Non-negative Factorization of the Occurrence Tensor from Financial Contracts

TL;DR

This paper introduces a non-negative tensor factorization algorithm tailored for analyzing occurrence data in financial networks, effectively handling sparse errors and uncovering embedded group structures.

Contribution

The paper presents a novel non-negative tensor factorization method using l0 norm for sparse errors, with an efficient optimization approach for nonconvex problems, applied to financial document data.

Findings

Effective in modeling sparse errors in occurrence tensors

Successfully applied to financial document dataset resMBS

Demonstrates robustness on synthetic and real data

Abstract

We propose an algorithm for the non-negative factorization of an occurrence tensor built from heterogeneous networks. We use l0 norm to model sparse errors over discrete values (occurrences), and use decomposed factors to model the embedded groups of nodes. An efficient splitting method is developed to optimize the nonconvex and nonsmooth objective. We study both synthetic problems and a new dataset built from financial documents, resMBS.