A new model of selecting most relevant ratings in recommender systems,Serhiy Morozov,Hossein Saiedian

A new model of selecting most relevant ratings in recommender systems


	BibTex \| RIS \| RefWorks Download

A major assumption of collaborative filtering is that similar users will always agree on a majority of items, regardless of their domain. This concept establishes strong connections among neighbors. However, it eliminates potentially good users on the premise that they are not similar enough. Furthermore, this assumption allows for the possibility of a neighbor to be chosen simply because he/she shares a lot of similar ratings in unrelated domains and offers little useful information in the active item domain. This effectively reduces the amount of useful information that is considered for each recommendation. We propose a new way to identify relevant ratings that relies on somewhat weaker, but more abundantly available neighbors.

Published in 2010.

Cumulative Annual

View Publication

( ieeexplore.ieee.org )

Keywords (2)

A new model of selecting most relevant ratings in recommender systems,Serhiy Morozov,Hossein Saiedian

References (10)

Scalable Collaborative Filtering with Jointly Derived Neighborhood Interpolation Weights (Citations: 79)

Decentralized Mediation of User Models for a Better Personalization (Citations: 8)

Cross-Domain Mediation in Collaborative Filtering (Citations: 12)

Empirical Analysis of Predictive Algorithms for Collaborative Filtering (Citations: 1477)

Comparing State-of-the-Art Collaborative Filtering Systems (Citations: 18)


	The following links allow you to view full publications. These links are maintained by other sources not affiliated with Microsoft Academic Search.