• 한국지능시스템학회논문지
• /
• 제11권4호
• /
• pp.293-301
• /
• 2001

본 논문은 신경회로망과 같은 경험적 모델에서 출력별로 신뢰 구간을 추정하는 세 가지 대표적인 방법을 검토하고, 검토한 방법을 뉴로-퍼지 모델에 적용하여 장단점을 비교 분석한다. 본 논문에서 고려한 출력별 신뢰 구간 계산 방법은 cross-validation을 이용한 stacked generalization, 회귀 모델에서 유도된 predictive error bar, 지역 표현하는 신경회로망의 특성에 기반한 local reliability measure이다. 간단한 함수 근사화 문제와 혼돈 시계열 예측 문제를 이용하여 모의 실험을 수행하고, 세 가지 신뢰도 추정 방법의 성능을 정량적, 정성적으로 비교 분석한다. 분석 결과를 기초로 각 방법의 장단점 및 특성을 고찰하고, 모델링 문제에서 모델의 출력별 신뢰 구간 계산 방법의 실제 적용 가능성을 탐색한다.

• 대한전기학회논문지:전력기술부문A
• /
• 제54권10호
• /
• pp.467-474
• /
• 2005

This paper presents a systematic method to compute a reliability measure for a short term electrical load forecasting system using neuro-fuzzy models. It has been realized that the reliability computation is essential for a load forecasting system to be applied practically. The proposed method employs a local reliability measure in order to exploit the local representation characteristic of the neuro-fuzzy models. It, hence, estimates the reliability of each fuzzy rule learned. The design procedure of the proposed short term load forecasting system is as follows: (1) construct initial structures of neuro-fuzzy models, (2) store them in the initial structure bank, (3) train the neuro-fuzzy model using an appropriate initial structure, and (4) compute load prediction and its reliability. In order to demonstrate the viability of the proposed method, we develop an one hour ahead load forecasting system by using the real load data collected during 1996 and 1997 at KEPCO. Simulation results suggest that the proposed scheme extends the applicability of the load forecasting system with the reliably computed reliability measure.

• 한국경제지리학회지
• /
• 제11권3호
• /
• pp.492-507
• /
• 2008

일반적으로 사용되는 많은 네트워크 분석방법들은 비공간적인 측면에서 네트워크를 인식하는 경향이 있다. 가령 네트워크의 신뢰성(reliability)을 측정하고 중요한 집합점(node)을 찾는 기본적인 문제에 있어서도, 이러한 분석방법들은 집합점의 속성이나 연결선(edge)의 거리와 같은 네트워크의 공간적인 요소를 배제한 채, 네트워크요소들의 위상적인 접합 여부만을 고려한다. 그에 따라 이러한 네트워크 분석은 실제 공간에서의 네트워크 특성을 반영하지 못하는 제한적인 결과만을 도출하게 한다. 그러나 본 연구는 국지네트워크의 자기상관 (local network autocorrelation measure) 값을 이용하여 이러한 문제의 해결을 시도하였다. 국가자기상관 값은 공간객체들의 유사성이나 군집성을 개별적으로 측정하여 각 객체들의 중요도를 나타낸다. 본 연구는 미국의 주요 인터넷 네트워크를 disjoint product method와 Getis-Ord의 G 수치를 이용하여 분석하였으며 그 과정에서 인구와 신뢰도를 변수로서 이용하였다. 그 결과 국지네트워크의 자기상관값은 주요한 집합점들의 국지적인 군집정도를 나타냈고, 이러한 연구 결과는 연구 촛점이 국지네트워크의 범위나 그 영향일 경우, 국지자기상관값의 이용이 더 실용적이고 현실적임을 보여준다.

• 대한지역사회영양학회지
• /
• 제11권1호
• /
• pp.116-123
• /
• 2006

The purposes of this study were to develop a tool for assessing internal branding performance and to identify the difference of internal branding performance between local and international hotels. The internal branding performance measure was developed through literature studies and face-to-face in-depth interviews with managers in various departments of 5 deluxe hotels. The questionnaires were distributed to 460 employees in food and beverage departments of 11 deluxe hotels in Seoul, and a total of 398 questionnaires were used for analysis ($86.5\%$). The statistical analyses were completed using SPSS Win (12.0) for descriptive analysis, reliability analysis and t-test, and AMOS (5.0) for confirmatory factor analysis. The results of this study showed that the measure developed in this study was reliable and valid for assessing internal branding performances of hotels. In addition, the internal branding performances of hotels were high at understanding of brand dimensions and relatively low at work and organizational culture and human resource management dimensions, and local hotels showed lower levels of performance than international hotels at the dimensions of work and organizational culture (t = -3.801, P < 0.001) and human resource management(t = -5.397, p < 0.001). In conclusion, it is considered that a strong internal branding strategy should be established for successful brand management, and the practical measure for internal branding performance would be a key factor.

