Network analysis is one of the main topics in modern data analysis, since it enables us to reason about systems by studying their inner relations, for example we can study a network by analyzing its edges. However, in many cases it is impossible to detect or measure the network directly, due to noisy data for example. We present a method for dealing with such systems, more concretely we present a probabilistic model called latent distance network, which we use to model news data from EventRegistry. In the end of the article we also present experimental results on predictions of latent distance model with methods of machine learning