Modified robust ridge estimates to determine outliers
Journal Title: Вісник Національного технічного університету "ХПІ". Серія: Проблеми удосконалювання електричних машин і апаратів. Теорія і практика - Year 2015, Vol 13, Issue 1122
Abstract
Least squares estimates (LSE), in a multiple linear regression when the predictors are strongly correlated, give low prediction accuracy. Ridge regression, being regularized version of the regression based on least squares based on the minimization of a quadratic loss function, sensitive to outliers. Smoothly redescending ψ-functions, which lead to asymptotically efficient estimates are considered. The Iteratively Re-weighted Least Squares (IRLS) method based on the considered ψ-functions is used to obtain the resulting robust ridge estimates to detect outliers. Simulation results confirm the received theoretical conclusions. The convergence is obtained to the final estimates of the coefficients with smaller number of iterations than without using ridge regression. The combined robust ridge estimates allow to obtain stable coefficients and residuals which help in determining true coefficients and outliers.
Authors and Affiliations
V. I. Gritsyuk
Keywords
Related Articles
- EP ID EP408784
- DOI -
- Views 111
- Downloads 0
Development latch short-circuit for overhead power line voltage of 110 - 150 kV
Proposed a simple design of electromechanical latch short-circuit for overhead power line voltage of 110-150 kV. For the selected constructions shows a method of calculating its technical characteristics. Experimentally...
Hydrogen cooling of high power turbogenerators
The use of hydrogen for cooling of high power machines has been substantiated. The relation for heat transfer coefficient between hydrogen and inner walls of the cooling channels in the conductors of the rotor winding th...
Application eddy-current frequency balance for monitoring of material properties by unprofitable converters
It is shown that electromagnetic converters based on the use of eddy-current frequency balance, allow almost completely to get rid of the influence (effect) of air gaps between the converter and a controllable detail. Th...
Modified robust ridge estimates to determine outliers
Least squares estimates (LSE), in a multiple linear regression when the predictors are strongly correlated, give low prediction accuracy. Ridge regression, being regularized version of the regression based on least squar...
Investigation of electromagnetic processes in a semiconductor converter with seven-zone regulation of output voltage and electromechanical load
New mathematical models of the electromagnetic processes in three-phase electric circuits of the semiconductor converters with many-server zoned regulation of phases voltages disregarding losses in semiconductor commutat...