Content uploaded by Xiaojiao yu
Author content
All content in this area was uploaded by Xiaojiao yu on Jul 16, 2017
Content may be subject to copyright.
Machine Learning Application in Online Leading Credit Risk
Prediction
Abstract
!
"!#
$ %
! &' !
(
&')#*!
+ ,-
. "*!
!
Keywords: Online lending, Big data, Random forest, XGBoost
1. Introduction
/
0
,#12.
'.
+
!+3
/!(!
+4!
5
+
+6*
7
* ! ! '
! 8 9
:8!+
( !
+ ; 7
!
+,-%
(
< ! +
=+
'!
>0+
!+
,,
! !
= +
,1< !$?!
@ A
! #
2
! ,3 < !
2
,4
?!
,5
B
"#
%!+!
&' $
8
( )#* ! ! 7C=!
2. Data and Variable De*nition
D
! !
%.
2.1 Lending Platform
#
!($ E!
$ ! !
9 1, $# F
',---!1---#!
,---6---
2.2 Phone Records
G ! ! ! 2
$ ,8
2.3 Third Party Data
;,!!"!(".
,
3. Models developed
% ! &'
:;;-@
A =
#
D 9-H
!3-H
1,,359
;-461
3.1. Random forest model parameter tuning
! ,6 G
=+
@IA! I! I +
*I+
(*I+
(
!@J
tp
tp+fp
!J
tp
tp+fn
A
=! G
5- 5--!
< 5---5--
G<,E
I!
<1BI,35!G
! < 3 ' . !
I5---!I1!I
,
<, G#
D>J5-!5--!5---
<1G#I
1!4!6
<3G#I
,!3!5
3.2 XGBoost *ne parameter tuning
&'&'
KK
.&'
< !
+ "
%$
"7!
"+.
,9 &'+"
&' 7 !
.I @ . A! @0 A!
I @ L
A! @
A! II @L
( A! @0, " A
@& + (A
7C=@A
7C=!
M .I
!.I1-!7C=
*!1
--,! =I -9!
-9! 8II ,! --,! -,!
7C=!-;4-64
1?&'
4. Experimental results
,-
&'?!
7 :;;- ! "*!
III6I,:::3,:I@#
6A!III,I,:::3,4I@
# ,# A!
III,1I,:::3
1-I @ # ,1 A!
>I8I
I&! III9I,:::3,1I! >I8I*I&!
II
I6-I,:::316I! ==!
III14I,:::314I ,-
<4
,- &' "*!
III,I
I,:::3,4I! ==!
III9I,:::3,1I!II
I3I,:::3,6I! III6I,:::3,:I!
*!=
=! IIIII6I!
III,1I,:::31-I
.
7"*
! 369H! ;
-,6HG ,- &'
* &'
!
,-
!
+<!.
!(
:;;-I
&' 7 &' ! "
" +
&'
0" < 5
)#* +
)#*
.)#*
-6494! &' -91-3* )#*
!
)#*! '
"!&'"
=! &'
+
"*
III6I,:::3,:I
III,I,:::3,4I
III,1I,:::31-I
>78%I8II&
III9I,:::3,1I
>78%I8I*I&
III6-I,:::316I
==
III14I,:::314I
-
--,
--1
--3
--4 369H
-,:H -,6H -,6H -,6H -,6H -,5H -,5H -,5H -,5H
@A,-
"*
III,I,:::3,4I
==
III9I,:::3,1I
III3I,:::3,6I
III6I,:::3,:I
*
==
IIIII6I
III,1I,:::31-I
-
----5
---,
---,5
---1
---15
---3
---35 -31H
-13H -13H -1-H
-,6H -,5H -,5H -,4H -,4H -,4H
@A&',-
<4,-@A@A&'
<5@A)#*@A)#*&'
5. Conclusion
! &'
.
% "
,-
&'
'
. "*!
!
)#*
+&'
6. Literature
@,A * * ? G1G % ? ' 1--:!
:6@1AD,9#45
@1A ' *=! & < % <
D 1 ' ! 1--;N
1@,AD3;#65
@3A B!?8% &!OGD' 8 0
1-!P=7 *!M
1;!>,D143#15:!1-,,
@4A Q(!B!Q0R< @1-,5A,D,;E
,-,,:6F4-:54#-,5#--,:#,
@5A 0= Q<,6@1---A,4;#,91
@6A 7 Q E B B! O7 8 *
7P!4%!G#?!C*!,;;:
@9A 0!R Q *!1-,1!
1!,:,#,:;
@:A 0 = ! O7 D
+ P Q
<! M ,6! > 1! 1---! ,4;#,91 D ,-,-,6F*-,6;#
1-9-@--A---34#-
@;A ER!??!S=QO=
P =>=P1--9
@,-A 7& 7 7 Q
8"!M,!>3!71-,,
@,,A 0>70!O7.
+!P%.
! M 36! > 1! 1--;! 3-1:#3-33 D
,-,-,6F1--:-,-,:
@,1A B!7!%0Q0!O0(
!P7%!M;;!
>,D4;#:4!1--;
@,3A 0!!7 )!%<G0!)*!O*
D
TP?)*< BG*!
Q<)*&!?C!,#41!1--;
@,4A <! *! & R Q! OE
T % GP>%
BG!,#63!1--:
@,5A B ?! = )! R B! * R @1-,5A O7 G1G
!O< ,@,AD,#:
@,6A '!0@1--,A!8!45!5#31
@,9A * '! G *! '! * ?"!
? Q = 7 @-;95#:::9A
M,6,#>,,!81-,9