User Interests Identification on Twitter Using a Hierarchical Knowledge Base

Document Type

Conference Proceeding

Publication Date


Find in a Library

Catalog Record


Twitter, due to its massive growth as a social networking platform, has been in focus for the analysis of its user generated content for personalization and recommendation tasks. A common challenge across these tasks is identifying user interests from tweets. Semantic enrichment of Twitter posts, to determine user interests, has been an active area of research in the recent past. These approaches typically use available public knowledge-bases (such as Wikipedia) to spot entities and create entity-based user profiles. However, exploitation of such knowledge-bases to create richer user profiles is yet to be explored. In this work, we leverage hierarchical relationships present in knowledge-bases to infer user interests expressed as a Hierarchical Interest Graph. We argue that the hierarchical semantics of concepts can enhance existing systems to personalize or recommend items based on a varied level of conceptual abstractness. We demonstrate the effectiveness of our approach through a user study which shows an average of approximately eight of the top ten weighted hierarchical interests in the graph being relevant to a user’s interests.


Presented at the 11th International Conference on the Semantic Web: Trends and Challenges, Crete, Greece, May 25-29, 2014.

Presentation slides can be viewed at

Video of the presentation can be found at



Catalog Record