基于信息熵協(xié)同過(guò)濾算法

基于信息熵的協(xié)同過(guò)濾算法匯報(bào)人：張佳閩南師范大學(xué)計(jì)算機(jī)學(xué)院2021/7/8預(yù)備知識(shí)1主要內(nèi)容基于用戶的算法2基于信息熵的協(xié)同過(guò)濾算法3實(shí)驗(yàn)結(jié)果4參考文獻(xiàn)52021/7/8協(xié)同過(guò)濾思想:根據(jù)用戶過(guò)去的瀏覽記錄，購(gòu)買記錄，評(píng)分記錄及其主頁(yè)標(biāo)注等信息，分析用戶潛在的興趣偏好，以給用戶提供個(gè)性化的推薦服務(wù)。例1：

the

mendation

of

movies

in

MTime2021/7/8國(guó)內(nèi)外研究現(xiàn)狀推薦系統(tǒng)二部圖 混合過(guò)濾 基于內(nèi)容 協(xié)同過(guò)濾基于模型 基于內(nèi)存基于項(xiàng)目基于用戶2021/7/8存在的問(wèn)題2021/7/8可拓展性（Scalability）：當(dāng)數(shù)據(jù)規(guī)模逐漸增大時(shí)，可能導(dǎo)致算法速度急劇下降，無(wú)法及時(shí)產(chǎn)生推薦。冷啟動(dòng)（Coldstart）：對(duì)于新加入的系統(tǒng)的項(xiàng)目或用戶，由于缺乏評(píng)價(jià)資源，沒(méi)辦法進(jìn)行推薦。稀疏性（Datasparsity）：收集到的用戶數(shù)據(jù)有限，系統(tǒng)的性能和準(zhǔn)確性較低。2021/7/8預(yù)備知識(shí)1主要內(nèi)容基于用戶的算法2基于信息熵的協(xié)同過(guò)濾算法3實(shí)驗(yàn)結(jié)果4參考文獻(xiàn)5基于用戶的算法2021/7/8基于用戶的算法2021/7/8基于用戶的算法2021/7/82021/7/8預(yù)備知識(shí)1主要內(nèi)容基于用戶的算法2基于信息熵的協(xié)同過(guò)濾算法3實(shí)驗(yàn)結(jié)果4參考文獻(xiàn)5假設(shè)間隔2021/7/8模型構(gòu)建2021/7/8模型構(gòu)建2021/7/82021/7/8預(yù)備知識(shí)1主要內(nèi)容基于用戶的算法2基于信息熵的協(xié)同過(guò)濾算法3實(shí)驗(yàn)結(jié)果4參考文獻(xiàn)5數(shù)據(jù)集及度量指標(biāo)2021/7/8數(shù)據(jù)集（數(shù)據(jù)可在下載）：選用公開(kāi)的MovieLens數(shù)據(jù)集（ML數(shù)據(jù)集）和HetRec2011-MovieLens數(shù)據(jù)集（HML數(shù)據(jù)集）對(duì)算法有效性進(jìn)行驗(yàn)證。其中，ML數(shù)據(jù)集包含943個(gè)用戶在1682個(gè)電影上的100000條評(píng)分記錄，每個(gè)用戶至少有20條評(píng)分記錄，評(píng)分矩陣的稀疏等級(jí)為0.9370。鑒于HML數(shù)據(jù)集中評(píng)分記錄過(guò)多，隨機(jī)選擇了其中的404個(gè)用戶在1300個(gè)電影上的39259條評(píng)分記錄，每個(gè)用戶至少有3條評(píng)分記錄，評(píng)分矩陣的稀疏等級(jí)為0.9254。度量指標(biāo)：采用推薦算法中常見(jiàn)的平均絕對(duì)偏差(Mean實(shí)驗(yàn)結(jié)果預(yù)測(cè)結(jié)果優(yōu)化用戶間的相似性大小和信息熵差異對(duì)推薦結(jié)果的影響2021/7/8實(shí)驗(yàn)結(jié)果算法比較2021/7/8［1］ADOMAVICIUSG,TUZHILIN

A.

Toward

the

next

generation

of

mender

systems:

a

survey

of

the

state-of-the-art

and

possibleextensions

[J].

IEEE

Transactions

on

Knowledge

and

Data

Engineering,

2005,

17(6):

734-749.［2］LIYanen,

ZHAI

Chengxiang,

ChenYe.

Exploiting

rich

user

information

for

one-class

collaborative

filtering

[J].

Knowledge

andInformation

Systems,

2014,

38(2):

277-301.［3］SHI

Yue,

LARSON

M,

HANJALIC

A.

Collaborative

filtering

beyond

the

user-item

matrix:

A

survey

of

the

state

of

the

art

and

futurechallenges

[J].

ACM

Computing

Surveys,

2014,

47(1):

3:1-3:45.［4］BOBADILLA

J,

ORTEGA

F,

HEMANDO

A,

et

al.

mender

systems

survey

[J].

Knowledge-Based

Systems,

2013,

46:

109-132.［5］BREESE

J

S,

HECKERMAN

D,

KADIE

C.

Empirical

Analysis

of

Predictive

Algorithms

for

Collaborative

Filtering[C]//Proc

of

the

14thConference

on

Uncertainty

in

Artificial

Intelligence.

Madison,

USA:

UAI,

1998:

43-52.［6］RESNICK

P,

IACOVOU

N,

SUCHAK

M,

et

al.

