14
Estimation: chapter 2+3 Minimum variance unbiased estimation + the CRLB Natasha Devroye [email protected] http://www.ece.uic.edu/~devroye Spring 2011 • Find a few estimators Estimation: a first example • Estimate the DC level, A, of a signal given noisy measurements x[0], x[1], ... x [N-1] where x[n] are samples of this! • Compare their performance { • mean? • variance? • pdf?

Estimation: chapter 2+3 - UIC EngineeringDevroye/Courses/ECE531/Lectures/B1c2+3.pdfMinimum variance unbiased estimation Give a counter-example! (b1pg.20) The Cramer-Rao Lower Bound

  • Upload
    others

  • View
    2

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Estimation: chapter 2+3 - UIC EngineeringDevroye/Courses/ECE531/Lectures/B1c2+3.pdfMinimum variance unbiased estimation Give a counter-example! (b1pg.20) The Cramer-Rao Lower Bound

Estimation: chapter 2+3

Minimum variance unbiased estimation + the CRLB

Natasha [email protected]://www.ece.uic.edu/~devroye

Spring 2011

• Find a few estimators

Estimation: a first example

• Estimate the DC level, A, of a signal given noisy measurements x[0], x[1], ... x[N-1] where

x[n] are samples of this!

• Compare their performance { • mean?• variance?• pdf?

Page 2: Estimation: chapter 2+3 - UIC EngineeringDevroye/Courses/ECE531/Lectures/B1c2+3.pdfMinimum variance unbiased estimation Give a counter-example! (b1pg.20) The Cramer-Rao Lower Bound

Estimation: a first example

• Estimators of the DC level, A

Estimation: definitions

Page 3: Estimation: chapter 2+3 - UIC EngineeringDevroye/Courses/ECE531/Lectures/B1c2+3.pdfMinimum variance unbiased estimation Give a counter-example! (b1pg.20) The Cramer-Rao Lower Bound

Estimation: definitions

How would you pick a ``good’’ estimator?

Vector versions....

Minimum variance unbiased estimation

Why?

So?

Page 4: Estimation: chapter 2+3 - UIC EngineeringDevroye/Courses/ECE531/Lectures/B1c2+3.pdfMinimum variance unbiased estimation Give a counter-example! (b1pg.20) The Cramer-Rao Lower Bound

Minimum variance unbiased estimation

Give a counter-example! (b1pg.20)

The Cramer-Rao Lower Bound

• the CRLB give a lower bound on the variance of ANY UNBIASED estimator

• does NOT guarantee bound can be obtained

• IF find an estimator whose variance = CRLB then it’s MVUE

• otherwise can use Ch.5 tools (Rao-Blackwell-Lehmann-Scheffe Theorem and Neyman-Fisher Factorization Theorem) to construct a better estimator from any unbiased one - possibly the MVUE if conditions are met

Page 5: Estimation: chapter 2+3 - UIC EngineeringDevroye/Courses/ECE531/Lectures/B1c2+3.pdfMinimum variance unbiased estimation Give a counter-example! (b1pg.20) The Cramer-Rao Lower Bound

The Cramer-Rao Lower Bound (CRLB)

• Use?

• Intuition?

The Cramer-Rao Lower Bound (CRLB)

Page 6: Estimation: chapter 2+3 - UIC EngineeringDevroye/Courses/ECE531/Lectures/B1c2+3.pdfMinimum variance unbiased estimation Give a counter-example! (b1pg.20) The Cramer-Rao Lower Bound

CRLB examples

CRLB proof

Page 7: Estimation: chapter 2+3 - UIC EngineeringDevroye/Courses/ECE531/Lectures/B1c2+3.pdfMinimum variance unbiased estimation Give a counter-example! (b1pg.20) The Cramer-Rao Lower Bound

CRLB T or F

What is I(θ)?

• Why “information”?

• non-negative• additive for independent observations

Page 8: Estimation: chapter 2+3 - UIC EngineeringDevroye/Courses/ECE531/Lectures/B1c2+3.pdfMinimum variance unbiased estimation Give a counter-example! (b1pg.20) The Cramer-Rao Lower Bound

Vector form of the CRLB

Vector form of the CRLB

Page 9: Estimation: chapter 2+3 - UIC EngineeringDevroye/Courses/ECE531/Lectures/B1c2+3.pdfMinimum variance unbiased estimation Give a counter-example! (b1pg.20) The Cramer-Rao Lower Bound

Vector form of the CRLB examples

What can we conclude?

