A COMPARISON OF REGRESSION METHODS FOR IMPROVED ESTIMATION IN SAMPLING
Keywords:
nverse term model, linear regression estimator, Polynomial regression, simple random sampling.Abstract
A linear model with an inverse term is proposed for estimation of population mean and population total in regression analysis. A comparison has been made in precisions of estimates of parameters, considering ordinary linear regression estimate, regression estimates using second degree polynomial for relationship between dependent variable ( y )and auxiliary variable ( x ) and also with estimates obtained in regression method of estimation incorporating inverse term model for describing the relationship between y and x. The gain in efficiency is also demonstrated with the help of a data set in which comparison of various estimates of regression method of estimation has also been made with simple random sampling.
Downloads
References
Bates, D.M. and Watts, D.G. (1988). Non-linear Regression Analysis and Its
Applications. John Wiley, New York.
Cochran, W.G. (1999). Sampling Techniques. John Wiley & Sons.
Des Raj. (1972). The Design of Sample Surveys. McGraw-Hill, New York.
Draper, N.R. and Smith, H. (1998). Applied Regression Analysis. John Wiley,
New York.
