• Journal of the Korean Data and Information Science Society
• /
• 제13권2호
• /
• pp.65-75
• /
• 2002

The support vector machine (SVM), first developed by Vapnik and his group at AT &T Bell Laboratories, is being used as a new technique for regression and classification problems. In this paper we present an approach to estimating approximate prediction intervals for SVM regression based on posterior predictive densities. Furthermore, the method is illustrated with a data example.

• Journal of the Korean Statistical Society
• /
• 제34권1호
• /
• pp.73-83
• /
• 2005

Thompson's (1990) adaptive cluster sampling is a promising sampling technique to ensure effective representation of rare or localized population units in the sample. We consider the problem of simultaneous estimation of the numbers of earners through a number of rural unorganized industries of which some are concentrated in specific geographic locations and demonstrate how the performance of a conventional Rao-Hartley-Cochran (RHC, 1962) estimator can be improved upon by using auxiliary information in the form of generalized regression (greg) estimators and then how further improvements are also possible to achieve by adopting adaptive cluster sampling.

• Communications for Statistical Applications and Methods
• /
• 제12권2호
• /
• pp.413-423
• /
• 2005

Consider the problem of estimating the underlying regression function from a set of noisy data which is contaminated by a long tailed error distribution. There exist several robust smoothing techniques and these are turned out to be very useful to reduce the influence of outlying observations. However, no matter what kind of robust smoother we use, we should choose the smoothing parameter and relatively less attention has been made for the robust bandwidth selection method. In this paper, we adopt the idea of robust location parameter estimation technique and propose the robust cross validation score functions.

• 한국데이터정보과학회:학술대회논문집
• /
• 한국데이터정보과학회 2004년도 추계학술대회
• /
• pp.53-59
• /
• 2004

Support vector machine(SVM) approach to regression can be found in information science literature. SVM implements the regularization technique which has been introduced as a way of controlling the smoothness properties of regression function. In this paper, we propose a new estimation method based on quadratic loss SVM for a linear fuzzy regression model of Tanaka's, and furthermore propose a estimation method for nonlinear fuzzy regression. This approach is a very attractive approach to evaluate nonlinear fuzzy model with crisp input and output data.

• 대한기계학회:학술대회논문집
• /
• 대한기계학회 2003년도 추계학술대회
• /
• pp.1822-1827
• /
• 2003

End-milling have been used widely in industrial system because it is effective to a material manufacturing with various shape. Recently the end-milling processing is needed the high-precise technique with good surface roughness and rapid time in precision machine part and electronic part. The optimum surface roughness has an effect on end-milling condition such as, cutting direction, spindle speed, feed rate and depth of cut, etc. Therefore this study was carried out to presume for mutual relation of end-milling condition to get the optimum surface roughness by regression analysis. The results shown that coefficient of determination($R^{2}m$) of regression equation has a fine reliability over 80% and regression equation of surface rough is made by regression analysis.

• 한국정보시스템학회지:정보시스템연구
• /
• 제4권
• /
• pp.27-39
• /
• 1995

This study tests the fuzzy linear regression model to see if there is a performance difference between it and the classical linear regression model. These results show that FLR was better as f forecasting technique when compared with CLR. Another important find in the test of the two different regression methods is that they generate two different predicted P/E ratios from expected value test, variance test and error test of two different regressions, though we can not see a significant difference between two regression models doing test in error measurements (GMRAE, MAPE, MSE, MAD). So, in this financial setting we can conclude that FLR is not superior to CLR, comparing and testing between the t재 different regression models. However, FLR is better than CLR in the error measurements.

• 한국공작기계학회논문집
• /
• 제12권5호
• /
• pp.46-52
• /
• 2003

End-milling have been used widely in industrial system because it is effective to a material manufacturing with various shape. Recently the end-milling processing is needed the high-precise technique with good surface roughness and rapid time in precision machine part and electronic part. The optimum surface roughness has an effect on end-milling condition such as, cutting direction spindle speed, feed rate and depth of cut, etc. Therefore this study was carried out to presume for mutual relation of end-milling condition to get the optimum surface roughness by regression analysis. The results shown that coefficient of determination($\textrm{R}^2$) of regression equation has a fine reliability of 87.5% and regression equation of surface rough is made by regression analysis.