CRLB for transformations

Page 10: Estimation: chapter 2+3 - UIC EngineeringDevroye/Courses/ECE531/Lectures/B1c2+3.pdfMinimum variance unbiased estimation Give a counter-example! (b1pg.20) The Cramer-Rao Lower Bound

Vector CRLB for transformations

Example of vector CRLB with transformation

Page 11: Estimation: chapter 2+3 - UIC EngineeringDevroye/Courses/ECE531/Lectures/B1c2+3.pdfMinimum variance unbiased estimation Give a counter-example! (b1pg.20) The Cramer-Rao Lower Bound

CRLB for General Gaussian Case

• When observations are Gaussian and one knows the dependence of the mean and covariance matrix on the unknown parameters, we know the closed form of the CRLB (or Fisher information matrix):

CRLB for Gaussians examples

Page 12: Estimation: chapter 2+3 - UIC EngineeringDevroye/Courses/ECE531/Lectures/B1c2+3.pdfMinimum variance unbiased estimation Give a counter-example! (b1pg.20) The Cramer-Rao Lower Bound

Example: range estimation

!"#$

%&'(

)*+

,)+)+)+ !"

#"

! $$##%#"#&

!

!!

!

"#$$"%#!"!#$

-.&/0%12345607833 "+#)333333/9/:7.939;7.3#&<,($"=

>7?71;73>7@67?219/8333 "+#'( !!)

A7&05.73-1%73B76&C8 !!

!"#$%&'()*$+,-.*

D&/E61%127E

FG1273H&5001&/

#$"

"+#)

#$

"+#''( !!)D4I

J3K%L

&+#)

/%0$,1$!2"3

)'*(*

#$45

678$9$:)('.$6;<&=)<&,($/>,?*.=

Continuous time signal model

!"#$

%&'(

)*+

,)+)+)+ !"

#"

! $$##%#"#&

!

!!

!

"#$$"%#!"!#$

-.&/0%12345607833 "+#)333333/9/:7.939;7.3#&<,($"=

>7?71;73>7@67?219/8333 "+#'( !!)

A7&05.73-1%73B76&C8 !!

!"#$%&'()*$+,-.*

D&/E61%127E

FG1273H&5001&/

#$"

"+#)

#$

"+#''( !!)D4I

J3K%L

&+#)

/%0$,1$!2"3

)'*(*

#$45

678$9$:)('.$6;<&=)<&,($/>,?*.=

Taken from http://www.ws.binghamton.edu/fowler/fowler%20personal%20page/EE522.htm

Page 13: Estimation: chapter 2+3 - UIC EngineeringDevroye/Courses/ECE531/Lectures/B1c2+3.pdfMinimum variance unbiased estimation Give a counter-example! (b1pg.20) The Cramer-Rao Lower Bound

Example: range estimation

Discrete time signal model

Taken from http://www.ws.binghamton.edu/fowler/fowler%20personal%20page/EE522.htm

!"#$

%&'()*+,-*./+! 0+#"1!"2*3

#4$5+0+#6$!7

89+:;<=*+

>&?22<&@+AB<2*

C&.+"! " !"#

#DD#DE547654 #$%#!$ %$$#$&$' ( !&

)"

#$%&'(&$)%*

&#"1!

!*!

#"! !"1!

+,-&'(&$)%*

"#.! "! " !"

#

''(

'')

*

#++%

#%++%#!

#++

$

#54

#5476

#E54

54

%$+$$#

+$$$$#$&

$$$#

$'

(

(((

(

&

&4$F&(5G ;&2++ @B@HI*.B+2&'()*2+2=&.=<@J+&=+$(

/0123&45678067'1&-9:&,7210;&<'=3;

!"#$

%&'()*+,-*./+! 0+#"1!"2*3

#4$5+0+#6$!7

89+:;<=*+

>&?22<&@+AB<2*

C&.+"! " !"#

#DD#DE547654 #$%#!$ %$$#$&$' ( !&

)"

#$%&'(&$)%*

&#"1!

!*!

#"! !"1!

+,-&'(&$)%*

"#.! "! " !"

#

''(

'')

*

#++%

#%++%#!

#++

$

#54

#5476

#E54

54

%$+$$#

+$$$$#$&

$$$#

$'

(

(((

(

&

&4$F&(5G ;&2++ @B@HI*.B+2&'()*2+2=&.=<@J+&=+$(

/0123&45678067'1&-9:&,7210;&<'=3;

Example: range estimation

Taken from http://www.ws.binghamton.edu/fowler/fowler%20personal%20page/EE522.htm

!"#$

%&'()**+,(-.)/0)10(2345(16-7+.(8&1(-9:/)+(;(<=%>

!!

!!

"

# $#

"%

# "$#

"%

#

"

#

&&'

())*

+

,

,#

&&

'

(

))

*

+

,

,

#

&&'

())*

+

,

"$,#

&&'

())*

+

,

,-

#

?

@

@

#@

@

# @

@

#

?

@

@

ABAB

ABCDE

AFG)1B

!

" "#

!"

"" "#

!"

"" $

$

%

" $

$

$

#

#&

#

#&

"&"&

$

$ $

$

$

..

