• 전력전자학회:학술대회논문집
• /
• 전력전자학회 2004년도 전력전자학술대회 논문집(2)
• /
• pp.738-741
• /
• 2004

The overall performance of AC servo system is greatly affected by the uncertainties of unpredictable mechanical parameter variations and external load disturbances. Therefore, to compensate this problem, it is necessary to know different parameters and load disturbances subjected to position/speed control. This paper proposes an online identification method of mechanical parameters/load disturbances for AC servo system using Support Vector Regression (SVR). The proposed methodology advocates analytic parameter regression directly from the training data, rather than adaptive controller and observer approaches commonly used in motion control applications. The experimental results demonstrate that the proposed SVR algorithm is appropriate for control of unknown servo systems even with large measurement noise.

• 응용통계연구
• /
• 제26권3호
• /
• pp.511-520
• /
• 2013

이상치가 존재하는 회귀모형에서 이상치를 탐색하거나 로버스트 추정량에 대한 연구는 매우 중요하다. 이러한 연구는 leave-one-out를 이용하여 회귀계수를 추정하고 잔차를 이용하여 오차 분산을 추정하여 이상치를 탐색하는데 있다. 본 연구는 회귀모형에서 회귀계수를 추정하지 않고 오차 분산을 추정할 수 있는 Rice 추정량의 적용을 소개한 것이다. 특히, 단순회귀모형에서 이상치의 유무에 따라 Rice 추정량의 통계적 성질을 비교하고 이상치 탐색에 있어 어떤 장점이 있는지를 탐색한 연구이다.

• 한국정밀공학회지
• /
• 제22권2호
• /
• pp.70-78
• /
• 2005

In this paper, a novel rotor flux estimation method of an induction motor using support vector regression(SVR) is presented. Two well-known different flux models with respect to voltage and current are necessary to estimate the rotor flux of an induction motor. Training of SVR which the theory of the SVR algorithm leads to a quadratic programming(QP) problem. The proposed SVR rotor flux estimator guarantees the improvement of performance in the transient and steady state in spite of parameter variation circumstance. The validity and the usefulness of proposed algorithm are throughly verified through numerical simulation.

• Communications for Statistical Applications and Methods
• /
• 제25권5호
• /
• pp.523-544
• /
• 2018

Bivariate distributions play a fundamental role in survival and reliability studies. We consider a regression model for bivariate survival times under right-censored based on the bivariate Kumaraswamy Weibull (Cordeiro et al., Journal of the Franklin Institute, 347, 1399-1429, 2010) distribution to model the dependence of bivariate survival data. We describe some structural properties of the marginal distributions. The method of maximum likelihood and a Bayesian procedure are adopted to estimate the model parameters. We use diagnostic measures based on the local influence and Bayesian case influence diagnostics to detect influential observations in the new model. We also show that the estimates in the bivariate Kumaraswamy Weibull regression model are robust to deal with the presence of outliers in the data. In addition, we use some measures of goodness-of-fit to evaluate the bivariate Kumaraswamy Weibull regression model. The methodology is illustrated by means of a real lifetime data set for kidney patients.

• Communications for Statistical Applications and Methods
• /
• 제11권3호
• /
• pp.485-494
• /
• 2004

An algorithm is proposed to identify multiple outliers in linear regression. It is based on the clustering of residuals from the least median of squares estimation. A cut-height criterion for the hierarchical cluster tree is suggested, which yields the optimal clustering of the regression outliers. Comparisons of the effectiveness of the procedures are performed on the basis of the classic data and artificial data sets, and it is shown that the proposed algorithm is superior to the one that is based on the least squares estimation. In particular, the algorithm deals very well with the masking and swamping effects while the other does not.

• 대한전자공학회논문지SP
• /
• 제45권5호
• /
• pp.25-34
• /
• 2008

동일한 위치에서 같은 장면을 담고 있지만 서로 다른 시간에 획득된 두 장의 영상을 서로 비교하여 움직이는 자동차등에 의한 겹침과 같은 이상점의 집합을 검출할 수 있다. 영상들의 서로 다른 밝기 특성에 의한 영향을 줄이기 위하여 다항식 회귀 모델에 근거한 밝기 보정을 하였다. 이상점 집합으로 인한 영향을 약화시키면서 정확한 이상점 검출을 위하여 회귀분석을 단순히 반복하는 알고리듬을 도입하였다. 본 논문에서는 회귀분석을 반복하는 알고리듬의 성능을 잡음분산의 추정의 수렴 특성을 관찰하므로 분석하였으며, 교정 상수를 잡음분산 추정에 사용하여 강인한 검출이 가능하도록 하였다. 합성 영상과 실제 영상에 검출 알고리듬을 실험하여 그 강인성을 보였다.

