International Journal of Statistics and Applied Mathematics 2017; 2(6): 172-186
ISSN: 2456-1452 Maths 2017; 2(6): 172-186
© 2017 Stats & Maths www.mathsjournal.com Received: 12-09-2017 Accepted: 14-10-2017 Arpita Lakhre
School of mathematics, Statistics and Computational Science, Central University of Rajasthan, Ajmer, Rajasthan, India Rajesh Tailor
S.S in statistics, Vikram University, Ujjain, Madhya Pradesh, India
Correspondence Arpita Lakhre
School of mathematics, Statistics and Computational Science, Central University of Rajasthan, Ajmer, Rajasthan, India
Improved ratio-cum-product estimators of finite population mean using known parameters of two
auxiliary variates in double sampling
Arpita Lakhre and Rajesh Tailor
Abstract
Use of auxiliary information has been in practice to improve the efficiency of the estimators of parameters. Ratio, product and regression methods are good examples of use of auxiliary information.
Ratio, product and regression type estimators essentially require the knowledge of population mean of auxiliary variates. But many times, the information on population mean of the auxiliary variate is not available. In this type of situations, double sampling is used. Ajagaonkar (1975) and Sisodia and Dwivedi (1982) discussed problem of estimation using single auxiliary variate whereas Khan and Tripathi (1967), Rao (1975) and Singh and Namjoshi (1988) considered the use of multi auxiliary variates in double sampling.
Singh (1967) used information on two auxiliary variates and envisaged a ratio-cum-product estimator of finite population mean of the study variate assuming that the population mean of the auxiliary variates are known. Upadhyaya and Singh (1999) proposed some ratio type estimators using coefficient of variation and coefficient of kurtosis. Tailor et al. (2011) suggested ratio-cum-product estimators using coefficient variation and coefficient of kurtosis of two auxiliary variates in simple random sampling. In this paper, authors study the Tailor et al. (2011) ratio-cum-product estimators in double sampling.
Keywords: Ratio-cum-product estimator, double sampling, population mean, Bias, Mean squared error
Introduction
This paper considers the problem of estimation of finite population mean in double sampling.
In this paper, two ratio-cum-product estimators of finite population mean, using known coefficient of variation and coefficient of kurtosis of two auxiliary variates, have been suggested. Suggested estimators have been compared with usual unbiased estimator, classical ratio and product estimators in double sampling and double sampling versions of Singh (1967)
[3] and Upadhyaya and Singh (1999) estimators. To judge the performance of the suggested estimators over other considered estimators an empirical study also has been carried out.
Let us consider a finite population
U U
1, U
2,..., U
N
of sizeN
. Lety
,x
1andx
2 be the study variate and auxiliary variates taking valuesy
ix
1i andx
2i respectively on) ,..., 2 , 1
( i N
U
i
Let the auxiliary variatesx
1 andx
2 be positively and negatively correlated with the study variate y respectively.Let us define
Ni
y
iY N
1
1
: Population mean of the study variatey
,
Ni
x
iX N
1 1 1
1
: Population mean of the auxiliary variatex
1,
Ni
x
iX N
1 2 2
1
: Population mean of the auxiliary variatex
2.Let
ni i
n y y
1
/
,
ni i
n x x
1 1
1
/
and
ni
i
n x x
1 2
2
/
be the unbiased estimators of population meanY
,X
1 andX
2 respectively.
ni
x
ix n
1 1 /
1
1
: First phase sample mean of the auxiliary variatex
1 based on sample sizen
/,
ni
x
ix n
1 2 /
2
1
: First phase sample mean of the auxiliary variatex
2 based on sample sizen
/,
ni
y
iy n
1
1
: Second phase sample mean of the study variatey
based on sample sizen
,
ni
x
ix n
1 1 1
1
: Second phase sample mean of the auxiliary variatex
1 based on sample sizen
,
ni
x
ix n
1 2 2
1
: Second phase sample mean of the auxiliary variatex