• The Journal of the Acoustical Society of Korea
• /
• 제28권3E호
• /
• pp.109-117
• /
• 2009

Identity recognition in real environment with a reliable mode is a key issue in human computer interaction (HCI). In this paper, we present a robust person identification system considering score-based optimal reliability measure of audio-visual modalities. We propose an extension of the modified reliability function by introducing optimizing parameters for both of audio and visual modalities. For degradation of visual signals, we have applied JPEG compression to test images. In addition, for creating mismatch in between enrollment and test session, acoustic Babble noises and artificial illumination have been added to test audio and visual signals, respectively. Local PCA has been used on both modalities to reduce the dimension of feature vector. We have applied a swarm intelligence algorithm, i.e., particle swarm optimization for optimizing the modified convection function's optimizing parameters. The overall person identification experiments are performed using VidTimit DB. Experimental results show that our proposed optimal reliability measures have effectively enhanced the identification accuracy of 7.73% and 8.18% at different illumination direction to visual signal and consequent Babble noises to audio signal, respectively, in comparison with the best classifier system in the fusion system and maintained the modality reliability statistics in terms of its performance; it thus verified the consistency of the proposed extension.

• 한국광학회:학술대회논문집
• /
• 한국광학회 1990년도 제5회 파동 및 레이저 학술발표회 5th Conference on Waves and lasers 논문집 - 한국광학회
• /
• pp.242-247
• /
• 1990

By self-homodyne method we measured the relative phase fluctuation of a light wave. Balanced detection system can eliminate local oscillator excess noise and multiport detection makes it possible ot measure the phase change of the signal beam. Deriving the SB(Signal Beam) and the LO(Local Oscillator) from the same laser source, we find the SB maintain constant phase relative to that of the LO. We have introduced a phase fluctuation in the SB by modulating the beam path of the SB. The measured phase fluctuation agreed well with the predicted one, thereby we confirmed the reliability of our system.

• 한국식생활문화학회지
• /
• 제21권1호
• /
• pp.65-70
• /
• 2006

The purpose of this study were to measure brand personalities of deluxe hotels in Seoul, and to identify the difference of brand personality between local and international hotels. The questionnaires developed for this study were distributed to 460 employees in kitchen and F&B departments of 11 deluxe hotels in Seoul. A total of 398 questionnaires were used for anaylsis(86.5%) and the statistical analyses were completed using SPSS Win(12.0) for descriptive analysis, reliability analysis and t-test, and AMOS(5.0) for confirmatory factor analysis. The results of this study showed that deluxe hotels have brand personalities relatively strong at 'affection', 'sophistication', 'competence' dimensions, and there was a significant difference by hotel nationality. The brand personality scores of international brand hotels perceived by employees were high at the 'excitement'(p<0.001), 'sophistication'(p<0.001), and 'competence'(p<0.01) dimensions, whereas local hotels were considered more obedient(p<0.01). Overall, it could be a key factor for successful brand management that establish a distinctive brand personality, and a localized brand personality measure will lead to more desirable decision making.

• 한국신재생에너지학회:학술대회논문집
• /
• 한국신재생에너지학회 2006년도 추계학술대회
• /
• pp.153-156
• /
• 2006

The domestic solar radial ion data have been measured at 16 different sites all over the country since the beginning of 1980. It is very important that the fundamental data for the estimation and assessment of local solar radiation can be secured this project. In order to estimate available energy resource from solar radiation, it is necessary to have enough data, more than 30 years In any country. However since we have collected solar radiation(global radial ion including direct normal radiation) data only for 10 years we still need to measure insolation to secure the reliability and standardization of measured local solar radial ion data. In brief, the major activities on this R&D include rout me maintenance of the national network for insolation data measurement, evaluation of the collected data, and reliability enhancement for assessing the quality of solar radiation data as well.

• 대한기계학회:학술대회논문집
• /
• 대한기계학회 2003년도 춘계학술대회
• /
• pp.1130-1135
• /
• 2003

The need to increase the reliability of a structural system has been significantly brought in the procedure of real designs to consider, for instance, the material properties or geometric dimensions that reveal a random or incompletely known nature. Reliability based design optimization of a real system now becomes an emerging technique to achieve reliability, robustness and safety of these problems. Finite element analysis program and the reliability analysis program are necessary to evaluate the responses and the probabilities of failure of the system, respectively. Moreover, integration of these programs is required during the procedure of reliability based design optimization. It is well known that reliability based design optimization can often have so many local minima that it cannot converge to the specified probability of failure. To overcome this problem, barrier function method in reliability based design optimization is suggested. To illustrate the proposed formulation, reliability based design optimization of a bracket is performed. AMV and FORM are employed for reliability analysis and their optimization results are compared based on the accuracy and efficiency.

• 패션비즈니스
• /
• 제16권4호
• /
• pp.75-87
• /
• 2012

This study was to investigate the conversion intent of fashion store among college undergraduates. The conversion intent in internet shopping mall and local store was associated with store attributes and relationship quality. A random sample of undergraduate students completed a questionnaire that contained measure of Store attributes, Relationship quality, Conversion intent of fashion store. The results revealed that store attributes of local store were store comport/salesman service, product diversity, store location, price and the store attributes of internet shopping mall were product diversity/easy of navigation, store reliability, screen configuration, sales promotion. Also, the result founded that the conversion intent of fashion store is related to store comport/salesman service in local store and is related to product diversity/easy of navigation, screen configuration, promotion in internet shopping mall. Relationship quality is highly effected of the conversion intent of local store and internet shopping mall. We expect that this research results can be used as a basic material for another study and setting up the marketing strategies in fashion store.