• The Journal of Asian Finance, Economics and Business
• /
• 제9권7호
• /
• pp.1-15
• /
• 2022

Combining the strength of both Mixed Data Sampling (MIDAS) Regression and realized variance measures, this paper seeks to investigate two objectives: (1) evaluate the post-sample performance of the proposed weekly Realized Variance-MIDAS (RVar-MIDAS) in one-week ahead volatility forecasting against the established Generalized Autoregressive Conditional Heteroskedasticity (GARCH) model and the less explored but robust STES (Smooth Transition Exponential Smoothing) methods. (2) comparing forecast error performance between realized variance and squared residuals measures as a proxy for actual volatility. Data of seven private equity mutual fund indices (generated from 57 individual funds) from two different time periods (with and without financial crisis) are applied to 21 models. Robustness of the post-sample volatility forecasting of all models is validated by the Model Confidence Set (MCS) Procedures and revealed: (1) The weekly RVar-MIDAS model emerged as the best model, outperformed the robust DAILY-STES methods, and the weekly DAILY-GARCH models, particularly during a volatile period. (2) models with realized variance measured in estimation and as a proxy for actual volatility outperformed those using squared residual. This study contributes an empirical approach to one-week ahead volatility forecasting of mutual funds return, which is less explored in past literature on financial volatility forecasting compared to stocks volatility.

• International Journal of Computer Science & Network Security
• /
• 제22권7호
• /
• pp.301-307
• /
• 2022

Energy consumption has grown alongside dramatic population increases. Statistics show that buildings in particular utilize a significant amount of energy, worldwide. Because of this, building energy prediction is crucial to best optimize utilities' energy plans and also create a predictive model for consumers. To improve energy prediction performance, this paper proposes a ResNet-LSTM model that combines residual networks (ResNets) and long short-term memory (LSTM) for energy consumption prediction. ResNets are utilized to extract complex and rich features, while LSTM has the ability to learn temporal correlation; the dense layer is used as a regression to forecast energy consumption. To make our model more robust, we employed Huber loss during the optimization process. Huber loss obtains high efficiency by handling minor errors quadratically. It also takes the absolute error for large errors to increase robustness. This makes our model less sensitive to outlier data. Our proposed system was trained on historical data to forecast energy consumption for different time series. To evaluate our proposed model, we compared our model's performance with several popular machine learning and deep learning methods such as linear regression, neural networks, decision tree, and convolutional neural networks, etc. The results show that our proposed model predicted energy consumption most accurately.

• 응용통계연구
• /
• 제18권2호
• /
• pp.435-443
• /
• 2005

로지스틱 회귀모형에서 결정계수는 선형 회귀모형보다 다양하게 정의되며 그 값들도 매우 작아 로지스틱 회귀모형 평가기준으로 사용되는 통계량이 라고 할 수 없다. Liao와 McGee(2003)는 부적절한 설명변수의 추가 또는 표본크기의 변화에 민감하지 않은 두 종류의 수정 결정계수를 제안하였다. 본 연구에서는 실제자료에 적용한 로지스틱 회귀모형에서 수정 결정계수를 포함한 네 종류의 결정계수들을 변수선택의 기준으로 사용하여 기존의 변수선택 방법인 전진선택, 후진제거, 단계적 선택방법, AIC 통계량 등을 사용한 방법들과 비교하여 그 적절함과 효율성을 토론한다.

• Asian-Australasian Journal of Animal Sciences
• /
• 제8권5호
• /
• pp.505-513
• /
• 1995

The major factors affecting productivity on daily farms in Queensland, Australia, were determined using the stepwise linear regression approach. The data were obtained from a survey conducted on the total population of daily farms in Queensland in 1987. These data were divided into six major dailying regions. The technique was applied using 12 independent variables believed by a panel of experienced research and extension personnel to exert the most influence on milk production. The regression equations were all significant (p < 0.001) with the percentage coefficients of determination ranging from 62 to 76% for equations developed using' total farm milk: production as the dependent variable. Three of the variables affecting total farm milk: production were found to be common to all six regions. These were; the amount of supplementary energy fed, the area set aside to irrigate winter feed and the size of the area used for dailying. Higher production farms appeared to be more efficient in that they consistently produced milk production levels higher than those estimated from the regression equation for their region. Other methods of analysis including robust regression and non linear regression techniques were unsuccessful in overcoming this problem and allowing development of a model appropriate for farms at all levels of production.