2 based on sample sizen
,
Ni i
y
y Y
S N
1
2
2
( )
1
1
: Population mean square of the study variatey
,
Ni i
x
x X
S N
1
2 1 1
2
( )
1 1
1 : Population mean square of the auxiliary variate
x
1,
Ni i
x
x X
S N
1
2 2 2
2
( )
1 1
21 : Population mean square of the auxiliary
variate
x
2,
Ni
i i
yx
y Y x X
S N
1
1
1
)
)(
1 ( 1
1 : Population covariance between the study variate
y
and auxiliary variatex
1
Ni
i i
yx
y Y x X
S N
1
2
2
)
)(
1 ( 1
2 : Population covariance between the study variate
y
and auxiliary variatex
2,
Ni
i i
x
x
x X x X
S N
1
2 2 1
1
)( )
1 ( 1
2
1 : Population covariance between the auxiliary variate
x
1 andx
2,2 2
1 1 1
x y
yx
yx
S S
S
: Population correlation coefficient between the study variatey
and auxiliary variatex
1,2 2
2 2 2
x y yx
yx
S S
S
: Population correlation coefficient between the study variate
y
and auxiliary variatex
2,2 2
2 1
2 1 2
1
x x
x x x
x
S S
S
: Population correlation coefficient between the auxiliary variate
x
1 and auxiliary variatex
2,Y C
y S
y: Population coefficient of variation of the study variate
y
,1
1
1
X
C
x S
x: Population coefficient of variation of the auxiliary variate
x
1,2
2
2
X
C
x S
x: Population coefficient of variation of the auxiliary variate
x
2,
21 1
4 1 1 1
2
( )
) ) (
( X X
X x X
i
i : Coefficient of kurtosis of the auxiliary variatex
1,
22 2
4 2 2 2
2
( )
) ) (
( X X
X x X
i
i : Coefficient of kurtosis of the auxiliary variatex
2,Cochran (1940) [2] envisaged classical ratio estimator for estimating the population mean
Y
when study variatey
and auxiliary variatex
1 are positively correlated as
1
ˆ
1x y X
Y
R (1.1)In case of negative correlation between the study variate
y
and the auxiliary variatex
2, the classical product estimator was given by Robson (1957) [6] as
2
ˆ
2X y x
Y
P (1.2)Assuming that the population mean
X
1 and coefficient of variationx1
C
of the auxiliary variatex
1 are known, Sisodia and Dwivedi (1981) defined a ratio type estimator as
1 1
1
ˆ
1x x
SDR
x C
C y X
Y
, (1.3)When the correlation coefficient between the study variate
y
and auxiliary variatex
2 is negative product type estimator using coefficient of variationx2
C
is expressed as
2 2
2
ˆ
2x x
SDP
X C
C y x
Y
. (1.4)Singh et al. (2004) [10] defined ratio and product type estimators using coefficient of kurtosis
2( x
1)
and
2( x
2)
respectively as
) (
) ˆ (
1 2 1
1 2 1
x x
x y X
Y
SER
, (1.5)and
) (
) ˆ (
2 2 2
2 2 2
x X
x y x
Y
SEP
. (1.6)Upadhyaya and Singh (1999) utilized both coefficient of kurtosis as well as coefficient of variations of auxiliary variates and suggested two ratio and two product type estimators of population mean
Y
as
) (
) ˆ (
1 2 1
1 2 1
1
1 1
x C
x
x C
y X Y
x x R
US
, (1.7)
1 1
) (
) ˆ (
1 2 1
1 2 1 2
x x R
US
x x C
C x y X
Y
(1.8)
) (
) ˆ (
2 2 2
2 2 2
1
2 2
x C
X
x C
y x Y
x x R
US
.(1.9)
2 2
) (
) ˆ (
2 2 2
2 2 2 2
x x P
US
X x C
C x y x
Y
(1.10)Singh (1967) [3] utilized information on known population means
X
1 andX
2 of auxiliary variatesx
1 andx
2 respectively and envisaged a ratio-cum-product estimator of population meanY
as
2 2
1
ˆ
1X x x y X
Y
SRP (1.11)The problem of estimating the population mean
Y
of y when the population meansX
1 andX
2 ofx
1 andx
2 respectively are known, has been discussed by many researchers including Singh and Tailor (2005) [9],Tailor and Tailor (2008) [12], Tailor et al.(2011a) [13], Tailor (2012) [11] and many others. When information is not available on
X
1 andX
2 in advance, double sampling procedure is used. The standard double sampling procedure is described as
(i) a first phase sample
S
1 of fixed size n' is drawn form U to observe onlyx
1 andx
2 to estimateX
1 andX
2 respectively then (ii) a second phase sampleS
2 of fixed size n is drawn either fromS
1 from first phase sample or directly from the population.These two cases may be designated as
Case I: As a sub sample from the first phase sample and Case II: Drawn independently to the first phase sample.
