• Journal of Power Electronics
• /
• 제15권6호
• /
• pp.1456-1467
• /
• 2015

This paper presents a novel close-loop control scheme based on small signal modeling and weighted composite voltage feedback for a three-phase input and single-phase output Matrix Converter (3-1MC). A small non-polar capacitor is employed as the decoupling unit. The composite voltage weighted by the load voltage and the decoupling unit voltage is used as the feedback value for the voltage controller. Together with the current loop, the dual-loop control is implemented in the 3-1MC. In this paper, the weighted composite voltage expression is derived based on the sinusoidal pulse-width modulation (SPWM) strategy. The switch functions of the 3-1MC are deduced, and the average signal model and small signal model are built. Furthermore, the stability and dynamic performance of the 3-1MC are studied, and simulation and experiment studies are executed. The results show that the control method is effective and feasible. They also show that the design is reasonable and that the operating performance of the 3-1MC is good.

• Journal of the Korean Data and Information Science Society
• /
• 제24권2호
• /
• pp.311-320
• /
• 2013

본 연구의 목적은 학생 (인구)수를 예측하기 위한 방법을 제시하는데 있다. 일반화된 가중비례모형들을 제시하고, 제시된 모형들을 이용하여 2029년까지의 학생 (인구)수를 추계하였다. 몬테칼로 모의실험의 결과 제시된 가중비례추정모형은 인구예측에 있어서 상당한 신뢰성과 예측력을 보인다. 결론적으로 추정된 고등학교 3학년 학생수는 대학들의 입학정원 수와 비교하면, 2019년부터 고3 학생수가 적게 나타나지만, 현재 대학들의 실제 입학자 수와 비교하면 2014년부터 입학생 수가 부족한 현상이 나타난다. 그러므로 입학정원과 입학정원 외의 신입생 수를 줄이지 않는 경우에는 2015년부터 입학생을 채우지 못하는 학교들이 생겨나게 될 것이다.

• 한국지능시스템학회논문지
• /
• 제19권2호
• /
• pp.206-211
• /
• 2009

본 논문은 가중 퍼지소속함수 기반 신경망(Neural Network with Weighted Fuzzy Membership Functions, NEWFM)기반의 자동 특징 추출기법을 사용하여 1일 후의 주식 예측을 하는 방안을 제안하고 있다. 비중복면적 분산측정 법에 의해 중요도가 가장 낮은 특징입력을 자동적으로 하나씩 제거하면서 최소의 특징입력을 선택하였다. 특징입력으로써 CPP$_{n,m}$(Current Price Position of the day n)과 최근 32일간의 CPP$_{n,m}$을 웨이블릿 변환한 38개의 계수들 중 비중복면적 분산측정법에 의해서 자동적으로 추출된 2개의 계수가 사용되었다 제안된 방법으로 1989년부터 1998년까지의 실험군을 사용한 결과로써 60.93%의 예측율을 나타내었다.

• 대한디지털의료영상학회논문지
• /
• 제14권1호
• /
• pp.21-29
• /
• 2012

The purpose of this study is to evaluate the mutual relations by measuring SNR from T2 weighted image and ADC values on the basis of the stiffness values from liver tissues. This study was conducted that total 37 people(23 of males and 11 of females) were taken the liver MRI examination and average age was $54.5{\pm}12.7$ years old. The equipment was MAGNETOM Skyra 3.0T (SIEMENS, Erlangen, Germany) and 32 channel body-array coil. The examination were conducted with HASTE T2 weighted image by axial plane, Spin-echo EPI (echo planner image) DWI (b-value = 800) and Magnetic resonance elastography. The ROIs (region of interest: 200-300 $mm^2$) were established on the basis of the first axial stiffness image corresponded 95% confidence interval from axial stiffness image and then were measured values. After drawing the grid lines, signals were measured SNR from T2 weighted image and ADC values on the same locations that were analysed other 3 planes respectively. The results were showed correlation (0.057) that were increased to SNR from T2 weighted image by increasing stiffness value that no significant difference statistically p = 0.003. Other results were showed correlations (-0.301) that were decreased to ADC values by increasing stiffness values that no significant difference statistically p = 0.088. In the 3.0T equipment, the results may be error in much the same fashion as the 1.5T from ADC values by evaluation of fibrosis stage. However, Magnetic resonance elastography would be useful method that is used to diagnose exactly liver fibrosis stages in the 3.0T.

• Journal of Information Processing Systems
• /
• 제16권2호
• /
• pp.329-342
• /
• 2020

In recent years, emotional text classification is one of the essential research contents in the field of natural language processing. It has been widely used in the sentiment analysis of commodities like hotels, and other commentary corpus. This paper proposes an improved W-LDA (weighted latent Dirichlet allocation) topic model to improve the shortcomings of traditional LDA topic models. In the process of the topic of word sampling and its word distribution expectation calculation of the Gibbs of the W-LDA topic model. An average weighted value is adopted to avoid topic-related words from being submerged by high-frequency words, to improve the distinction of the topic. It further integrates the highest classification of the algorithm of support vector machine based on the extracted high-quality document-topic distribution and topic-word vectors. Finally, an efficient integration method is constructed for the analysis and extraction of emotional words, topic distribution calculations, and sentiment classification. Through tests on real teaching evaluation data and test set of public comment set, the results show that the method proposed in the paper has distinct advantages compared with other two typical algorithms in terms of subject differentiation, classification precision, and F1-measure.

