• 한국경영과학회:학술대회논문집
• /
• 한국경영과학회 1998년도 추계학술대회 논문집
• /
• pp.36-39
• /
• 1998

It is very difficult for the large-scale retailers, who deal with tens of thousands of items, to price all the items dynamically reflecting all the constraints and policies. In spite of its importance, the prices are determined by human experts because the process of setting the prices of all the items is not established yet. To solve this problem, we adopt a mixed model that combines three typical pricing models: cost-plus model, competition-oriented model, and demand-oriented model. Since each model an be converted to a set of constraints in point and interval forms, solving the pricing problem with the three groups of models requires an algorithm which can solve the problem with weighted constraints of intervals and points. So we have devised an algorithm named “Point Determination Algorithm”. From the rules that represents tile models, the constraints are extracted to be solvable by tile Point Determination Algorithm. A prototype KAPA (Knowledge Assisted pricing Advisor) is developed with this idea using the expert system environment UNIK - a tool developed by KAIST. According to the experiment with 76 items in comparison with 53 human pricing experts we confirmed that the KAPA can perform highly consistent with human experts. This implies KAPA system is applicable to pricing millions of items dynamically.

• 한국음향학회지
• /
• 제18권8호
• /
• pp.42-47
• /
• 1999

본 논문에서는 시변가산유색잡음에 오염된 음성신호의 향상을 위한 MIMM(mixture interacting multiple model) 알고리즘을 제안 한다. 제안된 방법에서 음성신호는 혼합 은닉필터모델(hidden filter model: HFM)로 모델링되며, 잡음신호는 하나의 은닉필터로 모델링 된다. MIMM 알고리즘은 혼합 은닉필터모델에 의한 다중 Kalman 필터링에 기초한 회귀계산이기 때문에 계산량이 많아, Kalman 필터링 식의 구조적 측면에서 효율적인 계산이 가능하도록 알고리즘을 구현했다. 시뮬레이션 결과, 제안된 방법이 기존의 결과 [4,5]에 비하여 성능향상이 이루어 졌음을 보여 준다.

• 제어로봇시스템학회논문지
• /
• 제19권11호
• /
• pp.1022-1028
• /
• 2013

It is well knowns that based on the CSP (Common Spatial Pattern) algorithm, the linear projection of an EEG (Electroencephalography) signal can be made to spaces that optimize the discriminant between two patterns. Sharing disadvantages from linear time invariant systems, CSP suffers from the non-stationary nature of EEGs causing the performance of the classification in a BCI (Brain-Computer Interface) system to drop significantly when comparing the training data and test data. The author has suggested a simple idea based on the parallel model of CSP filters to improve the performance of BCI systems. The model was tested with a simple CSP algorithm (without any elaborate regularizing methods) and a perceptron learning algorithm as a classifier to determine the improvement of the system. The simulation showed that the parallel model could improve classification performance by over 10% compared to conventional CSP methods.

• 한국멀티미디어학회논문지
• /
• 제7권1호
• /
• pp.80-90
• /
• 2004

본 논문에서는 패치별 EGI 분포를 이용한 3D 메쉬 모델(mesh model) 워터마킹 알고리즘을 제안하였다. 제안한 알고리즘에서는 기하학적 변형에 견고하기 위하여 3D 메쉬 모델을 6개 패치로 분할한다. 그리고 위상학적 변형에 견실한 특성을 가지는 EGI 분포를 각 패치별로 구한다. 그리고 동일한 워터마크 비트열을 각 패치의 EGI 분포 중에서 길이가 큰 면체에 투영된 메쉬 법선 벡터들에 각각 삽입한다. 그러므로 제안한 알고리즘은 기하학적 및 위상학적 변형에 견고하며, 특히 워터 마크 추출시 복잡한 재표본화 과정 및 원 모델이 필요 없다. 본 논문에서 제안한 워터마킹 알고리즘의 성능 평가를 위한 모의 실험에서 워터마크가 삽입된 모델의 비가시성 및 다양한 공격에 대한 견고성이 우수함을 확인하였다.

• 대한교통학회지
• /
• 제21권1호
• /
• pp.65-75
• /
• 2003

교통혼잡 문제가 점점 심각해짐에 따라 대중교통의 중요성은 날로 부각되며, 대중교통을 지원하기 위한 정책들이 속속 입안되고 있어 대중교통을 심도 있게 분석할 수 있는 틀의 개발은 필연적이라 할 수 있다. 이에 본 연구는 대중교통의 특성을 고려하는 대중교통 통행배정모형의 개발을 목적으로 수행되었다. 본 연구에서 개발한 대중교통 통행배정모형은 기존의 대중교통 통행배정모형이 개별차량과 다른 대중교통의 특성을 정확히 반영하고 있지 못하다는 한계를 극복하고자, 최적경로 탐색에는 유전자 알고리즘(Genetic Algorithm)을 통행량 배정에는 로짓모형을 기반으로 한 확률적 통행량 배정모형(Stochastic Network Loading Model)을 이용하였다.

