• Proceedings of the Korean Society of Computer Information Conference
• /
• 2009.01a
• /
• pp.25-30
• /
• 2009

패턴 인식에서 분류기 모형으로 많이 사용되는 퍼지 가중치 평균 분류기는 가중치를 적절히 설정함으로써 뛰어난 분류 성능을 얻을 수 있다는 장점이 있다. 그러나 일반적으로 가중치는 인식 문제 분야의 특성이나 해당 전문가의 지식이나 주관적 경험을 기반으로 설정되므로 설정된 가중치의 일관성과 객관성을 보장하기가 어려운 문제점을 갖고 있다. 따라서 이 논문에서는 퍼지 가중치 평균 분류기의 가중치를 설정하기 위한 객관적 기준을 제시하기 위하여 특징값들 간의 통계적 정보를 이용한 가중치 설정 기법들을 제안하였다. 제안한 기법들을 이용하여 UCI machine learning repository 사이트에서 제공되는 표준 데이터들 중의 하나인 Iris 데이터 세트를 이용하여 실험하고 그 결과를 비교, 분석하였다.

• Environmental Sciences Bulletin of The Korean Environmental Sciences Society
• /
• v.2 no.1
• /
• pp.1-8
• /
• 1998

Biomass of non-living brown seaweed Sargassum fluitans pretreated by different methods is capable of taking up more than $10\%$ (11 meq/g) of its dry weight in aluminum at a pH of 4.5. It is indicated that the biomass sequestered the aluminum in the form of polynuclear aluminum species. The fraction of $Al(OH)_3$ Precipitated in the aluminum nitrate solution without biomass at pH 4.5 increased as the Al concentration increased. Aluminum-alginate complex precipitated in the solution as alginate was partially released from the biomass. External colloidal precipitate occurring in native and protonated S. fluitans biomass sorption systems caused a significant difference in Al sorption isotherms determined by standard and desorption methods, respectively, Sodium ions added for pH adjustment were not sorbed at all in the presence of aluminum ions.

• East Asian Economic Review
• /
• v.21 no.2
• /
• pp.167-197
• /
• 2017

I build a small open economy (SOE) dynamic stochastic general equilibrium (DSGE) model to investigate the effect of a heterogeneous wage contract between regular and temporary workers on a macroeconomic volatility in a financially fragile economy. The imperfect financial market condition is captured by a quadratic financial adjustment cost for borrowing foreign assets, and the labor market friction is captured by a Nash bargaining process which is only available to the regular workers when they negotiate their wages with the firms while the temporary workers are given their wage which simply equals the marginal cost. As a result of impulse responsesto a domestic productivity shock, the higher elasticity of substitution between two types of workers and the lower weight on the regular workers in the firm's production process induce the higher volatilities in most variables. This is reasoned that the higher substitutability creates more volatile wage determination process while the lower share of the regular workers weakens their Nash bargaining power in the contract process.

• ETRI Journal
• /
• v.33 no.1
• /
• pp.71-77
• /
• 2011

Up to 10% of the global population suffers from color vision deficiency (CVD) [1], especially deuteranomaly and protanomaly, the conditions in which it is difficult to discriminate between red and green hues. For those who suffer from CVD, their career fields are restricted, and their childhood education is frustrating. There are many optical eye glasses on the market to compensate for this disability. However, although they are attractive due to their light weight, wearing these glasses will decrease visual brightness and cause problems at night. Therefore, this paper presents a supplementary device that comprises a head-mounted display and an image sensor. With the aid of the image processing technique of digital color space adjustment implemented in a high-speed field-programmable gate array device, the users can enjoy enhanced vision through the display without any decrease in brightness.

• Journal of KIISE:Information Networking
• /
• v.29 no.5
• /
• pp.495-500
• /
• 2002

We propose a new estimation method for rate adjustment in the face of a packet loss in the TFRC protocol, a TCP-Friendly congestion control protocol for UDP flows. Previous methods respond in a sensitive way to a single packet loss, resulting in oscillatory transmission behavior. This is an undesirable for multimedia services demanding constant bandwidth. The proposed TFRC provides more smooth and fair (against TCP flows) transmission through collective response based on multiple packets loss events. We show our "Exponential smoothing method" performs better than known "Weight smoothing method" in terms of smoothness and fairness.

• Communications for Statistical Applications and Methods
• /
• v.12 no.1
• /
• pp.29-38
• /
• 2005

In sample design, determining the weights of estimates becomes usually great influence on the result. In this article, raking methods are applied to different domain and depending on the range of the domain and sample size, the results of estimates are explained and compared. For the comparison, we use the MSE, MAE, MSPE and MAPE with Actual State of Minor Enterprisers Human Resources Survey data in 2001. The simulation result shows that more elaborate method is superior to the widely used method as expected but the difference is not quite significant.

• Communications for Statistical Applications and Methods
• /
• v.23 no.1
• /
• pp.47-55
• /
• 2016

Stratified cluster sampling has been widely used for effective parameter estimations due to reductions in time and cost. The probability proportional to size (PPS) sampling method is used when the number of cluster element are significantly different. However, simple random sampling (SRS) is commonly used for simplicity if the number of cluster elements are almost the same. Also it is known that the ratio estimator produces a good performance when the total number of population elements is known. However, the two stage cluster estimator should be used if the total number of elements in population is neither known nor accurate. In this study we suggest a composite estimator by combining the ratio estimator and the two stage cluster estimator to obtain a better estimate under a certain population circumstance. Simulation studies are conducted to compare the superiority of the suggested estimator with two other estimators.

• Proceedings of the Korea Concrete Institute Conference
• /
• 2006.05a
• /
• pp.582-585
• /
• 2006

Recently, many methods to evaluate the water content in concrete were developed. In this study, using a commercially avaliable equipment by the unit volume weight method, the water content in concrete was estimated when the moisture content in sand was changed. In order to evaluate the accuracy and the usefulness of the method, the designed water content and the measured water content were compared. The compressive strength of concrete with the variation of moisture content in sand was measured as well. As a result, it appears that in order to predict the water content in concrete, the adjustment coefficient of aggregate should be exactly estimated. It is shown that the equipment tends to underestimate the water content in concrete.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.15 no.5
• /
• pp.137-144
• /
• 2016

This paper aims to ensure describe the a spring-back prevention technique for improving shape fixability by using an ultra-high strength steel sheet with 980 MPa to develop a lightweight seat rail parts. Ultra-high strength steel gives a potential for considerable weight reduction and a cost-effective way to produce energy efficient vehicles. The influence of a spring-back of seat rail parts on the shape fixability in forming processes was investigated to be solved by an adjustment of the appropriate tool design and process parameters. The computed results for improving shape fixability were in good agreement with the experimental results.

• Proceedings of the IEEK Conference
• /
• 1999.06a
• /
• pp.1017-1020
• /
• 1999

Recently, a new learning algorithm for multilayer neural networks has been proposed 〔1〕. In the new learning algorithm, each output neuron is considered as a function of weights and the weights are adjusted so that the output neurons produce desired outputs. And the adjustment is accomplished by taking gradients. However, the gradient computation was performed numerically, resulting in a long computation time. In this paper, we derive the all necessary equations so that the gradient computation is performed analytically, resulting in a much faster learning time comparable to the backpropagation. Since the weight adjustments are accomplished by summing the gradients of the output neurons, we will call the new learning algorithm “multi-gradient.” Experiments show that the multi-gradient consistently outperforms the backpropagation.