Content uploaded by Nadeem Javaid
Author content
All content in this area was uploaded by Nadeem Javaid on Feb 22, 2022
Content may be subject to copyright.
Outline
!"#
$ %!
& '(!
) "
* +"
, !!
- .
Preliminary/
Classier 0
..12
Supervised Learning 3!"3
#4354
Smoothing parameter6!##"3
.7.6!3#"!#
.84
Modeling parameter83
Preliminary/
Parametric963"4(.!
34!"3:3
".!".!63!
.!;3!
!#6;.
Non-Parametric9<=63"!
!!!>
3!.#.""
Preliminary/
Local weighted least squares:
?3
Based on the rst law of geography: "3
"6#
K-Nearest-Neighbour:
!"
@#"#
4!#@#
.#33"#
#!4##ABC
$
Introduc'on
•@ "! #Cleveland and Susan
J.Devlin,**
• 3 =! 6 6
does not learn4(.!.!=!4!
•!".4
!#
• 3 a memory-based method that performs a
regression around a point of interest using only training dataDDAA
!5
5.!!(#E!
5(!3#.E!
5.!#.
!!(=".
F..E4
!!("
&
LWR Working 1/2
•3!"
•=!
•@(<!3!/3!
•3=#0=6=#6
=#
•%5
+G!
7="
7!!#
H%.
$ %.(!
& %!
)
LWR Working 2/2
•<=!
4!
•%!!!.(!3
!63E!6.
•%!.
•I.!3(=
!#3JK63
•7!!3
*
Improvements Base Learners
•<L" M 3 "4
@."
. ! . ! 3
•B<<algorithm..!3#
=!#.66
3!..3.
• 3 N'%% N'%% > ! #
363!46>86#3>3
.#."#
,
Steps in LWR
•H"%!J"K
•%".%!O!P-=
•%.#/.(-3#.J
• +3(
•$+".!Q
•&R(-SQ
-
'ET
'ET
'ET
Takes the average in parts and find the weight.
Efficiently classifies the data having different weights at different
points.
Example
O
LWR Advantages
•Advantages
•.U(#!0!.4"U(#
•<!=!E=:33"
•V
•@"!43
•'84"G!##G!
Disadvantages of LWR
.!.!!.6!#3
.!3!!
'8
@6#"63
-03##!
@"63!!
4
# = !#
F4
(4
F#%H.%!!!
LWR Applica'ons
References
WX?!5//#33!/-*/-)/-/$==3==/
WX45///F5/I//+3/,*&$-,,!.
WX?!5//!// ,$-&&&/
W X?!5//333./=3====!/
W$X?!5//333./==3==/
W&X?!5///Y!/=3==3==#==.-)*,&,,
W)X?!5///,/=3==/
W*X?!5//333!/=3====!/
W,X?!5//(/E/ & /=3=="===
W-X?!5//333.//#/--*-/-& $,,**- )*&,Z5[5(R\-3\-\F\-\-6!\-.\-\-\-
WX?!5//333/.//!;/;/!#/" /,&=/)Z5[5(R\-3\-\-\-6AA\-\-\-!
WX?!5//333=!/;/;/$**&,/"3
WX?!
5//!/[$-//-&/3/(!!]RZ5[5(R\-3\-\-\-6\-!
\-\-\-
W X?!5//333"/"*,*/#/!//!
W$X?!5//333/.//!;/= /#/!0//3#/3
$
Thank You !!!
&