/

/

.

/

/

./

/

H+7:(9/I()/0(J66*(

/&/KL61&(.61M-

NO*+&9.(2)+P7+7-QQQ R-6()**1&O9M).9&/>(/$'('$ "

$

ST6/(0&(PT)/:6(&8(G)19)U+6-QQ

!"#$%&'()*+")*,#&-!./

Page 14: Estimation: chapter 2+3 - UIC EngineeringDevroye/Courses/ECE531/Lectures/B1c2+3.pdfMinimum variance unbiased estimation Give a counter-example! (b1pg.20) The Cramer-Rao Lower Bound

Example: range estimation

Taken from http://www.ws.binghamton.edu/fowler/fowler%20personal%20page/EE522.htm

!"#$

!

"

#

!

"

#

"#

$

%&&

&!

'

$%&

&

&!

'

%&&

&! !!!

!!! "#

$%&

'((

)

"#

$%&

'((

)

"#

$%&

'((

*

+

%

&

%

&

%

&

&

'(

&"

#

'(

&"

'(#')*+,(

,-

-../012.+03412.3+567826.2.0+449 +33,:;<2./02=>267?18,+49

!)!"

! %&&!$%

& '(

. /

. /

!!

!

0

01

0

01

)

+

%&()

%(()(

'

$

$

%(()(

'

$

#

!

!

"

#

!

#!

&

&&

%

&&

'(

'(&

&"

#

'(&

&"

#')*+,(

2

2- @A2AB1:,10

C2D+,.1*+4

D+,.1*+4. /

!

!0

01

0

01)%&()

%(()(*+,!

&

&&

'(

'(&2

!"#$%"&'&()&*"'+,-".

*+,! 6.2EFGH2IJK22(LM'-2?>312:N2EHOFK

/'%0"&1+2$*'2$3%&4/5(&673%289

Example: range estimation

Taken from http://www.ws.binghamton.edu/fowler/fowler%20personal%20page/EE522.htm

!"#$

%&'()*+,-&-*'.-/&*0-*/11'2-*/+*+,-*3456*7(*+,-*.-8/9:

;<&-=#

;>2/1< ?

?!"#

$

%&'( !"#

@7*)-+*+,-*3456*7(*+,-*!"#$%A*B&-*C+1/(&DE*7D*F/1G&H*1-&B8+:*

;<GI"

;>2/1< ?

?

?

!"#%&'(

)(

!"

$*(+%

(*(+%

$( ##

>

?

> $$%

&''(

)

*

*+ 0'+,* ( J*)#

$,"*?

3456*'&*'(2-1&-89*F17F71+'7(/8*+7:

K LM4*N-/&B1-

K 4NL*6O*N-/&B1-

3456*'&*'(2-1&-89*F17F71+'7(/8*+7:

K LM4*N-/&B1-

K 4NL*6O*N-/&B1-

L7*+,-*3456*+-88&*B&A

K 3,77&-*&')(/8*0'+,*8/1)-*%!"#

K P(&B1-*+,/+*LM4*'&*8/1)-

K 6-++-1*7(*M-/1Q9"8/1)-*+/1)-+&

K O,'=,*'&*Q-++-1R**

S T7BQ8-*+1/(&G'++-.*-(-1)9R

S T7BQ8-*4NL*Q/(.0'.+,R

&"#$%'()*+,"*+-#'.&/0'12-#*34

!"#$

%&'()*+,-&-*'.-/&*0-*/11'2-*/+*+,-*3456*7(*+,-*.-8/9:

;<&-=#

;>2/1< ?

?!"#

$

%&'( !"#

@7*)-+*+,-*3456*7(*+,-*!"#$%A*B&-*C+1/(&DE*7D*F/1G&H*1-&B8+:*

;<GI"

;>2/1< ?

?

?

!"#%&'(

)(

!"

$*(+%

(*(+%

$( ##

>

?

> $$%

&''(

)

*

*+ 0'+,* ( J*)#

$,"*?

3456*'&*'(2-1&-89*F17F71+'7(/8*+7:

K LM4*N-/&B1-

K 4NL*6O*N-/&B1-

3456*'&*'(2-1&-89*F17F71+'7(/8*+7:

K LM4*N-/&B1-

K 4NL*6O*N-/&B1-

L7*+,-*3456*+-88&*B&A

K 3,77&-*&')(/8*0'+,*8/1)-*%!"#

K P(&B1-*+,/+*LM4*'&*8/1)-

K 6-++-1*7(*M-/1Q9"8/1)-*+/1)-+&

K O,'=,*'&*Q-++-1R**

S T7BQ8-*+1/(&G'++-.*-(-1)9R

S T7BQ8-*4NL*Q/(.0'.+,R

&"#$%'()*+,"*+-#'.&/0'12-#*34