Search papers, labs, and topics across Lattice.
This paper introduces a novel approach to learning Random Geometric Graphs (RGGs) from multivariate datasets within probabilistic metric spaces, allowing for flexibility across various data types and distributions. The authors define a new distance function based on a disparity variable that captures the connectedness of graph vertices, leading to the concept of Soft RGGs where edges exist with a certain probability. Through a Rejection Sampling technique, they demonstrate the ability to learn edge probabilities and the expected degree distribution, validated with multiple real-world datasets.
Learning RGGs in probabilistic metric spaces reveals that edges can exist with a defined probability, transforming how we understand graph connectivity in complex datasets.
We present a new data-driven learning of a Random Geometric Graph (RGG) of a multivariate dataset, where the graph is drawn in a probabilistic metric space. This graph learning works for generic datasets, irrespective of the type of the observables; their probability distributions; or size of the data. We identify a metric of the space that the graph is drawn in, as a probability distribution of a random variable that we introduce, namely, a variable that represents the disparity between the connectedness of two vertices of the graph, and the correlation between the two random variables that are attached to the respective vertex. It is the closed-form {\it{cdf}} of this disparity variable that we advance as the distance function of the host space of the learnt RGG, such that the edge exists between any two nodes, if this inter-nodal distance falls short of a chosen cutoff probability. Drawing the RGG in this probabilistic space leads to the graph being an Soft RGG, such that any edge - if it exists - exists with an identified probability. We forward a simple Rejection Sampling-based technique for learning the probability of any edge. The expected degree distribution of a vertex of this RGG is identified as local, and dependent on the inter-observable correlation matrix. If said correlation matrix is not known, it can be learnt given the data, using its closed-form posterior probability density function, that we forward. We illustrate our graph learning method by learning multiple RGGs of highly multivariate real datasets.
