NEW ROBUST BASED RATIO TYPE ESTIMATORS OF POPULATION MEAN USING TRANSCENDENTAL FUNCTION

In practical situations, sometimes it may happen that the auxiliary variable does not exist or the auxiliary variable exists but we cannot get the auxiliary information due to limited budget. So there is a need to use the parameters of the study variable as supplementary information [Subramani (2016)]. In this connection the author suggested a ratio type estimator using median of study variable. Since median is least influenced by outliers, the present study proposes robust based exponential ratio type estimators of population mean on the basis of known median of the study variable when some observations in the data are outliers or out respondents. The efficiency of the proposed estimators has been compared both theoretically and empirically with the sample mean estimator, linear regression estimator of Watson (1937), ratio estimators of Cochran (1940), Reddy (1974), Bahl and Tuteja (1991) and Subramani (2016). The said comparison reveals that there is a significant gain in efficiency by the new proposed robust based exponential estimators.