没有合适的资源?快使用搜索试试~ 我知道了~
论文研究-众筹项目的个性化推荐:面向稀疏数据的二分图模型.pdf
需积分: 10 1 下载量 32 浏览量
2019-09-20
20:25:41
上传
评论
收藏 1.26MB PDF 举报
温馨提示
试读
13页
论文研究-众筹项目的个性化推荐:面向稀疏数据的二分图模型.pdf, 二分图模型是一种全局优化算法,本文将二分图模型应用于直接推荐众筹项目,使用PersonalRank算法迭代计算网络节点的全局关联度,从而推荐那些基于余弦相似度的协同过滤不能有效推荐的项目,适用性更加广泛. 更进一步,提出将二分图模型与协同过滤算法相结合,首先把网络结构划分为二分图,采用二分图算法得到的两类节点(用户节点,项目节点)之间的全局相似度,再结合协同过滤算法,得到基于二分图模型的协同过滤算法. 实验表明,在众筹项目推荐中,由于数据极端稀疏,适宜采用二分图模型来进行相似度计算并进行推荐.
资源推荐
资源详情
资源评论
37
4
Vol.37, No.4
2017
4
Systems Engineering — Theory & Practice Apr., 2017
doi: 10.12011/1000-6788(2017)04-1011-13
:TP18
:A
Æ
:
1
,
2
,
3
,
4
(1.
,
362021; 2. Eller College of Management, University of Arizona, Tucson 85721;
3. Rady School of Management, University of California, San Diego 92093; 4.
,
200092)
,
!"
#
$
,
%
Per-
sonalRank
&'()*+,-./0
,
12!"3
4
50-
6789:
!"-
$
,
;
Æ
<=>?
.
<@
A
,
BC
678
DE
,
F
G)*
DHIJ
,
K
LM-NO+,
(
P+,
,
$+,
)
Q-
50
,
DE
678
,
LM4
-
678
.
R
ST
,
#
$!"
,
!
UV
WX
"
Y
,
;
#
K
Z@
$
50([@
$
!"
.
#
;
!"
%
\
;
;
)*DH
Personalized recommendation of crowd-funding campaigns:
A bipartite graph approach for sparse data
WANG Wei
1
,CHENWei
2
,ZHUKevin
3
, WANG Hongwei
4
(1. College of Business Administration, Huaqiao University, Quanzhou 362021, China; 2. Eller College of Management,
University of Arizona, Tucson 85721, USA; 3. Rady School of Management, University of California, San Diego 92093, USA;
4. School of Economics and Management, Tongji University, Shanghai 200092, China)
Abstract Bipartite graph is a global optimal algorithm, which enables direct recommendation of crowd-
funding campaigns. In our method, PersonalRank is applied to calculate global similarity for a network in
an iterative manner. It can be applied to recommendations where Cosine similarity function is ineffective.
Furthermore, we propose a bipartite graph based collaborative filtering (CF) by combining CF and Per-
sonalRank. The nodes are classified into one of the following two types: user nodes and item nodes. For
any two types of nodes, the new model calculates the global similarity between the nodes by PersonalRank,
and obtains the recommendation list through CF algorithm. Experiment results show that the bipartite
graph based CF achieves better performance for the extremely sparse data from crowd-funding community.
Keywords crowd-funding; recommendation system; bipartite graph; network structure
1
]
,
Kickstarter
8604863
230850
,
: 2015-12-22
^_
:
(1982–),
,
,
,
,
,
Æ
:
,
, E-mail: wwang@hqu.edu.cn;
(1983–),
,
,
,
Æ
:
,
, E-mail: weichen@email.arizona.edu;
(1963–),
,
,
,
,
Æ
:
, E-mail: kxzhu@ucsd.edu;
:
(1973–),
,
,
,
,
,
,
Æ
:
, E-mail: hwwang@tongji.edu.cn.
:
Æ
·
(16SKGC-QG14);
(71
601082, 71371144);
(FJ2016B075)
Foundation item: Huaqiao University’s Academic Project Supported by the Fundamental Research Funds for the Central
Universities (16SKGC-QG14); National Natural Science Foundation of China (71601082, 71371144); Social Science Planning
Project of Fujian Province, China (FJ2016B075)
`
ab
:
,
,
,
.
:
[J].
