• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.58 no.11
• /
• pp.2136-2141
• /
• 2009

Transmission planning is an important part of power system planning to meet an increasing demand for electricity. The objective of transmission expansion is to minimize operational and construction costs subject to system constraints. There is inherent uncertainty in transmission planning due to errors in forecasted demand and fuel costs. Therefore, transmission planning process is not reliable if the uncertainty is not taken into account. The paper presents a systematic method to find the optimal location and amount of transmission expansion using Cross-Entropy (CE) incorporating uncertainties about future power system conditions. Numerical results are presented to demonstrate the performance of the proposed method.

• Journal of Advanced Navigation Technology
• /
• v.20 no.4
• /
• pp.380-385
• /
• 2016

Shadows are common phenomena observed in natural scenes and often bring a major problem that is affected negatively in colour image analysis. It is important to detect the shadow areas and should be considered in the pre-processing of computer vision. In this paper, the method of shadow detection is proposed using cross entropy and intensity image, and is performed in single image based on the satellite images. After converting the color image to a gray level image, the shadow candidate region has been estimated the optimal threshold value by cross entropy, and then the final shadow region has been detected using intensity image. For the validity of the proposed method, the satellite images is used to experiment. Some experiments are conducted so as to verify the proposed method, and as a result, shadow detection is well performed.

• Earthquakes and Structures
• /
• v.12 no.6
• /
• pp.673-687
• /
• 2017

A three-dimensional; structural health monitoring; vertical; planar; cross-sample entropy; multiscaleA three-dimensional structural health monitoring (SHM) system based on multiscale entropy (MSE) and multiscale cross-sample entropy (MSCE) is proposed in this paper. The damage condition of a structure is rapidly screened through MSE analysis by measuring the ambient vibration signal on the roof of the structure. Subsequently, the vertical damage location is evaluated by analyzing individual signals on different floors through vertical MSCE analysis. The results are quantified using the vertical damage index (DI). Planar MSCE analysis is applied to detect the damage orientation of damaged floors by analyzing the biaxial signals in four directions on each damaged floor. The results are physically quantified using the planar DI. With progressive vertical and planar analysis methods, the damaged floors and damage locations can be accurately and efficiently diagnosed. To demonstrate the performance of the proposed system, performance evaluation was conducted on a three-dimensional seven-story steel structure. According to the results, the damage condition and elevation were reliably detected. Moreover, the damage location was efficiently quantified by the DI. Average accuracy rates of 93% (vertical) and 91% (planar) were achieved through the proposed DI method. A reference measurement of the current stage can initially launch the SHM system; therefore, structural damage can be reliably detected after major earthquakes.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.12 no.1
• /
• pp.88-97
• /
• 2013

Most of the traditional spectrum sensing methods consider only the current detected data sets of Primary User (PU). However previous state of PU is a kind of conditional probability that strengthens the reliability of the detector. Therefore, in the cross entropy spectrum sensing method, relationship of the previous and current spectrum sensing is considered to detect PU signal more effectively. But these cross entropy spectrum sensing methods only consider the ideal system. In other words, PU always occupy the channel during the same period. However, PU can occupy the channel either for a longer or a shorter period than the ideal case in the real system. For this reason, the spectrum sensing performance can be varied. In this paper, we propose the method that can maintain the performance of spectrum sensing in the real system and we confirm the results with the help of simulation.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.1
• /
• pp.49-54
• /
• 2024

Data imbalance is a common issue encountered in classification problems, stemming from a significant disparity in the number of samples between classes within the dataset. Such data imbalance typically leads to problems in classification models, including overfitting, underfitting, and misinterpretation of performance metrics. Methods to address this issue include resampling, augmentation, regularization techniques, and adjustment of loss functions. In this paper, we focus on loss function adjustment, particularly comparing the performance of various configurations of loss functions (Cross Entropy, Balanced Cross Entropy, two settings of Focal Loss: 𝛼 = 1 and 𝛼 = Balanced, Asymmetric Loss) on Multi-Class black-box video data with imbalance issues. The comparison is conducted using the I3D, and R3D_18 models.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.7 no.5
• /
• pp.1044-1051
• /
• 2003

