• 한국시뮬레이션학회논문지
• /
• 제9권3호
• /
• pp.43-51
• /
• 2000

In regression model, we estimate the unknown parameters by using various methods. There are the least squares method which is the most general, the least absolute deviation method, the regression quantile method and the asymmetric least squares method. In this paper, we will compare each others with two cases: firstly the theoretical comparison in the asymptotic sense and then the practical comparison using Monte Carlo simulation for a small sample size.

• 복식문화연구
• /
• 제5권4호
• /
• pp.130-158
• /
• 1997

This study outputs calculation of regression of each items for production of torso basic pattern according to 6 body types as the result of another study and intends to present drawing method of torso model by short measure method modified and supplied and supplied by experiments of wearing clothing. SAS(Statistical Analysis System) is used for figures management and methods for analysis used are Frequency Analysis, Means Analysis, Regression Analysis, Correlation Analysis, etc. Results are as follows. 1. Correlation analysis is used to output the size necessary for torso prototype drawing by sort measure method and waist front length, back length, crotch length, shoulder point-cerricale-shoulder point, bust circumference, waist circumference, weight, etc, are set up as representative items calculation of regression of each type is suggested. 2. In the result of experiment of the first wearing clothing intended for 5 in each type and the whole 30, to develop torso prototype drawing method by short measure method, as we find some problems of the shape and propriety of neck root circumference line, the position of shoulder point, pulling or hold armpit parts, waist circumference line, the degree of dissatisfaction is high, so the second experiment of wearing clothing is propriety of each part is improved, all items except the length and quantity of shoulder dart, waist in back bodice, clearance quantity of hip circumference, and the place of shoulder line in side bodice. So, it was modifed and supplied and then the third torso prototyped drawing method by shout measure method was suggested. The third prototype drawing method was suggested, by modifying and supplying.

• 한국지반공학회논문집
• /
• 제38권12호
• /
• pp.103-112
• /
• 2022

본 연구에서는 시간-침하량 계측 데이터를 기반으로 한 기존 침하 예측 이론식을 확인하였다. 기존 계측 기반 침하 예측 이론식 중 쌍곡선법 및 Asaoka법이 정확도가 높게 나타났으며, 이외 방법은 정확도가 낮은 것으로 확인되었다. 이러한 분석 결과를 토대로 기존 침하 예측 방법의 한계점을 도출하였으며, 이러한 한계점을 보완할 수 있는 개선방안으로써 가중 비선형 회귀분석을 통한 침하 예측 방법을 제시하였다.

• Journal of Information Processing Systems
• /
• 제18권2호
• /
• pp.268-281
• /
• 2022

The development of information technology is bringing many changes to everyday life, and machine learning can be used as a technique to solve a wide range of real-world problems. Analysis and utilization of data are essential processes in applying machine learning to real-world problems. As a method of processing data in machine learning, we propose an approach based on applying multiple linear regression models by interlacing data to the task of classifying similar software. Linear regression is widely used in estimation problems to model the relationship between input and output data. In our approach, multiple linear regression models are generated by training on interlaced feature data. A combination of these multiple models is then used as the prediction model for classifying similar software. Experiments are performed to evaluate the proposed approach as compared to conventional linear regression, and the experimental results show that the proposed method classifies similar software more accurately than the conventional model. We anticipate the proposed approach to be applied to various kinds of classification problems to improve the accuracy of conventional linear regression.

• Journal of the Korean Statistical Society
• /
• 제26권4호
• /
• pp.493-505
• /
• 1997

The analysis of binary data appears to many areas such as statistics, biometrics and econometrics. In many cases, data are often collected in which some observations are incomplete. Assume that the missing covariates are missing at random and the responses are completely observed. A method to Bayesian analysis of the binary regression model with incomplete data is presented. In particular, the desired marginal posterior moments of regression parameter are obtained using Meterpolis algorithm (Metropolis et al. 1953) within Gibbs sampler (Gelfand and Smith, 1990). Also, we compare logit model with probit model using Bayes factor which is approximated by importance sampling method. One example is presented.

