• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1998년도 하계학술대회 논문집 B
• /
• pp.509-511
• /
• 1998

Design of fuzzy logic controllers encounters difficulties in the selection of optimized membership function and fuzzy rule base, which is traditionally achieved by tedious trial-and-error process. In this paper We proposed a new method to generate fuzzy logic controllers throught genetic algorithm(GA). The controller design space is coded in base-7 strings chromosomes, where each bit gene matches the 7 discrete fuzzy value. The developed approach is subsequently applied to the design of proportional plus integral type fuzzy controller for a do-servo motor control system. It was presented in discrete fuzzy linguistic value, and used a membership function with Gaussian curve. The performance of this control system is demonstrated higher than that of a conventional PID controller and fuzzy logic controller(FLC).

• The Journal of the Acoustical Society of Korea
• /
• 제18권4E호
• /
• pp.43-50
• /
• 1999

A fast converging coding algorithm that adaptively selects the modes of macroblocks is introduced. For a given frame, the optimal modes are selected based on the decision curves that minimize the overall distortion at a given bit rate. The method proposed in this paper is different from the conventional ones in that it does not manipulate the quantizer to meet the target bit rate but it satisfies the target bit rate by finding optimal modes of macroblocks which result consistent visual quality. Lagrange multiplier of the unconstrained cost function is controlled to trigger decision curves to generate appropriate modes to meet bit rate and the curve is obtained by utilizing simulated annealing optimization technique. The algorithm is implemented within H.261 video codec and simulation results demonstrate superior visual quality.

• 한국컴퓨터정보학회논문지
• /
• 제22권11호
• /
• pp.111-116
• /
• 2017

This paper proposes a data mining approach to predicting stock price direction. Stock market fluctuates due to many factors. Therefore, predicting stock price direction has become an important issue in the field of stock market analysis. However, in literature, there are few studies applying data mining approaches to predicting the stock price direction. To contribute to literature, this paper proposes comparing single classifiers and ensemble classifiers. Single classifiers include logistic regression, decision tree, neural network, and support vector machine. Ensemble classifiers we consider are adaboost, random forest, bagging, stacking, and vote. For the sake of experiments, we garnered dataset from Korea Stock Exchange (KRX) ranging from 2008 to 2015. Data mining experiments using WEKA revealed that random forest, one of ensemble classifiers, shows best results in terms of metrics such as AUC (area under the ROC curve) and accuracy.

• Journal of the Korean Association of Oral and Maxillofacial Surgeons
• /
• 제29권1호
• /
• pp.43-47
• /
• 2003

Anterior set back segmental surgery has been used for shortening the period of orthodontic treatment in case of bimaxillary or maxillary protrusion. In most cases, it requires pre-operative orthodontic treatment. Through properly performed leveling and tooth aligning, the operative porcedure can be easier and post-operative occlusal stability can be increased. But it takes time for orthodontic treatment. Recently, we have been using anterior segmental surgery before orthodontic treatment and have reliable results from that. Therefore, we have to consider arch shape, curve of Spee, tooth selection to be extracted for obtaining of post-operative occlusal stability without pre-operative orthodontic treatment.

• 한국자동차공학회논문집
• /
• 제11권5호
• /
• pp.60-69
• /
• 2003

The simplified two-color method is proposed which can estimate the temperature distribution and the soot density of the whole flame with the image analysis of the high-speed photographs. The factors influenced on its processing were examined, for example, the selection of the wave length, the kind of films, the preparation of the calibration curve between the radiance of flame and the luminance temperature. The simplified two-color method reported in this paper can be used as a tool for the improvement of the combustion process in direct injection diesel engine.

