Network Lens: Node Classification in Topologically Heterogeneous Networks
This addresses the problem of identifying diverse behaviors in large heterogeneous networks for network analysis applications, though it appears incremental in approach.
The paper tackles node classification in topologically heterogeneous networks by using variable-sized 'lenses' to capture local structure and weighting network signatures for predictions, achieving peak accuracy of ~42% (vs. 11% random) on networks with ~100,000 and ~1,000,000 nodes.
We study the problem of identifying different behaviors occurring in different parts of a large heterogenous network. We zoom in to the network using lenses of different sizes to capture the local structure of the network. These network signatures are then weighted to provide a set of predicted labels for every node. We achieve a peak accuracy of $\sim42\%$ (random=$11\%$) on two networks with $\sim100,000$ and $\sim1,000,000$ nodes each. Further, we perform better than random even when the given node is connected to up to 5 different types of networks. Finally, we perform this analysis on homogeneous networks and show that highly structured networks have high homogeneity.