Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Harnessing the Real-Time Web for Personalized News Recommendation

4715c0947b4e0ca3bec820d8051aa45a?s=47 Gianmarco De Francisci Morales
February 09, 2012
Research
2
140

Harnessing the Real-Time Web for Personalized News Recommendation

Presentation of my article at WSDM'12

4715c0947b4e0ca3bec820d8051aa45a?s=128

Gianmarco De Francisci Morales

February 09, 2012
Tweet

More Decks by Gianmarco De Francisci Morales

See All by Gianmarco De Francisci Morales
4715c0947b4e0ca3bec820d8051aa45a?s=48 gdfm
0
8
4715c0947b4e0ca3bec820d8051aa45a?s=48 gdfm
0
38
4715c0947b4e0ca3bec820d8051aa45a?s=48 gdfm
0
100
4715c0947b4e0ca3bec820d8051aa45a?s=48 gdfm
0
68
4715c0947b4e0ca3bec820d8051aa45a?s=48 gdfm
2
230
4715c0947b4e0ca3bec820d8051aa45a?s=48 gdfm
1
200
4715c0947b4e0ca3bec820d8051aa45a?s=48 gdfm
0
410
4715c0947b4e0ca3bec820d8051aa45a?s=48 gdfm
2
530
4715c0947b4e0ca3bec820d8051aa45a?s=48 gdfm
1
110

Other Decks in Research

See All in Research
753eb5a167cf43a033413f08e63f3632?s=48 osx
0
130
27e6303669f854882d672f3cd3fcb796?s=48 dasayan05
0
440
0d4ef9af8e4f0cf5c162b48ba24faea6?s=48 ept
0
14k
Bb6c3fc8c577710c72d03aeb4fa56bf6?s=48 masakat0
0
110
608638a6cfefbf0c2294a30e55634f6d?s=48 mihozono
2
1.6k
38e2af7f8bdad4f2087ab3d42b627e33?s=48 shuntaros
0
650
A8e5603efe2e904e5643163d1e0cb0d6?s=48 butsugiri
39
11k
Fba2bc6aeb14d00ab84710be1903da49?s=48 taishi
0
170
C612ab39597a17ba5948cae54d13f99f?s=48 mkimura
2
1.5k
76b8d066c4a1b9b6624abbf1544ed7ba?s=48 higgi13425
0
470
8a7e83d2e447783ab6d824f553429a09?s=48 stakaya
0
1.2k
Cd3d2cb2dadf5488935fe0ddaea7938a?s=48 monochromegane
1
1.2k

Featured

See All Featured
0c6fd6a08d0f898159cda7cafcacf07f?s=48 afnizarnur
180
14k
Dafc4723e9a1c067796c0fec6f509774?s=48 lemiorhan
633
49k
B83d5b590577969ee79a5ad845411a7d?s=48 ammeep
660
55k
Bfd6b681ec6e8c9bef82ff0521c364f7?s=48 kastner
56
2k
7cbd3f5f922d798cc312eaf596de7ae6?s=48 iamctodd
28
2.8k
245cee81a9c424266e5e401d844ea881?s=48 lara
595
62k
25c7c18223fb42a4c6ae1c8db6f50f9b?s=48 mojombo
364
63k
8081b26e05bb4354f7d65ffc34cbbd67?s=48 chriscoyier
147
20k
61857dafbd287b3027c4dcea9008ad3c?s=48 carmenhchung
43
2k
C51b39fa3c79ccb99dcb8a076bc98287?s=48 brianwarren
87
5.3k
06f8b41980eb4c577fa40c41d5030c19?s=48 keathley
25
910
A60068bce2e73de3a37ca9d2dbe36092?s=48 bryan
105
11k

Transcript

  1. Harnessing the Real-Time Web for Personalized News Recommendation Gianmarco De

    Francisci Morales ✱ ✦✤ Aristides Gionis ✦ Claudio Lucchese ✤ ✱ IMT, Lucca ✤ ISTI-CNR, Pisa ✦ Yahoo! Research, Barcelona WSDM 2012

  2. The Problem Information Overload

  3. -0.2 0 0.2 0.4 0.6 0.8 1 1.2 M ay-01

    h20 M ay-02 h00 M ay-02 h04 M ay-02 h08 M ay-02 h12 M ay-02 h16 M ay-02 h20 M ay-03 h00 M ay-03 h04 M ay-03 h08 news twitter clicks The Motivation Twitter as source of news recommendation Number of mentions of “Osama Bin Laden”

  4. 0 2 4 6 8 10 12 14 1 2

    3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Average DCG Rank T.Rex+ T.Rex Popularity Content Social Recency Click count The Result Click prediction! Topic overlap between clicked and suggested news article