• Tribology and Lubricants
• /
• 제20권3호
• /
• pp.132-139
• /
• 2004

Many researches about the contacts between cam and follower have investigated EHL film thickness either without dynamic loading effect or only with curve fitting formula such as Dowson-Hamrock's, because including squeeze film effect makes it hard to obtain convergence and stability of computation. Therefore, inaccurate information about minimum film thickness without dynamic loading condition causes inappropriate design of cam profiles and wrong selection of cam and follower materials. In this work, computation tools both for kinematics and dynamics of valve train system of push-rod type and for fluid film thickness with elastic deformation on the basis of dynamic loading condition with multigrid multi-level method is developed. The computational results of minimum film thickness with the respects of both static and dynamic loading conditions are compared for the contact of flat follower over the entire cycle.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 2005년도 추계학술대회 논문집
• /
• pp.781-785
• /
• 2005

Cutting forces and tool deflection in end milling are represented as the closed form of tool rotational angle and cutting conditions. The discrete cutting forces caused by tool entry and exit are continued using the Fourier series expansion. Tool deflection is predicted by direct integration of the distributed loads on cutting edges. Cutting conditions, tool geometry, run-outs and the stiffness of tool clamping pan are considered for cutting forces and tool deflection estimation. Compared to numerical methods, the presented method has advantages in short prediction time and the effects of feeding and run-outs on cutting forces and tool deflection can be analyzed quantitatively. This research can be effectively used in real time machining error estimation and cutting condition selection for error minimization since the ferm accuracy is easily predicted by tool deflect ion curve.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제13권3호
• /
• pp.1722-1737
• /
• 2019

We developed an insider threat detection model to be used by organizations that repeat tasks at regular intervals. The model identifies the best combination of different feature selection algorithms, unsupervised learning algorithms, and standard scores. We derive a model specifically optimized for the organization by evaluating each combination in terms of accuracy, AUC (Area Under the Curve), and TPR (True Positive Rate). In order to validate this model, a four-year log was applied to the system handling sensitive information from public institutions. In the research target system, the user log was analyzed monthly based on the fact that the business process is processed at a cycle of one year, and the roles are determined for each person in charge. In order to classify the behavior of a user as abnormal, the standard scores of each organization were calculated and classified as abnormal when they exceeded certain thresholds. Using this method, we proposed an optimized model for the organization and verified it.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2010년도 추계학술발표대회
• /
• pp.1012-1015
• /
• 2010

VANET(Vehicle Ad hoc NETwork)에서는 차량들이 불규칙한 움직임과 빠른 이동성을 가진다. 이러한 VANET 의 특성 때문에 메시지를 목적지까지 올바르게 전송하기 위해 전송 차량의 다음 홉 선택은 매우 중요하다. 본 논문에서는 기존에 직선도로나 교차로에서 사용되는 메시지 전송 프로토콜로는 $90^{\circ}$커브에서 메시지 전송 실패를 가져올 수 있는 경우에 대해 기술하고, 이를 해결하기 위해 커브에서 중앙선을 기준으로 전송 차량이 있는 도로와 그 반대편 도로의 범위를 구하여 전송 차량이 다음 홉 선택 시 자신과 동일한 도로에 위치한 차량만 고려하도록 함으로써 신뢰성 있는 다음 홉 선택이 이루어지도록 하는 방안을 제안한다.

• Communications for Statistical Applications and Methods
• /
• 제31권2호
• /
• pp.203-212
• /
• 2024

The area under the ROC curve (AUC) is one of the most common criteria used to measure the overall performance of binary classifiers for a wide range of machine learning problems. In this article, we propose a L₁-penalized AUC-optimization classifier that directly maximizes the AUC for high-dimensional data. Toward this, we employ the AUC-consistent surrogate loss function and combine the L₁-norm penalty which enables us to estimate coefficients and select informative variables simultaneously. In addition, we develop an efficient optimization algorithm by adopting k-means clustering and proximal gradient descent which enjoys computational advantages to obtain solutions for the proposed method. Numerical simulation studies demonstrate that the proposed method shows promising performance in terms of prediction accuracy, variable selectivity, and computational costs.