ACTU 475 MDT S1 1444 - Corr

King Saud University Midterm Exam, S1-1444H
College of Sciences ACTU 475

Department of Mathematics Credibility Theory and Loss Distributions
Time: 90 minutes - Marks: 30
‫ـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ ـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ‬
Answer the following questions.
Q1: [5+5]
Determine the mean excess loss and limited expected value functions for the model of the age at death
that defined as:
 0, x 0

F (x )   0.01x , 0  x  100
 1, x  100

Q2: [5+5]
(a) Let  have a gamma distribution and let X  have a Weibull distribution with conditional survival
x 
function S
X
(x  )  e . Determine the unconditional or marginal distribution of X.
(b) Show that the Weibull distribution has a scale parameter.
Q3: [5+5]
(a) Using the following data set,
Pa yment ra nge 0-7500 7500-17500 17,500-32,500 32,500-67,500 67,500-125,000 125,000-300,000 Over 300,000
Number of 99 42 29 28 17 9 3
pa yments
Determine the loglikelihood function.
(b) Let X have a Pareto distribution with parameters  and  . Let Y  ln(1  X /  ). Determine the
name of the distribution of Y and its parameters.
‫ــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ‬
1
The Model Answer
Q1: [5+5]
(i)
eX (d ) 
 d
S (x )dx
S (d )
100
=

d
(1  0.01x )dx
1  0.01d
100
1  (1  0.01x ) 2 
  
0.01(1  0.01d )  2 d
1  0.01d 100  d
eX (d ) = 
0.02 2
(ii)
u
E (X  u )   xf (x )dx  u[1  F (u )]

u
  x (0.01)dx  u (1  0.01u )
0
 0.005u 2  u  0.01u 2
 E (X  u )  u (1  0.005u )
or
u
E (X  u )   S (x )dx
0
u
=  (1  0.01x )dx
0
u
 0.01x 2 
 x  
 2 0
 E (X  u )  u  0.005u 2
or
By using the following formula
E (X  u )  E (X )  e (u )S (u )
2
u 100
E (X )   x (0.01)dx   x (0.01)dx
0 0
100
x 2 
= 0.01  
 2 0
 E(x) = 50
Also, e (u )  1002u and S (u )  1  0.01u

E (X  u )  E (X )  e (u )S (u )
 50 - (100-u
2 )(1 - 0.01u )
 50 - 0.01
2 (100  u )
2
 E (X  u )  u  0.005u 2
Q2: [5+5]
(a)
let  gamma ( , ), X  weibull( , )


 S X  (x  )  e  x
 A(x )  x 
S X (x )  E [e A(x ) ]
=M  [A(x )]
 S X (x )  M  [x  ]
and M  (z )  (1   z ) 
 S X (x )  (1   x  ) 
which is a Burr distribution with parameters

1
   ,  
(b)
For two parameter Weibull distribution -  , 


FX (x )  1  e  (x / )
Let Y  cX , where c  0 Then
3
FY (y )  Pr(cX  y )  Pr(X  yc )

FY (y )  1  e  (y /c )
which is also a Weibull distribution, with parameters  and c .
 is a scale parameter.
Q3: [5+5]
(a) The loglikelihood function is
l ( )  99 ln[F (7500)  F (0)]  42 ln[F (17,500)  F (7500)]

 29 ln[F (32,500)  F (17,500)]  28ln[F (67,500)  F (32,500)]
+17 ln[F (125, 000)  F (67,500)]  9 ln[F (300, 000)  F (125, 000)]
+3ln[(1  F (300, 000)]
l ( )  99 ln[1  exp(7,500 /  )]  42 ln[exp( 7,500 /  )  exp( 17,500 /  )]

 29ln[exp(17,500 /  )  exp(32,500 /  )]  28ln[exp(32,500 /  )  exp(67,500 /  )]
+17 ln[exp(67,500 /  )  exp(125, 000 /  )]  9 ln[exp(125, 000 /  )  exp(300, 000 /  )]
+3ln[exp(300, 000 /  )]
(b)
X Pareto ( ,  )

  
 FX (x )  1   
 x  
For Y  ln(1  X /  ), FY (y )  Pr(Y  y )
FY (y )  Pr[ln(1  X /  )  y ]
 FY (y )  Pr[1  X /   e y ]
 Pr[X   (e y  1)]

  
 1  y 
  (e  1)   

1
 FY (y )  1   y 
e 
 1  e  y ,
which is the distribution function of the exponential distribution with parameter 1/  .

ACTU 475 MDT S1 1444 - Corr

Uploaded by

Copyright:

Available Formats

ACTU 475 MDT S1 1444 - Corr

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ACTU 475 MDT S1 1444 - Corr

Uploaded by

Copyright:

Available Formats

King Saud University Midterm Exam, S1-1444H

College of Sciences ACTU 475

(b) Show that the Weibull distribution has a scale parameter.

(a) Using the following data set,

Determine the loglikelihood function.

By using the following formula

Also, e (u )  1002u and S (u )  1  0.01u

let  gamma ( , ), X  weibull( , )

which is a Burr distribution with parameters

For two parameter Weibull distribution -  , 

Let Y  cX , where c  0 Then

which is also a Weibull distribution, with parameters  and c .

l ( )  99 ln[F (7500)  F (0)]  42 ln[F (17,500)  F (7500)]

l ( )  99 ln[1  exp(7,500 /  )]  42 ln[exp( 7,500 /  )  exp( 17,500 /  )]

which is the distribution function of the exponential distribution with parameter 1/  .

You might also like