Least Squares Percentage Regression
نویسنده
چکیده
Percentage error (relative to the observed value) is often felt to be more meaningful than the absolute error in isolation. The mean absolute percentage error (MAPE) is widely used in forecasting as a basis of comparison, and regression models can be fitted which minimize this criterion. Unfortunately, no formula exists for the coefficients, and models for a given data set may not be unique. We instead explore least squares regression based on the percentage error. We are able to derive exact expressions for the regression coefficients when the model is linear in these coefficients. Another advantage of this approach over MAPE is that the solution is unique. Furthermore, it has been shown that this approach provides strongly consistent coefficient estimates, and is superior to ordinary least squares when the data does not possess constant variance. For the practitioner we demonstrate that percentage regression models can easily be fitted using ordinary regression software, as well as by spreadsheets, using simple transformation of the data. Finally, we show that when the relative error is normally distributed, least squares percentage regression provides maximum likelihood estimates.
منابع مشابه
A robust least squares fuzzy regression model based on kernel function
In this paper, a new approach is presented to fit arobust fuzzy regression model based on some fuzzy quantities. Inthis approach, we first introduce a new distance between two fuzzynumbers using the kernel function, and then, based on the leastsquares method, the parameters of fuzzy regression model isestimated. The proposed approach has a suitable performance to<b...
متن کاملOn Mappings Between Electronic Noses
We consider the task of finding a mapping between two eNoses that employ two different sensor technologies, quartz microbalance and conducting polymers. Such a mapping is a model that predicts the response of one eNose based on the response of the other. eNose mappings are important for odor communication and synthesis, as well as for eNose data integration. We investigated a number of methods ...
متن کاملEvaluation of hybrid fuzzy regression capability based on comparison with other regression methods
In this paper, the difference between classical regression and fuzzy regression is discussed. In fuzzy regression, nonphase and fuzzy data can be used for modeling. While in classical regression only non-fuzzy data is used. The purpose of the study is to investigate the possibility of regression method, least squares regression based on regression and linear least squares linear regression met...
متن کاملDetermination of 137Ba Isotope Abundances in Water Samples by Inductively Coupled Plasma-optical Emission Spectrometry Combined with Least-squares Support Vector Machine Regression
A simple and rapid method for the determination of 137Ba isotope abundances in water samples by inductively coupled plasma-optical emission spectrometry (ICP-OES) coupled with least-squares support vector machine regression (LS-SVM) is reported. By evaluation of emission lines of barium, it was found that the emission line at 493.408 nm provides the best results for the determination...
متن کاملDetermination of Protein and Moisture in Fishmeal by Near-Infrared Reflectance Spectroscopy and Multivariate Regression Based on Partial Least Squares
The potential of Near Infrared Reflectance Spectroscopy (NIRS) as a fast method to predict the Crude Protein (CP) and Moisture (M) content in fishmeal by scanning spectra between 1000 and 2500 nm using multivariate regression technique based on Partial Least Squares (PLS) was evaluated. The coefficient of determination in calibration (R2C) and Standard Error of Calibra...
متن کامل