This paper surveys the EBP(Error Back Propagation) learning, the Cross Entropy function and the LBL(Layer By Layer) learning, which are used for learning the MLP(Multi Layer Perceptrons). We compare the merits and demerits of each learning method in the handwritten digit recognition. Although the speed of EBP learning is slower than other learning methods in the initial learning process, its generalization capability is better. Also, the speed of Cross Entropy function that makes up for the weak points of EBP learning is faster than that of EBP learning. But its generalization capability is worse because the error signal of the output layer trains the target vector linearly. The speed of LBL learning is the fastest speed among the other learning methods in the initial learning process. However, it can't train for more after a certain time, it has the lowest generalization capability. Therefore, this paper proposes the standard of selecting the learning method when we apply the MLP.

• Annual Conference on Human and Language Technology
• /
• 2017.10a
• /
• pp.56-59
• /
• 2017

CNN(Convolutional Neural Network)을 이용하여 발화 주제 다중 분류 task를 multi-labeling 방법과, cluster 방법을 이용하여 수행하고, 각 방법론에 MSE(Mean Square Error), softmax cross-entropy, sigmoid cross-entropy를 적용하여 성능을 평가하였다. Network는 음절 단위로 tokenize하고, 품사정보를 각 token의 추가한 sequence와, Naver DB를 통하여 얻은 named entity 정보를 입력으로 사용한다. 실험결과 cluster 방법으로 문제를 변형하고, sigmoid를 output layer의 activation function으로 사용하고 cross entropy cost function을 이용하여 network를 학습시켰을 때 F1 0.9873으로 가장 좋은 성능을 보였다.

• Journal of the Korean Statistical Society
• /
• v.26 no.1
• /
• pp.131-146
• /
• 1997

In this paper the problem of measuring the stochastic dependence among sets fo random variates is considered, and attention is specifically directed to forming a single well-defined measure of the dependence among sets of normal variates. A new information theoretic measure of the dependence called dependence index (DI) is introduced and its several properties are studied. The development of DI is based on the generalization and normalization of the mutual information introduced by Kullback(1968). For data analysis, minimum cross entropy estimator of DI is suggested, and its asymptotic distribution is obtained for testing the existence of the dependence. Monte Carlo simulations demonstrate the performance of the estimator, and show that is is useful not only for evaluation of the dependence, but also for independent model testing.

• Journal of the Semiconductor & Display Technology
• /
• v.22 no.2
• /
• pp.40-44
• /
• 2023

The defect detection based on deep learning requires minimal loss and high accuracy to pinpoint product defects. In this paper, we confirm the loss rate of deep learning training based on disc-shaped artificial scaffold images. It is intended to compare the performance of Cross-Entropy functions used in object detection algorithms. The model was constructed using normal, defective artificial scaffold images and category cross entropy and sparse category cross entropy. The data was repeatedly learned five times using each loss function. The average loss rate, average accuracy, final loss rate, and final accuracy according to the loss function were confirmed.

• Journal of the Korean Society of Safety
• /
• v.38 no.1
• /
• pp.1-8
• /
• 2023

It is extremely important to design safety-guaranteed industrial processes because such process determine the ultimate outcomes of industrial activities, including worker safety. Application of artificial intelligence (AI) in industrial safety involves modeling industrial safety systems by using vast amounts of safety-related data, accident prediction, and accident prevention based on predictions. As a preliminary step toward realizing AI-based industrial safety in Korea, this study discusses neural network-based modeling of industrial safety systems. The input variables that are the most discriminatory relative to the output variables of industrial safety processes are selected using two information-theoretic measures, namely entropy and cross entropy. Normalized frequency and severity of industrial accidents are selected as the output variables. Our simulation results confirm the effectiveness of the proposed neural network model and, therefore, the feasibility of extending the model to include more input and output variables.