In double sampling, the usual ratio and product estimators of population mean
Y
are respectively defined as
1 ) 1
ˆ
(x y x Y
Rd, (1.12)
and
'2 ) 2
ˆ
(x y x Y
Pd, (1.13)
where
y
,x
1 andx
2 are sample means based on second phase sample of size n whereas
/
1 1 1
1
ni
x
ix n
and
ni
x
ix n
1 2 2
1
are the first phase sample means of
x
1 andx
2, which are unbiased estimates of population meansX
1 andX
2 respectively.In double sampling, Sisodia and Dwivedi (1981) ratio type and Pandey and Dubey (1988) [4] product type estimators of population mean
Y
are defined as
1 1
1 ) 1
ˆ
(x d x
SDR
x C
C y x
Y
, (1.14)and
2 2
2 ) 2
ˆ
(x d x
SDP
x C
C y x
Y
. (1.15)In double sampling, Singh et al. (2004) [10] ratio and product type estimators of population mean
Y
are defined as
( )
) ˆ (
1 2 1
1 2 ) 1
(
x x
x y x
Y
SERd
, (1.16)and
) (
) ˆ (
2 2 2
2 2 ) 2
(
x x
x y x
Y
SEPd
. (1.17)Double sampling versions of Upadhyaya and Singh (1999) ratio type estimators are
) (
) ˆ (
1 2 1
1 2 /
) 1 (
1
1 1
x C
x
x C
y x Y
x d x
R
US
, (1.18)and
1 1
) (
) ˆ (
1 2 1
1 2
| / ) 1
( 2
x d x
R
US
x x C
C x y x
Y
. (1.19)Double sampling version of Upadhyaya and Singh (1999) product type estimators are define as
) (
) ˆ (
2 2 2
2 2 /
) 2 (
1
2 2
x C
x
x C
y x Y
x d x
P
US
, (1.20)
and
2 2
) (
) ˆ (
2 2 2
2 2 / ) 2
( 2
x d x
P
US
x x C
C x y x
Y
. (1.21)In double sampling, Singh (1967) [3] ratio-cum-product estimator
Yˆ
RP of population meanY
is expressed as
'2 2
1 ' ) 1
ˆ
(x x x y x
Y
RPd . (1.22)
The biases and mean squared errors
ˆ
(d)Y
R ,ˆ
(d)Y
P ,ˆ
(d)Y
SDR,ˆ
(d)Y
SDP,ˆ
(d)Y
SER,ˆ
(d)Y
SEP,ˆ
(d1) RY
US ,ˆ
(d2) RY
US ,ˆ
(d1) PY
US ,ˆ
(d2) PY
US andˆ
(d)Y
RP in both case I and case II are given below:), 1 ( ˆ )
(
( ) 3 2 011
K
C f Y Y
B
Rd I
x
(1.23)), 1 ( ˆ )
(
( ) 3 2 022
K
C f Y Y
B
Pd I
x
(1.24)), (
ˆ )
(
( ) 3 1 2 1 011
t K
C t f Y Y
B
SDRd I
x
(1.25)), (
ˆ )
(
( ) 3 2 2 2 022
t K
C t f Y Y
B
SDPd I
x
(1.26)), (
ˆ )
(
( ) 3 3 2 3 011
t K
C t f Y Y
B
SERd I
x
(1.27)), (
ˆ )
(
( ) 3 4 2 4 022
t K
C t f Y Y
B
SEPd I
x
(1.28)), (
ˆ )
(
(1) 3 5 2 5 011
t K
C t f Y Y
B
USdR I
x
(1.29)), (
ˆ )
(
(1) 3 6 2 6 022
t K
C t f Y Y
B
USdP I
x
(1.30)), (
ˆ )
(
(2) 3 7 2 7 011
t K
C t f Y Y
B
USdR I
x
(1.31)), (
ˆ )
(
(2) 3 8 2 8 022
t K
C t f Y Y
B
USdP I
x
(1.32) ( 1 ) ( ) ,
ˆ )
(
( ) 3 2 01 2 02 122
1
K C K K
C f Y Y
B
RPd I
x
x
(1.33)), 1 ( ˆ )
(
( ) 1 2 011
K
C f Y Y
B
Rd II
x
(1.34)), (
ˆ )
(
( ) 2 2 1 022
f f K
C Y Y
B
Pd II
x
(1.35)), (
ˆ )
(
( ) 1 1 2 1 011
t K
C t f Y Y
B
SDRd II
x
(1.36)), (
ˆ )
(
( ) 2 2 2 2 1 022
f t f K
C t Y Y
B
SDPd II
x
(1.37)), (
ˆ )
(
( ) 1 3 2 3 011
t K
C t f Y Y
B
SERd II
x
(1.38)), (
ˆ )
(
( ) 4 2 24 1 022