• Archives of Pharmacal Research
• /
• 제11권2호
• /
• pp.99-113
• /
• 1988

Quantitation of ethenzamide, isopropylantipyrine and caffeine takes about 41 hrs by conventional GC method. Quantitation of allylisoprorylacetylurea takes about 40 hrs by conventional UV method. But quantitation of them takes about 6 hrs by DRIFT developing method. Each standard and sample sieved, powdered and acquired DRIFT spectrum. Out of them peak of each component was selected and ratio of each peak to standard peak was acquired, and then linear stepwise multiple regression was performed with these data and concentration. Reflectance value, Kubelka-Munk equation and Inverse-Kubelka-Munk equation were modified by us. Inverse-Kubelka-Munk equation completed the deficit of Kubelka-Munk equation. Correlation coefficients acquired by conventioanl GC and UV against DRIFT were more than 0.95.

• Communications for Statistical Applications and Methods
• /
• 제19권6호
• /
• pp.885-898
• /
• 2012

In this paper, we consider Bayesian approaches to partially linear models, in which a regression function is represented by a semiparametric additive form of a parametric linear regression function and a nonparametric regression function. We make a comparative study on the performance of widely used Bayesian partially linear models in terms of empirical analysis. Specifically, we deal with three Bayesian methods to estimate the nonparametric regression function, one method using Fourier series representation, the other method based on Gaussian process regression approach, and the third method based on the smoothness of the function and differencing. We compare the numerical performance of three methods by the root mean squared error(RMSE). For empirical analysis, we consider synthetic data with simulation studies and real data application by fitting each of them with three Bayesian methods and comparing the RMSEs.

• 한국지능시스템학회논문지
• /
• 제25권5호
• /
• pp.477-482
• /
• 2015

단기풍속 예측을 위한 진화적 선형 및 비선형 회귀분석 기반의 보정 기법을 비교한다. 모델의 체계적 오류를 교정하기 위한 효율적인 MOS(Model Output Statistics)의 개발이 필요하나, 기존의 선형회귀분석 기반의 보정기법은 다양한 기상요소의 복잡한 비선형 특성을 반영하기 힘들다. 이를 개선하기 위해서 유전 프로그래밍을 사용하여 풍속 예측에 대한 비선형 보정 수식을 생성하는 기법을 제안하고 기본 다중선형회귀분석법 및 Ridge, Lasso 회귀분석법과 비교한다. 더불어, 선형회귀분석법과 진화적 비선형회귀분석 기법의 인자 선택의 차이와 유사성을 비교하고 분석한다. 2007년~2013년의 KLAPS(Korea Local Analysis and Prediction System) 재분석자료를 사용하여 제주도와 부산지역의 격자점에 대한 실험을 수행한다.

• KIEAE Journal
• /
• 제15권5호
• /
• pp.13-20
• /
• 2015

Purpose: In order to upgrade the energy performance of existing building, energy audit stage should be implemented first because it is useful method to find where the problems occur and know how much time and cost consumption for retrofit. In overseas researches, three levels of audit is proposed whereas there are no standards for audit in Korea. Besides, most studies use dynamic simulation in detail like audit level 3 even though the level 2 can save time and cost than level 3. Thus, this paper focused on audit level 2 and proposed the audit method with the simple linear regression analysis model. Method: Two parameters were considered for the simple regression analysis, which were the monthly electric use and the mean outdoor temperature data. The former is a dependent variable and the latter is a independent variable, and the building's energy performance profile was estimated from the regression analysis method. In this analysis, we found the abnormal point in cooling season and the more detailed analysis were conducted about the three heat source equipments. Result: Comparing with real and predicted models, the total consumption of predicted model was higher than real value as 23,608 kWh but it was the results that was reflected the compulsory control in 2013. Consequently, it was analyzed that the revised model could save the cooling energy as well as reduce peak electric use than before.

• 설비공학논문집
• /
• 제12권12호
• /
• pp.1090-1097
• /
• 2000

Predicting building energy use can be useful to evaluate its energy performance. This study proposed empirical approach for predicting building energy use with regression analysis. For the empirical analysis, simple regression models were developed based on the historical energy consumption data as a function of daily outside temperature, the predicting equations were derived for different operational modes and day types, then the equations were applied for predicting energy use in a building. BY selecting a real building as a case study, the feasibilities of the empirical approach for predicting building energy use were examined. The results showed that empirical approach with regression analysis was fairly reliable by demonstrating prediction accuracy of $pm10%$ compared with the actual energy consumption data. It was also verified that the prediction by regression models could be simple and fairly accurate. Thus, it is anticipated that the empirical approach will be useful and reliable tool for many purposes: retrofit savings analysis by estimating energy usage in an existing building or the diagnosis of the building operational problems with real time analysis.