| Article ID | Journal | Published Year | Pages | File Type |
|---|---|---|---|---|
| 4960920 | Procedia Computer Science | 2017 | 7 Pages |
Abstract
In this paper, in order to improve IPTV recommendation system, a Spark-based IPTV implicit feedback scoring model is proposed based on the characteristics of user viewing behaviors. Firstly, this paper analyzes the data types and characteristics of IPTV user viewing behaviors. Secondly, this paper presents a new multi-features hybrid implicit feedback scoring model framework based on viewing duration and viewing ratio. Then, we design and implement the distributed multi-features hybrid implicit feedback scoring model on Spark cluster. Finally, we figure out reasonable weights of different features of implicit feedback through experiment. And the experiment shows that the Spark-based model performs better in time efficiency.
Keywords
Related Topics
Physical Sciences and Engineering
Computer Science
Computer Science (General)
Authors
Yue Jin, Gu Junhua, Zhang Suqi, Zhang Jian,