• Geomechanics and Engineering
• /
• 제21권6호
• /
• pp.583-598
• /
• 2020

This study is attempted to propose a new hybrid artificial intelligence model called integrative genetic algorithm with multivariate adaptive regression splines (GA-MARS) for settlement prediction of shallow foundations on sandy soils. In this hybrid model, the evolution algorithm - Genetic Algorithm (GA) was used to search and optimize the hyperparameters of multivariate adaptive regression splines (MARS). For this purpose, a total of 180 experimental data were collected and analyzed from available researches with five-input variables including the bread of foundation (B), length to width (L/B), embedment ratio (D_f/B), foundation net applied pressure (q_net), and average SPT blow count (N_SPT). In further analysis, a new explicit formulation was derived from MARS and its accuracy was compared with four available formulae. The attained results indicated that the proposed GA-MARS model exhibited a more robust and better performance than the available methods.

• 한국시뮬레이션학회:학술대회논문집
• /
• 한국시뮬레이션학회 2005년도 추계학술대회 및 정기총회
• /
• pp.71-76
• /
• 2005

XML is rapidly becoming one of the most widely adopted technologies for information exchange and representation on the World Wide Web. However, the large part of data is still stored in a relational database. Hence, we need to convert relational data into XML documents. The most important point of the conversion is to reflect referential integrities In relational schema model to XML schema model exactly. Until now, FT, NeT and CoT are suggested as existing approaches for conversion from the relational schema model to the XML schema model but these approaches only reflect referential integrities which are defined explicitly for conversion. In this paper, we suggest an algorithm for automatic extraction of implicit referential integrities such as foreign key constraints which is not defined explicitly in the initial relational schema model. We present translated XML documents by existing algorithms and suggested algorithms as comparison evaluation. We also compare suggested algorithm and conventional algorithms by simluation in accuracy part.

• Journal of Power Electronics
• /
• 제15권4호
• /
• pp.1139-1147
• /
• 2015

In this paper, an algorithm based on a model analysis of the online calculation of the root-mean-square (RMS) value of welding current for single-phase AC resistance spot welding (RSW) was developed. The current is highly nonlinear and typically non-sinusoidal, which makes the measuring and controlling actions difficult. Though some previous methods focused on this issue, they were so complex that they could not be effectively used in general cases. The electrical model of a single-phase AC RSW was analyzed, and then an algorithm for online calculation of the RMS value of the welding current was presented. The description includes two parts, a model-dependent part and a model-independent part. Using a previous work about online measurement of the power factor angle, the first part can be solved. For the second part, although the solution of the governing equation can be directly obtained, a lot of CPU time must be consumed due to the fact that it involves a lot of complex calculations. Therefore, a neural network was employed to simplify the calculations. Finally, experimental results and a corresponding analysis showed that the proposed algorithm can obtain the RMS values with a high precision while consuming less time when compared to directly solving the equations.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권11호
• /
• pp.4028-4042
• /
• 2021

Aiming at the problem of software defect prediction difficulty caused by insufficient software defect marker samples and unbalanced classification, a semi-supervised software defect prediction model based on a tri-training algorithm was proposed by combining feature normalization, over-sampling technology, and a Tri-training algorithm. First, the feature normalization method is used to smooth the feature data to eliminate the influence of too large or too small feature values on the model's classification performance. Secondly, the oversampling method is used to expand and sample the data, which solves the unbalanced classification of labelled samples. Finally, the Tri-training algorithm performs machine learning on the training samples and establishes a defect prediction model. The novelty of this model is that it can effectively combine feature normalization, oversampling techniques, and the Tri-training algorithm to solve both the under-labelled sample and class imbalance problems. Simulation experiments using the NASA software defect prediction dataset show that the proposed method outperforms four existing supervised and semi-supervised learning in terms of Precision, Recall, and F-Measure values.

• 한국군사과학기술학회지
• /
• 제7권4호
• /
• pp.5-12
• /
• 2004

The tracking algorithm based on the interacting multiple model(IMM) requires a considerable number of sub-models for the various maneuvering targets in order to have a good performance. But it is not feasible to use the nm algorithm in the real system because of the computational burden. Therefore, we need an algorithm which requires less computing resources while maintaining a good performance. In this paper, we propose a fuzzy interacting multiple model algorithm(FIMMA) for the tracking of maneuvering targets, which uses a minimal number of sub-models by considering the maneuvering properties and adjusts the mode transition probabilities by using the mode probability as a fuzzy input. In order to verify the performance of FIMMA, the developed algorithm is applied to the tracking of i borne targets. Simulation results show that the FIMMA is very effective in the tracking of maneuvering targets.