ACM Conference on Recommender Systems 2012 – Poster slam

September 11, Dublin, Ireland

Improvement of user-based CF

by using spectral clustering techniques

ACM Conference on Recommender Systems 2012 – Poster slam

September 11, Dublin, Ireland

Context: cluster-based CF

neighbours identified based on a clustering method

Spectral Clustering: Normalised Cut

ACM Conference on Recommender Systems 2012 – Poster slam

September 11, Dublin, Ireland

Context: cluster-based CF

neighbours identified based on a clustering method

Spectral Clustering: Normalised Cut

Better than k-Means: performance

0.00

0.02

0.04

0.06

0.08

0.10

0.12

50 100 150 200 250 300 350 400 450 500 550 600

P

@

5

k

k-Means NCut

ACM Conference on Recommender Systems 2012 – Poster slam

September 11, Dublin, Ireland

Context: cluster-based CF

neighbours identified based on a clustering method

Spectral Clustering: Normalised Cut

Better than k-Means: coverage

0

10

20

30

40

50

60

70

80

90

100

50 100 150 200 250 300 350 400 450 500 550 600

C

o

v

e

r

a

g

e

k

cov(k-Means) cov(Ncut)

ACM Conference on Recommender Systems 2012 – Poster slam

September 11, Dublin, Ireland

Context: cluster-based CF

neighbours identified based on a clustering method

Spectral Clustering: Normalised Cut

Better than k-Means

Also better than MF and standard UB

0.00

0.02

0.04

0.06

0.08

0.10

0.12

50 100 150 200 250 300 350 400 450 500 550 600

P

@

5

k

UB50 MF NCut

ACM Conference on Recommender Systems 2012 – Poster slam

September 11, Dublin, Ireland

Using Graph Partitioning Techniques

for Neighbour Selection in

User-Based Collaborative Filtering

Alejandro Bellogín Javier Parapar

Information Retrieval Group Information Retrieval Lab

Universidad Autónoma de Madrid University of A Coruña