Abstract

News Recommendation shapes public perception of events. The specific challenge in developing news recommender algorithms are the continuous changes in the set of relevant items, the diversity of the items, and the context-dependent user preferences. Different news portals (publishers) have different audience with different habits and preferences. For better news recommendations, recommender systems need specific adjustments shaped according to their audience. In this work we aim to explore the different usage manners based on contextual data for different publishers. We mine characteristic patterns and discuss how these findings must be considered when developing recommendation strategies.

@inproceedings{...,
author = {Benjamin Kille and "Ozlem "Ozg"obek and Andreas Lommatzsch},
title = {Defining Contextual Factors for News Consumption},
booktitle = {Proceedings of the ACM CIKM 2018 Workshops},
year = {2018},
isbn = {ISSN 1613-0073},
numpages = {7},
location = {, Turin, Italy},
doi = {},
publisher = {CEUR Workshop Proceedings},
address = {http://ceur-ws.org/},
}