WEB RECOMMENDATION SYSTEM BASED ON A MARKOV-CHAIN MODEL
François Fouss, Stéphane Faulkner, Manuel Kolp, Alain Pirotte, Marco Saerens · 2005
This work presents some general procedures for computing dissimilarities between nodes of a weighted, undirected, graph. It is based on a Markov-chain model of random walk through the graph. This method is applied on the architecture of a Multi Agent System (MAS), in which each agent can be considered as a node and each interaction between two agents as a link. The model assigns transition probabilities to the links between agents, so that a random walker can jump from agent to agent. Two quantities, called the average commute time and the pseudoinverse of the Laplacian matrix of the graph, provide proximity measures between any pair of agents. The model is applied on a collaborative filtering task where suggestions are made about which movies people should watch based upon what they watched in the past. For the experiments, we build a MAS architecture and instantiated the agents belief set from a real movie database.