Learning from Sparse Data by Exploiting Monotonicity Constraints

TL;DR

This paper introduces a method to incorporate qualitative monotonicity constraints into Bayesian network learning algorithms, improving accuracy when training data is sparse.

Contribution

It formalizes the interpretation of monotonicity knowledge as probability constraints and integrates this into Bayesian network learning, enhancing performance with limited data.

Findings

Improved accuracy with small training sets

Effective incorporation of qualitative monotonicity constraints

Enhanced Bayesian network learning performance

Abstract

When training data is sparse, more domain knowledge must be incorporated into the learning algorithm in order to reduce the effective size of the hypothesis space. This paper builds on previous work in which knowledge about qualitative monotonicities was formally represented and incorporated into learning algorithms (e.g., Clark & Matwin's work with the CN2 rule learning algorithm). We show how to interpret knowledge of qualitative influences, and in particular of monotonicities, as constraints on probability distributions, and to incorporate this knowledge into Bayesian network learning algorithms. We show that this yields improved accuracy, particularly with very small training sets (e.g. less than 10 examples).