A simple counting estimator of network agents' behaviors: Asymptotics. Marano, S. & Willett, P. In 2016 24th European Signal Processing Conference (EUSIPCO), pages 1523-1527, Aug, 2016.
A simple counting estimator of network agents' behaviors: Asymptotics [pdf]Paper  doi  abstract   bibtex   
Recent works address the problem of estimating agents' behaviors in complex networks, of which social networks are a prominent example. Many of the proposed techniques work but at the cost of a substantial computational complexity, which is not permitted when dealing with big data real-time analysis. This raises the question of whether a very simple nonparametric counting estimator works in practical problems. We propose such an estimator and investigate its asymptotic properties for large number of agents N and/or for large network observation time T. The asymptotic optimality of the estimator is proven and computer experiments are provided to assess its performance for finite values of N and T.

Downloads: 0