, 2017,
37(4): 1011–1023.
c
ab
: Wang W, Chen W, Zhu K, et al. Personalized recommendation of crowd-funding campaigns: A bipartite graph
approach for sparse data[J]. Systems Engineering — Theory & Practice, 2017, 37(4): 1011–1023.
1012
37
22525091
(www.kickstarter.com).
,
60%
,
,
!
,
[1]
.
,
"
!
.
!
#$
,Kickstarter
%&
"
99.99%.
%&
'
,
(
#$
"
)
Æ
.
,
*
!
+
"
'
"
,
,
#$%
-
%
!
,
,
&
)
'
!
+
"
,
&
"
'
.
,
%&
'
(
)
!
(
!
+
,
)
"
"
$
!
[2]
.
"
#
%&'
,
.
$
!
/)
!
$
"
0
.
, PageRank
)
.
*
$
"
)
.
PageRank
%
!
+
,
!
(
1
PageRank
)
#
!
"
&
.
PageRank
#
!
)
(
"
$!
2
#
+
)
3
'
4
%
5
(
.
(
$!
2
#
+
,
)
.
$
&!
-
#
'
,
$
%&%
,
$
,
%&%
.
)
(
,
(
$!
2
"
)
'
!
+
"
,
)
'
!
+
"
,
*
'
'
*
!
+
"
,
&
)*
+
'
%& !
.
-
'
*
,
.
%&'
6
,
$!
2
#
+
&
Æ
"
'
.
)
!
,
,
.
$!
2
#
+
'
-.
,
/
)
$
(
"
7
,
+
)
(
)
$
,
$
(
"
.
)
(
"
(
.
$
/
0
01
,
*
'
&
)
!
*
!
2
,
+
*
3
0
'
%&
4
!
!
.
)
$!
2
#
+
!
+
"
)
#
-
"
,
!
#
.
,
4
5
$!
2
#
+
"
,
)
,
&
"
'
.
-
'
*
,
.
%&'
/
,
6
,
8
1
(
$!
2
#
+
!
"
.
2
"
2.1
Æ
!
#
"#
$
"
)
, PageRank
-
6
)
. PageRank
“
7
2
$
.
*
.%
/
.
*
,
89
+
2
$
.
*
”
Æ
9
,
/
+
9
:
.
*
$
,
$
0
*
"
:
%
3
.
*
:
1
4
+
0
,
[3]
.
$!
2.
$
;
5
,
.
-
2
.
$
!
/
;
0
6
#
+
&
[4]
.
$!
2
PageRank
,
,
PageRank
1
.
$
<,
$
,
$!
2
!
2
+
,
,
+
$
<
%&%
,
,
+
$
%&%
[5]
.
.
$
)
(
=
7
$!
2
,
,
.
.
$
!
2
#
+
1
8
>
)
&
)
$
34
,
'
)
)
(
(
$
!
+
"
[6]
,
:
5
Æ
!
+
"
.
.
"
,
8
4
5
(
6
.$!
2
#
+
/7
?
2
#
+
)
3
9
"
,
7
?
"
)
/0
[7]
.
2.2
Æ
*
$
+
%&'
,#
"#
"
,
8
=
.
"
:
45
:
.
&
9
>
,
)
$
'
!
+
"
,
)
(
"
!
+
6
1
2
.
"
,
8
=
!
2
(item-based)
"
,
7
(user-based)
"
,
.
"
,
)
?@
@
3
*
,
A(
3
*
)
!
+
"
34
,
!
+
"
3
4
;:
"
)
.
'
'
42
6
1
,
B
.
'
%& !
,
)
2
<=
"
Æ
0
[2]
,
5
(
'
6
,
6
2
7
?
)
:
/
"
,
8
4
5
5
,6
7
6
;
"
,
#
+
,
Æ
-
2
+
>
[8]
.
.
5
(
(
!!
/
7
'-.
?
7
6
8.
"
,
#
+
[9]
,
.
9"
'
%& !
.
3
()
/
de
*+,
0
f
3.1
"#
1
g
(
2
1
,
9
8
A, B, C, D
.'
,
:
8
e, f, g, h
.'
.
-
(
*
!
+
"
"
,
)
,
A
!
*
C
7
B,
f
@A
)
&
"
1
A,
A
!
*
f
%
.
,
,
"
,
)
,f
,
)
&
"
1
A.
*
!
+
"
)
'
*
)
,
.
