Towards Information Theory of Knowledge Networks

Preprint
Publisher Website
Google Scholar
  • 6 April 2001
Abstract
We develop a new type of Information Theory that exploits the affinity between people's tastes and their opinions on products, services, or each other. Individual tastes, modeled by an array, can be reconstructed based only on the information contained in a sparsely connected network of opinions. Two distinct phase transitions occur as the density of known elements is gradually increased: above the first transition the macroscopic prediction becomes possible, while above the second - all unknown opinions can be uniquely reconstructed. We study one particularly simple Gaussian model of such network using field theoretical methods augmented by numerical simulations.

This publication has 0 references indexed in Scilit: