Iterative procedure for network inference

TL;DR

This paper presents an iterative method to accurately reconstruct the degree distribution of a network from data, accounting for false positive and false negative errors in link detection.

Contribution

It introduces an iterative procedure that refines network topology estimates by adjusting error probabilities, improving the accuracy of network inference from data.

Findings

Successfully reconstructs vertex degree distribution.

Adjusts for type I and II errors in network inference.

Provides a systematic iterative approach for network topology estimation.

Abstract

When a network is reconstructed from data, two types of errors can occur: false positive and false negative errors about the presence or absence of links. In this paper, the vertex degree distribution of the true underlying network is analytically reconstructed using an iterative procedure. Such procedure is based on the inferred network and estimates for the probabilities $\alpha$ and $\beta$ of type I and type II errors, respectively. The iteration procedure consists of choosing various values for $\alpha$ to perform the iteration steps of the network reconstruction. For the first step, the standard value for $\alpha$ of 0.05 can be chosen as an example. The result of this first step gives a first estimate of the network topology of interest. For the second iteration step the value for $\alpha$ is adjusted according to the findings of the first step. This procedure is iterated, ultimately leading to a reconstruction of the vertex degree distribution tailored to its previously unknown network topology.