%&
.
$
#
'
&
'
!
+
"
!
)
.
'
*
!
+
"
.
C
9
'
6
"
)
/
"
+
*
,
.
'
%&
5
<
6
:
&
#
>
#
-
.
4
,
:
:
1013
.
$!
2)
,
)
(
#
2
-
$
34
,
,
&
)
)
34
(
(
"
,
#
%
"
#
-
.
,
2
1
-
(
#
2
2
:$
$!
2
.
(
PersonalRank
$!
2
!
)
.
-
A
4
%
"
,
=
(
A
;
,
!
-.
B
)
,
-
62
-.
,
)
#
-
A
>
,
#
A
;
(
"
.
s(A, e)=0.07709,s(A, f)=0.01791,s(A, g)=0.09499,s(A, h)=0.26949.
/
<
A
-
h,
.
B
3
,
"
C
C
:
?
g,
e,
(
f.
)
,
PersonalRank
,
-.
!
+
"
,
#0D
5
.
2
2
,A
E
$
D
0
!
+
"
:
s(A, B)=0.13602,s(A, C)=0.13602,s(A, D)=0.04213.
F
!
+
"
*
'
,
=
<
>
'
!
+
"
,
.
(
*
,
=
<
>
'
,
#
$
'
*
!
+
"
(
"
!
*
).
(
2
2
,
0
A
7
D
=
,
,
*
'
7
=
<
>
!
+
"
'
G
)H
'
$
!
+
"
,
1
$
?
s(A, D)=0.
.
C
9
'
=
,
,
&
#
?
"
*
6
.
,
.
%&'
5
<
6
,
'
)
!
+
"
#
)
)
Æ
.
A
B
C
h
e
D
f
g
-
1
./
2
01
3
h
234
4
`
5
ib
5
j
-
-
2
67
-
k
8
5
j
-
'
1
9
#
D
I
!
8
!
5
.
"
,
J
)
8
(
2
#
+
,
>
(
$!
2
#
+
,
E
:&
-%
'
%&
!
.
;
1
l
<
678
=5
mn
679
o
?
>
F
@
E
?
Rakesh
@
Choo
[10]
2015
4
K
F
:
L
A
M
@
,
M
@
,
G
M
@
G
N
@
M
@
.
O
A
P
A
Q
A
B
HI
.
H
R
M
@
S
,
BC
D
.
J
B
I
E
B
P
,
C
F
J
G
K
.
An
@
Quercia
[1]
2014
N
@
C
G
H
Æ
L
K
I
D
J
.
E
T
L
M
K
E
T
N
E
.
E
R
C
J
F
U
D
.
M
K
N
@
,
C
O
D
P
.
Lu
@
Shuai
[11]
2014
N
@
C
G
Æ
L
K
I
D
J
.
L
E
R
C
J
.
Æ
L
E
Æ
GH
,
Æ
F
.
Stone
@
Zhang
[12]
2013
L
T
M
I
D
E
MN
N
O
J
O
V
K
.
QR
W
.
N
O
J
E
SE
T
F
T
.
KNN
D
Æ
M
K
ÆD
J
U
P
.
Zhou
@
Kuscsik
[13]
2010
K
G
DO
HIG
V
D
.
“
L
JK
”
L
HJ
K
I
D
J
Æ
L
J
M
G
.
P
W
S
Q
.
L
RU
S
Q
H
MN
N
Q
,
X
O
O
T
.
3.2
"#:;
I
p
(
Kickstarter
.'
(
Nothing-or-More
#
0
,
R
0
40%.
ST
P
:Q
Y
U
,
W
,
1
G:
V
.
)
&
$
P
!
,
,
)
&
.
8
.
(
5
<
,
9
.
"
,
&
4
0
R
0
.
,
RS
$!
2
#
+
2
Z
,
#
.
"
,
)
2
,
4
5
(
6
8
!
/
[\
"
.
1)
'
W
X
%&
]
,
)
-%
=
7
$!
2
,
&
)
$
!
+
"
.
2)
X
$!
2
#
+
!
+
"
,
4
5
$!
2
#
+
"
,
)
,
^
"
.
3)
(
Kickstarter
'
Y
,
-
J
$!
2
#
+
"
)
Æ
.
剩余12页未读,继续阅读
资源评论
weixin_38743737
- 粉丝: 376
- 资源: 2万+
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功