Stochastic Patching Process

TL;DR

The paper introduces the Stochastic Patching Process, a novel partition model that efficiently identifies dense regions in multi-dimensional data arrays, improving relational modeling by reducing unnecessary data dissection.

Contribution

It proposes a new parsimonious partition model called SPP that uses an enclosing strategy for better data region detection and can be extended to infinite arrays.

Findings

SPP outperforms existing models in relational data tasks.

SPP effectively reduces unnecessary partitioning in sparse regions.

The model demonstrates strong theoretical properties like self-consistency.

Abstract

Stochastic partition models tailor a product space into a number of rectangular regions such that the data within each region exhibit certain types of homogeneity. Due to constraints of partition strategy, existing models may cause unnecessary dissections in sparse regions when fitting data in dense regions. To alleviate this limitation, we propose a parsimonious partition model, named Stochastic Patching Process (SPP), to deal with multi-dimensional arrays. SPP adopts an "enclosing" strategy to attach rectangular patches to dense regions. SPP is self-consistent such that it can be extended to infinite arrays. We apply SPP to relational modeling and the experimental results validate its merit compared to the state-of-the-arts.