GroupLens

:

An

open

architecture

for

collaborative

filtering

of

netnews[C]//Proc

of

theACM

Conference

on

Computer

Supported

Cooperative

Work.

Chapel

Hill,

USA:

ACM,

1994:

175-186.［7］JAMALI

M,

ESTER

M.

TrustWalker:

a

random

walk

model

for

combining

trust-based

and

item-based

mendation[C]//Proc

of

the

15thACM

SIGKDD

International

Conference

on

Knowledge

Discovery

and

Data

Mining.

New

York,

USA:

ACM,

2009:

397-406.［8］SARWAR

B,

KARPIS

G,

KONSTAN

J,

et

al.

Item-based

collaborative

filtering

mendation

algorithms[C]//Proc

of

the

10th

InternationalConference

on

WorldWide

Web.

New

York,

USA:

ACM,

2001:

285–295.［9］AHN

H

J.

A

new

similarity

measure

for

collaborative

filtering

to

alleviate

the

new

user

cold-starting

problem

[J].

Information

Sciences,2008,

178(1):

37–51.［10］LIU

Haifeng,

HU

Zheng,

MIAN

Ahmad,

et

al.

A

new

user

similarity

model

to

improve

the

accuracy

of

collaborative

filtering

[J].Knowledge-Based

Systems,

2014,

56:

156-166.［11］李改,李磊.一種解決協(xié)同過(guò)濾系統(tǒng)冷啟動(dòng)問(wèn)題的新算法[J].山東大學(xué)學(xué)報(bào):工學(xué)版,2012,42(2):11-17.LI

Gai,

LI

Lei.

A

new

algorithm

of

cold-start

in

a

collaborative

filtering

system

[J].

Journal

of

Shandong

University:

Engineering

Science,

2012,42(2):

11-17.［12］鄧曉懿,金淳,韓慶平,等.基于情境聚類和用戶評(píng)級(jí)的協(xié)同過(guò)濾推薦模型[J].系統(tǒng)工程理論與實(shí)踐,2013,33(11):2945-2953.DENG

Xiaoyi,

JIN

Chun,

HAN

Qingping,

et

al.

Improved

collaborative

filtering

model

based

on

context

clustering

and

user

raking

[J].

SystemsEngineering—Theory

&

Practice,

2013,

33(11):

2945-2953.［13］鄧愛(ài)林,朱揚(yáng)勇,施伯樂(lè).基于項(xiàng)目評(píng)分預(yù)測(cè)的協(xié)同過(guò)濾推薦算法[J].軟件學(xué)報(bào),2003,14(9):1621-1628.DENG

Ailin,

ZHU

Yangyong,

SHI

Bole.

A

collaborative

filtering

mendation

algorithm

based

on

item

rating

prediction

[J].

Journal

of

Software,2003,

14(9):

1621-1628.［14］林耀進(jìn),胡學(xué)鋼,李慧宗.基于用戶群體影響的協(xié)同過(guò)濾推薦算法[J].情報(bào)學(xué)報(bào),2013,32(3):299-305.LIN

Y2a0o2ji1n,/7H/8U

Xuegang,

LI

Huizong.

Collaborative

filtering

mendation

algorithm

based

on

user

group

influence

[J].

Journal

of

The

ChinaSociety

For

Scientific

andTechnical

Information,

2013,

32(3):

299-305.參考文獻(xiàn)參考文獻(xiàn)2021/7/8［15］JEONG

B,

LEE

J,

CHO

H.

User

credit-based

collaborative

filtering

[J].

Expert

Systems

with

Applications,

2009,

36(6):

7309-7312.［16］ANAND

D,

BHARADWAJ

K

K.

Utilizing

various

sparsity

measures

for

enhancing

accuracy

of

collaborative

mender

systems

based

onlocal

and

global

similarities

[J].

Expert

Systems

with

Applications,

2011,

38

(5):

5101–5109.［17］BOUMAZA

A,

BRUN

A.

Stochastic

search

for

global

neighbors

selection

in

collaborative

filtering[C]//Proc

of

the

27th

AnnualACMSymposium

on

Applied

Computing.

New

York,

USA:

ACM,

2012:

232–237.［18］黃創(chuàng)光,印鑒,汪靜,等.不確定近鄰的協(xié)同過(guò)濾推薦算法[J].計(jì)算機(jī)學(xué)報(bào),2010,33(8):1369-1377.HUANG

Chuangguang,

YIN

Jian,

WANG

Jing,

et

al.

Uncertain

neighbors’

collaborative

filtering

mendation

algorithm

[J].

Chinese

Journal

ofComputers,

2010,

33(8):

1369-1377.［19］李聰,梁昌勇,馬麗.基于領(lǐng)域最近鄰的協(xié)同過(guò)濾推薦算法[J].計(jì)算機(jī)研究與發(fā)展,2008,45(9):1532-1538.LI

Cong,

LIANG

Changyong,

MA

Li.

A

collaborative

filtering

mendation

algorithm

based

on

domain

nearest

neighbor

[J].

Journal

of

ComputerResearch

and

Development,

2008,

45(9):

1532-1538.［20］LIU

Qi,

CHEN

Enhong,

XIONG

Hui,

et

al.

Enhancing

Collaborative

Filtering

by

User

Interest

Expansion

via

Personalized

Ranking

[J].IEEE

Transactions

on

Systems,

Man,

and

Cybernetics,

2012,

42(1):

218-233.［21］BOBADILLA