SALZA: Soft algorithmic complexity estimates for clustering and causality inference

TL;DR

This paper introduces practical estimators for various algorithmic complexities and directed information, enabling improved clustering and causality inference, with performance comparable to existing methods like NCD.

Contribution

It presents new estimators for conditional, simple, and joint algorithmic complexities, along with directed information estimators for causality inference, advancing the state-of-the-art in these areas.

Findings

Estimators perform well compared to NCD.

New directed information estimators effectively infer causality.

Method enhances clustering and causality analysis.

Abstract

A complete set of practical estimators for the conditional, simple and joint algorihmic complexities is presented, from which a semi-metric is derived. Also, new directed information estimators are proposed that are applied to causality inference on Directed Acyclic Graphs. The performances of these estimators are investigated and shown to compare well with respect to the state-of-the-art Normalized Compression Distance (NCD).