f t f K
C t Y Y
B
SEPd II
x
(1.39)), (
ˆ )
(
(1) 1 5 2 5 011
t K
C t f Y Y
B
USdR II
x
(1.40)), (
ˆ )
(
(1) 1 6 2 2 6 1 022
f t f K
C t f Y Y
B
USdP II
x
(1.41)), (
ˆ )
(
(2) 3 7 2 7 011
t K
C t f Y Y
B
USdR II
x
(1.42)), (
ˆ )
(
(2) 1 8 2 28 1 022
f t f K
C t f Y Y
B
USdP II
x
(1.43) ( 1 ) ( ) ,
ˆ )
(
( ) 1 2 01 2 2 12 1 022
1
K C f K f K
C f Y Y
B
RPd II
x
x
(1.44) ( 1 2 )
ˆ )
(
( ) 2 1 2 3 2 011
K
C f C f Y Y
MSE
Rd I
y
x
, (1.45) ( 1 2 )
ˆ )
(
( ) 2 1 2 3 2 022
K
C f C f Y Y
MSE
Pd I
y
x
, (1.46) ( 2 )
ˆ )
(
( ) 2 1 2 3 1 2 1 011
t K
C t f C f Y Y
MSE
SDRd I
y
x
, (1.47)
( 2 )
ˆ )
(
( ) 2 1 2 3 2 2 2 022
t K
C t f C f Y Y
MSE
SDPd I
y
x
, (1.48) ( 2 )
ˆ )
(
( ) 2 1 2 3 3 2 3 011
t K
C t f C f Y Y
MSE
SERd I
y
x
, (1.49) ( 2 )
ˆ )
(
( ) 2 1 2 3 4 2 4 022
t K
C t f C f Y Y
MSE
SEPd I
y
x
, (1.50) ( 2 )
ˆ )
(
(1) 2 1 2 3 5 2 5 011
t K
C t f C f Y Y
MSE
USdR I
y
x
, (1.51) ( 2 )
ˆ )
(
(1) 2 1 2 3 5 2 6 012
t K
C t f C f Y Y
MSE
USdP I
y
x
, (1.52) ( 2 )
ˆ )
(
(2) 2 1 2 3 7 2 7 011
t K
C t f C f Y Y
MSE
USdR I
y
x
, (1.53) ( 2 )
ˆ )
(
(2) 2 1 2 3 8 2 8 012
t K
C t f C f Y Y
MSE
USdP I
y
x
, (1.54) ( 1 2 ) ( 1 2 2 )
ˆ )
(
( ) 2 1 2 3 2 01 3 2 02 122
1
K f C K K
C f C f Y Y
MSE
SRPd I
y
x
x
, (1.55)
1 2 1 01
2 2 1 2 )
(
) ( ) 2
( ˆ
1
f f f K
C C f Y Y
MSE
Rd II
y
x
, (1.56)
1 2 1 02
2 2 1 2 )
(
) ( ) 2
( ˆ
2
f f f K
C C f Y Y
MSE
Pd II
y
x
, (1.57)
1 1 2 1 01
2 1 2 1 2 )
(
) ( ) 2
( ˆ
1
t f f f K
C t C f Y Y
MSE
SDRd II
y
x
, (1.58)
2 1 2 1 02
2 2 2 1 2 )
(
) ( ) 2
( ˆ
2
t f f f K
C t C f Y Y
MSE
SDPd II
y
x
, (1.59)
3 1 2 1 01
2 3 2 1 2 )
(
) ( ) 2
( ˆ
1
t f f f K
C t C f Y Y
MSE
SERd II
y
x
, (1.60)
4 1 2 1 02
2 4 2 1 2 )
(
) ( ) 2
( ˆ
2
t f f f K
C t C f Y Y
MSE
SEPd II
y
x
, (1.61)
5 1 2 1 01
2 5 2 1 2 )
(
1
) ( ) 2
( ˆ
1
t f f f K
C t C f Y Y
MSE
USdR II
y
x
, (1.62)
6 1 2 1 02
2 6 2 1 2 )
(
1
) ( ) 2
( ˆ
2
t f f f K
C t C f Y Y
MSE
USdP II
y
x
, (1.63)
7 1 2 1 01
2 7 2 1 2 )
(
2
) ( ) 2
( ˆ
1
t f f f K
C t C f Y Y
MSE
USdR II
y
x
, (1.64)
8 1 2 1 02
2 8 2 1 2 )
(
2
) ( ) 2
( ˆ
2
t f f f K
C t C f Y Y
MSE
USdP II
y
x
, (1.65)and
( ˆ ) 2 ) (
1 22
1 022
2 12) .
2 01 1 2 1 2 2 1 2 )
(
2
1
f f f K C f f f K f K
C C f Y Y
MSE
SRPd II
y
x
x
(1.66)where
)
(
1 11 1
C
xX t X
,)
(
2 22 2
C
xX t X
,)) ( (
1 2 11
3
X x
t X
,)) (
(
2 2 22
4
X x
t X
,)) (
(
1 2 11 5
1 1
x C
X C t X
x x
,)) (
(
2 2 22 6
2 2
x C
X C t X
x x
,) ) ( (
) (
1 1
2 1
1 2 1 7
C
xx X
x t X
,) ) ( (
) (
2 2
2 2
2 2 2 8
C
xx X
x t X
.x y yx
C K
01 C
,z y yz
C K
02 C
,z x xz
C K
12 C
n N
f 1 1
1 ,
N f n 1 1
2 and
f
3 f
1 f
2. 2. Suggested Ratio-Cum-Product EstimatorsTailor et al. (2011 a) proposed ratio-cum-product estimators of population mean