Real-time news recommendation with rich representation thumbnail
Pause
Mute
Subtitles
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Real-time news recommendation with rich representation

Published on Nov 07, 20133299 Views

News recommendation is an area of research where we deal with a non-stationary source of documents which are recommended to the users of the publishers' web sites. Predominant success metric is the at

Related categories

Chapter list

News Recommendation00:00
Outline00:27
Motivation01:14
Reuters01:18
Why news recommendation?02:54
Why news recommendation is hard?03:37
Example: Bloomberg.com - 105:05
Example: Bloomberg.com - 205:24
Approaches05:49
News Recommender Architecture - 106:37
News Recommender Architecture - 206:41
News Recommender Architecture - 306:44
News Recommender Architecture - 406:45
News Recommender Architecture - 506:49
News Recommender Architecture - 606:50
News Recommender Architecture - 706:52
Architecture Considerations07:38
Data Collection08:50
Example - 110:23
Example - 211:11
Sample User profile12:47
Recommendation Algorithm - 113:20
Recommendation Algorithm - 214:28
Recommendation Algorithm - 314:30
Features14:44
Content Features - 115:03
What do we know about an article?15:05
What can be derived from content?15:28
Extracting triples from text and linking to LOD15:52
Content Features - 218:01
Collaborative Features - 118:06
Article Popularity18:09
Co-visited articles18:13
Article readers18:33
Collaborative Features - 218:36
Demographic Features18:49
User Data (NYTimes) - 118:51
User Data (NYTimes) - 218:53
User Modeling19:28
Estimating user’s “Gender”19:32
Estimating user’s “Age”20:04
Estimating user’s “Income” (≥10 visits)20:16
Evaluation20:29
Performance comparison20:31
Which features matter? - 121:17
Which features matter? - 221:23
Conclusions22:01
What is cross-lingual recommendation?22:30
Bloomberg23:27
How CL-recommendations are being done?23:56
Content acquisition from RSS feeds25:43
Capturing publisher’s style “Bloombergness”26:19
Comparing publishing styles26:49
Empirical evaluation of CL-Recommendation27:33
News recommendation for social media?28:12
What is news recommendation for social media?28:46
Prototype: Relevant Bloomberg articles for different countries29:01