, the objective of this study is to minimize the areas impacted by Hazmat accidents by adopting experts' opinion in planning the route. We calculated weights using AHP (Analytic Hierarchy Process) and deduced the best route by applying this weights. Results showed that in the case of shortest route versus weighted route, the percentage of population damage has been decreased by 33.4% in the comparison between shortest route and optimally weighted route. And the percentage of environmental damage also has been decreased by 21.8%. Social damage has been decreased by 1521.7%. In the case of none weighted route versus weighted route, the percentage of population damage has been decreased by 2.6% when we adopted weighted route. Consequently, the recommended route with weighted risk assessment avoids densely populated area comparing with none weighted route. Further research needs to be carried out in order to figure out the specific cost-effectiveness analysis applying the equal cost unit for each factor.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.32 no.3C
• /
• pp.241-248
• /
• 2007

In predictive image coding, a LS (Least Squares)-based adaptive predictor is an efficient method to improve image edge predictions. This paper proposes a hybrid interpolation with weighted edge detector. A hybrid approach of switching between bilinear interpolation and EDI (Edge-Directed Interpolation) is proposed in order to reduce the overall computational complexity The objective and subjective quality is also similar to the bilinear interpolation and EDI. Experimental results demonstrate that this hybrid interpolation method that utilizes a weighted edge detector can achieve reduction in complexity with minimal degradation in the interpolation results.

• Survey Research
• /
• v.5 no.2
• /
• pp.31-48
• /
• 2004

We collected and analyzed 18 social surveys to assess the quality of samples with respect to region, gender, age-band, education level and occupation. We found in our samples that highly educated people and house wives are over-represented whereas low educated people, self-employed/blue collars and white collars are under-represented. To correct such sample distortions, we applied the iterative proportional weighting or the raking to our samples. We observed sizable changes in survey results. Also, the effective sample sizes were shrunken up to 20%-40%, that could be interpreted as the necessity of larger samples to meet the claimed sampling error limits.

• Journal of the Korea Society of Computer and Information
• /
• v.16 no.11
• /
• pp.153-161
• /
• 2011

As the problem of GOSST building belongs to NP compete domain, heuristics for the problem ask for immense amount execution time and computations in large scale inputs. In this paper, we propose an efficient mechanism for GOSST construction using space locality PTAS. For 40,000 input nodes with maximum weight 100, the proposed space locality PTAS GOSST with 16 unit areas can reduce about 4.00% of connection cost and 89.26% of execution time less than weighted minimum spanning tree method. Though the proposed method increases 0.03% of connection cost more, but cuts down 96.39% of execution time less than approximate GOSST method (SGOSST) without PTAS. Therefore the proposed space locality PTAS GOSST mechanism can work moderately well to many useful applications where a greate number of weighted inputs should be connected in short time with approximate minimum connection cost.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.22 no.5
• /
• pp.561-566
• /
• 2021

The optimum kernel size for error-distribution estimation with given error samples cannot be used in the weight adjustment of minimum Euclidean distance between error distributions (MED) algorithms. In this paper, a new adaptive kernel estimation method for convergence enhancement of MED algorithms is proposed. The proposed method uses the average rate of change in error power with respect to a small interval of the kernel width for weight adjustment of the MED learning algorithm. The proposed kernel adjustment method is applied to experiments in communication channel compensation, and performance improvement is demonstrated. Unlike the conventional method yielding a very small kernel calculated through optimum estimation of error distribution, the proposed method converges to an appropriate kernel size for weight adjustment of the MED algorithm. The experimental results confirm that the proposed kernel estimation method for MED can be considered a method that can solve the sensitivity problem from choosing an appropriate kernel size for the MED algorithm.

• Journal of the Korea Society of Computer and Information
• /
• v.16 no.4
• /
• pp.23-28
• /
• 2011

Ultrasound images for the abdominal muscles are complicated enough to have difficulty in interpreting their results. For better interpretation, magnifying the original image is necessary but its magnified image could be deteriorated and suffer from information loss. Thus, in this paper, we propose a magnifying method that reduces the gap between the original image and the magnified one in quality using a fuzzy method with weights for its brightness and interpolation. The proposed method extracts information of pixels in magnified image that have most similar characteristics of the original one by applying fuzzy membership function. In the process, the difference in the brightness between pixels of the magnified image and the original one using bilinear interpolation method and the weight value using the interpolation from multiplied values of four pixels are supplied to the fuzzy membership function. In this experiment, the proposed method reduces the cloudy phenomenon appears commonly compared to the bilinear interpolation method among those qualitative issues of image interpretation.