ASSIGNMENT 3

COMPARISON OF ESTIMATORS USING MEAN SQUARE ERROR CRITERION

Q1)
CODE:
p=seq(0.1,1,by=0.1);p
q=1-p
MSE1=(2*p*q)+(p*p)
MSE2=(p*q)/2
MSE3=(5*p*q)/9
MSE4=(13*p*q)/25
plot(p,MSE1,col=1,lty=1,type="l",xlab="p",ylab="MSE",ylim=c(0,1),main="MSE CURVES")
lines(p,MSE2,col=2,lty=2)
lines(p,MSE3,col=3,lty=3)
lines(p,MSE4,col=4,lty=4)
legend("topright",legend=c("MSE1","MSE2","MSE3","MSE4"),col=1:4,lty=1:4)
PLOT:

RESULT:
On comparing the
graph, MSE2 is a
better estimator.
Q2)
CODE:
lam=seq(0.1,1,by=0.1);p
MSE1=lam/15
MSE2=((15*lam)+(lam^2))/196
plot(lam,MSE1,col=1,lty=1,type="l",xlab="p",ylab="MSE",ylim=c(0,0.1),main="MSE CURVES")
lines(p,MSE2,col=2,lty=2)
legend("topright",legend=c("MSE1","MSE2"),col=1:2,lty=1:2)
PLOT:

RESULT:
On comparing the
graph, MSE1 is a
better estimator.
Q3)
CODE:
mu=seq(-3,3,by=1);mu
n=10
n1=c(0.1,0.1,0.1,0.1,0.1,0.1,0.1)
MSE1=(n+((mu)^2))/((n+1)^2)
MSE2=(n+(4*(mu)^2))/((n+2)^2)
MSE3=1/n
plot(mu,MSE1,col=1,lty=1,type="l",xlab="lambda",ylab="MSE",ylim=c(0,0.4),main="MSE
CURVES")
lines(mu,MSE2,col=2,lty=2)
lines(mu,n1,col=3,lty=3)
legend("topright",legend=c("MSE1","MSE2","MSE3"),col=1:3,lty=1:3)
PLOT:

RESULT:
On comparing the
graph, as n increases
MSE1 drastically
decreases and
remains constant.
But, when compared
to MSE2 and MSE3 it
is a better estimator.
Q4)
CODE:
t=c(0,1,2,3,4,5,6,7,8,9)
MSE1=t^2
MSE2=7*(t^2)
MSE3=(3/8)*(t^2)
MSE4=(t^2)/10
plot(t,MSE1,col=1,lty=1,type="l",xlab="theta",ylab="MSE",main="MSE CURVES")
lines(t,MSE2,col=2,lty=2)
lines(t,MSE3,col=3,lty=3)
lines(t,MSE4,col=4,lty=4)
legend("topright",legend=c("MSE1","MSE2","MSE3","MSE4"),col=1:4,lty=1:4)
PLOT:
Q5)
CODE:
p=seq(0.1,1,by=0.1);p
q=1-p
MSE1=P
MSE2=((2/3)*p*q)+(p^2)
MSE3=((7/9)*p*q)+(p^2)
MSE4=(3/4)*p*q
plot(p,MSE1,col=1,lty=1,type="l",xlab="p",ylab="MSE",ylim=c(0,1),main="MSE CURVES")
lines(p,MSE2,col=2,lty=2)
lines(p,MSE3,col=3,lty=3)
lines(p,MSE4,col=4,lty=4)
legend("topright",legend=c("MSE1","MSE2","MSE3","MSE4"),col=1:4,lty=1:4)
PLOT:

RESULT:
On comparing the
graph, MSE4 has a
decreasing curve,
hence it is a better
estimator.
 ASSIGNMENT 4
ESTIMATION OF PARAMETERS BY MAXIMUM LIKELIHOOD METHOD AND
METHOD OF MOMENT
Q3)
CODE:
x=c(594,689,585,1034,720,929,737,1026,791,836,952,898,1043,806,577,994)
n=16
#MLE_mu=sample mean
#MLE_var=sample variance
mu=sum(x)/n;mu
var=((1/n)*(sum(x^2)))-((mu)^2);var
OUTPUT:
x=c(594,689,585,1034,720,929,737,1026,791,836,952,898,1043,806,577,994)
> n=16
> #MLE_mu=sample mean
> #MLE_var=sample variance
> mu=sum(x)/n;mu
[1] 825.6875
> var=((1/n)*(sum(x^2)))-((mu)^2);var
[1] 25237.59

Q4)
CODE:
x=c(28,32,29,23,28,21,24,22,14,19,25,20,15,26,35,23,12,13,36,32)
n=20
#MLE_mu=sample mean-(sqrt(3)*sd)
#MLE_var=sample variance+(sqrt(3)*sd)
mu=sum(x)/n;mu
sd=sqrt(((1/n)*(sum(x^2)))-((mu)^2));sd
a=mu-(sqrt(3)*sd);a
b=mu+(sqrt(3)*sd);b
OUTPUT:
> x=c(28,32,29,23,28,21,24,22,14,19,25,20,15,26,35,23,12,13,36,32)
> n=20
> #MLE_mu=sample mean-(sqrt(3)*sd)
> #MLE_var=sample variance+(sqrt(3)*sd)
> mu=sum(x)/n;mu
[1] 23.85
> sd=sqrt(((1/n)*(sum(x^2)))-((mu)^2));sd
[1] 6.915743
> a=mu-(sqrt(3)*sd);a
[1] 11.87158
> b=mu+(sqrt(3)*sd);b
[1] 35.82842
>
Q8)
CODE:
x=c(29.74,37.09,33.94,25.89,21.54,31.01,28.71,29.52,22.89,30.15,33.25,22.80,20.26,26.06,20.
96,20.50,36.19,26.13,15.77,36.69)
n=20
#MLE_mu=sample mean
#MLE_var=sample variance
mu=sum(x)/n;mu
var=((1/n)*(sum(x^2)))-((mu)^2);var
OUTPUT:
>x=c(29.74,37.09,33.94,25.89,21.54,31.01,28.71,29.52,22.89,30.15,33.25,22.80,20.26,26.06,2
0.96,20.50,36.19,26.13,15.77,36.69)
> n=20
> #MLE_mu=sample mean
> #MLE_var=sample variance
> mu=sum(x)/n;mu
[1] 27.4545
> var=((1/n)*(sum(x^2)))-((mu)^2);var
[1] 36.1764
>