Tuning Personalized PageRank for Semantics-aware Recommendations based on Linked Open Data
published: July 10, 2017, recorded: June 2017, views: 1438
Report a problem or upload filesIf you have found a problem with this lecture or would like to send us extra material, articles, exercises, etc., please use our ticket system to describe your request and upload the data.
Enter your e-mail into the 'Cc' field, and we will keep you updated with your request's status.
In this article we investigate how the knowledge available in the Linked Open Data cloud (LOD) can be exploited to improve the effectiveness of a semantics-aware graph-based recommendation framework based on Personalized PageRank (PPR).
In our approach we extended the classic bipartite data model, in which only user-item connections are modeled, by injecting the exogenous knowledge about the items which is available in the LOD cloud. Our approach works in two steps: first, all the available items are automatically mapped to a DBpedia node; next, the resources gathered from DBpedia that describe the item are connected to the item nodes, thus enriching the original representation and giving rise to a tripartite data model. Such a data model can be exploited to provide users with recommendations by running PPR against the resulting representation and by suggesting the items with the highest PageRank score.
In the experimental evaluation we showed that our semantics-aware recommendation framework exploiting DBpedia and PPR can overcome the performance of several state-of-the-art approaches. Moreover, a proper tuning of PPR parameters, obtained by better distributing the weights among the nodes modeled in the graph, further improved the overall accuracy of the framework and confirmed the effectiveness of our strategy.
Link this pageWould you like to put a link to this lecture on your homepage?
Go ahead! Copy the HTML snippet !