• 한국추진공학회:학술대회논문집
• /
• 한국추진공학회 2004년도 제22회 춘계학술대회논문집
• /
• pp.524-527
• /
• 2004

Hybrid rocket had many advantage with compared to solid and liquid rockets. However, the engines have not yet been used in practical rocket systems, due mainly to the disadvantage of hybrid combustion, such as low fuel regression rate. In this study, lab-scale hybrid motor was designed and manufactured. And the methods of regression rate improvement were considered. Test firings with thrusts up to 300 N were conducted with GOX and transparent PMMA. Thrust was calculated with the pressure of the combustion chamber and the regression rate was measured in with variation of oxidizer flow rate. The regression rates showed a strong dependency on GOX mass flux. The frequency analysis technique of the bulk-mode oscillation of motor was applied to a hybrid rocket motor and was based on the principle that this frequency was inversely proportional to the square root of the chamber volume. Several problems and solutions of operating hybrid rocket were presented.

• 한국세라믹학회지
• /
• 제36권8호
• /
• pp.813-819
• /
• 1999

Ion implantation, photo-lithography, Ar ion milling과 hot press 법을 이용한 micro-fabrication techrique을 사용하여 고순도 알루미나 단결정인 사파이어의 내부에, 조절된 Ca의 첨가량을 갖고 있는, crack과 비슷한 형태의 기공들을 형성시켰다. 이 bi-cryslal을 각각의 온도에서 열처리하여 Ca 이온이 고온에서 알루미나의 morphology와 hcaling에 미치는 영향을 관찰하였다. 열처리 온도가 올라감에 따라서 crack-like pore의 내부에 hcxagonal bridging ligaments가 생성되었는데, 열처리 온도와 Ca의 첨가량이 증가할수록 크기가 커지는 것을 관찰할 수 있었고, 생성된 hexagonal bndgmg ligaments는 열처리가 진행됨에 따라 서서히 커지면서 모서리가 둥글어지는 현상을 관찰할 수 있었다. Bicrystal 내부에 형성된 crack-like pore는 열처리가 진행되면서 edge regression. ligamcnt growth 및 flow의 3가지의 특징적인 형태로 진행되었다. 이때 edge regression은 상대적으로 저온에서부터 전체 crack-like pore에서 서서히 일어나기 시작하였으며, ligament growth는 일부 crack-like pore에서 진행되있으며, 대단히 빠른 속도로 crack healing이 진행됨을 추정할 수 있었다. Flow는 $1800^{\circ}C$ 이상의 고온에서 모든 crack-like pore에 걸쳐서 느리게 일어남을 알 수 있었다.

• 한국전산유체공학회:학술대회논문집
• /
• 한국전산유체공학회 2000년도 춘계 학술대회논문집
• /
• pp.159-165
• /
• 2000

In this paper, numerical study has been done for the improvement of the superdetonative ram accelerator performance and for the design optimization of the system. The objective function to optimize the premixture composition is the ram tube length required to accelerate projectile from initial velocity $V_o$ to target velocity $V_e$. The premixture is composed of $H_2,\;O_2,\;N_2$ and the mole numbers of these species are selected at design variables. RSM(Response Surface Methodology) which is widely used for the complex optimization problems is selected as the optimization technique. In particular, to improve the non-linearity of the response and to consider the accuracy and efficiency of the solution, design space stretching technique has been applied. Separate sub-optimization routine is introduced to determine the stretching position and clustering parameters which construct the optimum regression model. Two step optimization technique has been applied to obtain the optimal system. With the application of stretching technique, we can perform system optimization with a small number of experimental points, and construct precise regression model for highly non-linear domain. The error to compared with analysis result is only $0.01\%$ and it is demonstrated that present method can be applied more practical design optimization problems with many design variables.