Liang Xiang, Quan Yuan, Shiwan Zhao, Li Chen, Xiatian Zhang, Qing Yang and Jimeng Sun Institute of...

Liang Xiang, Quan Yuan, Shiwan Zhao, Li Chen,Xiatian Zhang, Qing Yang and Jimeng Sun

Institute of Automation Chinese Academy of Sciences, IBM Research Department of Computer Science, Hong Kong Baptist11University

IBM T.J. Watson Research Center

Temporal Recommendation on Graphs via Long- and Short-term Preference Fusion

KDD 2010

Intelligent Database Systems Lab.

School of Computer Science & Engineering

Seoul National University

Center for E-Business TechnologySeoul National UniversitySeoul, Korea

Presented by Sung Eun, Park11/04/2010

Copyright 2010 by CEBT

Contents

Introduction

STG Construction

Recommendation on STG

Injected Preference Fusion

Temporal Personalized Random Walk

Complexity Analysis

Experiment

Conclusion

2


Introduction

Motivation

Overall behavior of a user may be determined by her long-term interest

But at any given time, a user is also affected by her short-term interest due to transient events, such as new product releases and special personal occasions such as birthdays

Goal

Explicitly model users’ long-term and short-term preferences for recommendation on implicit datasets

1. How to represent the two types of user preferences?

2. How to model the interaction between the two types of preferences?

3


STG Construction

Transform “<user, item, time> triples” into “<user, item> and <session, item>” by dividing the time into bins and binding the bins with corresponding users

Session-based Temporal Graph (STG)

a directed bipartite graph G(U, S, I,E,w)

where U denotes the set of user nodes, S is the set of session nodes, and I is the set of item nodes, E is the set of edges and weight

– Session node

Dividing the time slices into bins

and binding the bins with corresponding

users

4


STG Construction

item1

item2

item3

item4

user1

1 1 1 0

user2

0 1 0 1

5

u1

u2

i1

i2

i3

i4

Session2 (u1,T2)

Session1 (u1,T1)

Session3 (u2,T0)

u’s long-term preferences

user node vu connects to all items viewed by user u, denoted as N(u)

u’s short-term preferences

vut only connects to items user u viewed at time t, denoted as N (u, t)

if we start walking from the user node vu , we will pass through N(u) and then reach unknown items which are similar to items in N(u)

if we walk from the session node vut , we will reach unknown items similar to items in N(u, t)


Injected Preference Fusion(IPF)

6

An algorithm based on STG, which balancing the impact of long-term and short-term preferences when making recommendation

Basic Idea

Injected preferences into both user node(β) and session node(1-β)

Then in propagation process, the preferences were propagated to an unknown item node

Finally the nodes which get most preferences will be recommended to current user


Injected Preference Fusion(IPF)

7

u1

u2

i1

i2

i3

i4

Session2 (u1,T2)

Session1 (u1,T1)

Session3 (u2,T0)

Preference β Preference 1-β

user item session

ηuηs

Making recommendation for U1 at time T1 : Session Temporal Graph(STG)

STG edge weight definition

1 1


Datasets(Bookmarking)

Evaluation Metric

Hit Ratio : put the latest item of each user into test set, then generate a list of N(N=10)items for everyone at time t. If the test item appears in the recommendation list, we call it hit

Compared Algorithms

Temporal User KNN

Temporal Item KNN

Temporal Personalized Random walk

Injected Preferences Fusion ( IPF )

8

CiteULIke Delicious

User 4607 8,861

Item 16,054 3,257

User-item pair 109,346 59,694

Sparsity 99.85% 99.97%


Experiments

β‘s impact-balance the injected preferences on user and session node

Optimal results were get when β belongs to [0.1,0.6]

Proves the impacts of both long-term and short-term preferences in making good recommendation

β=0, only short-term

β =1, onlylong-term

β =0, only short-term

β =1, onlylong-termCiteULike Delicious


Experiments

η‘s impact-balance the injected preferences on user and session node

Proves that effectiveness of balancing long-term and short-term preferences in propagation process

Since X-axis is the logarithm value, it means the optimal hit ratio can be get for a wide range of η

η=0, item-itemconnected only by session

CiteULike Deliciousη=1, item-itemconnected only by user

η=0, item-itemconnected only by session

η=1, item-itemconnected only by user


Session size‘s impact on Hit Ratio of IPF

The result is not very sensitive to the size of time window

Users’ interest on research topics(CiteULike)drift more slowly than interests on browsing web pages(Delicious)

Experiments


Overall Accuracy Comparison

– TItemKNN : Time weighted item-based CF

– TUserKNN : Time weighted user-based DF

– TPRW : Temporal Personalized Random Walk

– MS-IPF : Multi Source IPF

User Temporal Item KNN as baseline

On CiteULike, MS-IPF improves TItemKNN up to 15.02%

On Delicious, MS-IPF improves TItemKNN up to 34.45%

Experiments

CiteULike Delicious


Conclusion

propose a Session-based Temporal Graph (STG) which incorporates temporal information to model long-term and short-term preferences simultaneously

Based on STG framework, we propose a new algorithm, Injected Preference Fusion (IPF), to balance the impacts of users’ long-term and short-term preferences for accurate recommendation.

To further demonstrate the model generality, we extend the personalized Random Walk for temporal recommendation based on STG.

Q&A

Thank you

14

Liang Xiang, Quan Yuan, Shiwan Zhao, Li Chen, Xiatian Zhang, Qing Yang and Jimeng Sun Institute of...

Documents

Transcript of Liang Xiang, Quan Yuan, Shiwan Zhao, Li Chen, Xiatian Zhang, Qing Yang and Jimeng Sun Institute of...