• 인터넷정보학회논문지
• /
• 제9권1호
• /
• pp.129-135
• /
• 2008

본 논문은 가중 퍼지소속함수 기반 신경망(Neural Network with Weighted Fuzzy Membership Functions, NEWFM)을 사용하여 생성된 퍼지규칙과 비중복면적 분산 측정법에 의해 추출된 최소의 특징입력을 이용하여, 1일 후의 KOSPI 예측을 하는 방안을 제안하고 있다. NEWFM은 KOSPI의 최근 32일 동안의 CPPn,m(Current Price Position of day n for n-1 to n-m days)을 이용하여 1일 후의 KOSPI 상승과 하락을 예측한다. 특징입력으로써 CPPn,m과 최근 32일간의 CPPn,m을 웨이블릿 변환한 38개의 계수들 중 비중복면적 분산 측정법을 적용하여 추출된 5개의 계수가 사용되었다. 제안된 방법으로 1991년부터 1998년까지의 실험군을 사용한 결과 평균 67.62%의 예측율을 나타내었다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권3호
• /
• pp.894-907
• /
• 2022

The features extracted by convolutional neural networks are more descriptive of images than traditional features, and their convolutional layers are more suitable for retrieving images than are fully connected layers. The convolutional layer features will consume considerable time and memory if used directly to match an image. Therefore, this paper proposes a feature weighting and region integration method for convolutional layer features to form global feature vectors and subsequently use them for image matching. First, the 3D feature of the last convolutional layer is extracted, and the convolutional feature is subsequently weighted again to highlight the edge information and position information of the image. Next, we integrate several regional eigenvectors that are processed by sliding windows into a global eigenvector. Finally, the initial ranking of the retrieval is obtained by measuring the similarity of the query image and the test image using the cosine distance, and the final mean Average Precision (mAP) is obtained by using the extended query method for rearrangement. We conduct experiments using the Oxford5k and Paris6k datasets and their extended datasets, Paris106k and Oxford105k. These experimental results indicate that the global feature extracted by the new method can better describe an image.

• 대한전기학회논문지
• /
• 제35권5호
• /
• pp.193-198
• /
• 1986

In this paper, 1-168 hours ahead load prediction algorithm is developed for power system economic weekly operation. Total load is composed of three components, which are base load, week load and weather-sensitive load. Base load and week load are predicted by moving average and exponential smoothing method, respectively. The days of moving average and smoothing constant are optimally determined. Weather-sensitive load is modeled by linear form. The paramiters of weather load model are estimated by exponentially weighted recursive least square method. The load prediction of special day is very tedious, difficult and remains many problems which should be improved. Test results are given for the day of different types using the actual load data of KEPCO.

• International Journal of Control, Automation, and Systems
• /
• 제5권5호
• /
• pp.492-500
• /
• 2007

This paper focuses on the problem of guaranteed cost control for a class of uncertain linear delay systems with actuator failures. When actuators suffer "serious failure" the never failed actuators can not stabilize the system, based on switching strategy of average dwell time method, under the condition that activation time ratio between the system without actuator failure and the system with actuator failures is not less than a specified constant, a sufficient condition for exponential stability and weighted guaranteed cost performance are developed in terms of linear matrix inequalities (LMIs). Finally, as an example, a river pollution control problem illustrates the effectiveness of the proposed approach.

• 응용통계연구
• /
• 제19권1호
• /
• pp.171-181
• /
• 2006

모비율 차이의 구간 추정에서 표준으로 인식되고 있는 Wald 신뢰구간은 모비율 구간 추정과 마찬가지로 포함확률의 근사성에서 문제가 있다는 것이 알려져 있다. 이에 대한 대안으로 모비율 차이의 신뢰구간에 대한 많은 연구가 있어 왔으나 대부분의 신뢰구간은 매우 복잡한 과정을 통해 얻어지게 되어 있어 실용성에 대한 문제가 제기될 수 있다. 이와 비교하여 Agresti와 Caffo(2000)에 의해 제시된 신뢰구간은 매우 간편한 식에 의해 구할 수 있어 이해하기 쉽고 포함확률과 포함확률의 평균절대오차에 있어 다른 복잡한 신뢰 구간과 필적할 수 있다. 그러나 Agresti-Caffo 신뢰 구간은 포함확률이 명목 신뢰수준을 상회하는 보수적인 구간으로 알려져 있다. 본 논문에서는 이승천(2005)에서 이항비율의 신뢰구간을 구하기 위해 사용된 가중 Polya 사후분포를 이용하여 두 모비율 차이의 신뢰구간을 구하였다. 이렇게 구하여진 신뢰구간은 간편성은 물론 Agresti-Caffo 신뢰구간의 보수성을 개선하였다.