Learning Web Page Scores by Error Back-Propagation

Learning Web Page Scores by Error Back-Propagation,Michelangelo Diligenti,Marco Gori,Marco Maggini

Learning Web Page Scores by Error Back-Propagation   (Citations: 17)
BibTex | RIS | RefWorks Download
In this paper we present a novel algorithm to learn a score distribution over the nodes of a labeled graph (directed or undirected). Markov Chain theory is used to define the model of a random walker that converges to a score distribution which depends both on the graph connectivity and on the node la- bels. A supervised learning task is defined on the given graph by assigning a target score for some nodes and a training algorithm based on error back- propagation through the graph is devised to learn the model parameters. The trained model can as- sign scores to the graph nodes generalizing the cri- teria provided by the supervisor in the examples. The proposed algorithm has been applied to learn a ranking function for Web pages. The experimental results show the effectiveness of the proposed tech- nique in reorganizing the rank accordingly to the examples provided in the training set.
Cumulative Annual
View Publication
The following links allow you to view full publications. These links are maintained by other sources not affiliated with Microsoft Academic Search.
Sort by: