[Paper Review] A construction of an optimal base for conditional attribute and attributional condition implications in triadic contexts

Romuald Kwessy Mouona, Blaise B. Koguep Njionou+3 more|arXiv (Cornell University)|Jan 4, 2026

Decision-Making and Behavioral Economics0 citations

TL;DR

The paper develops methods to construct optimal bases for conditional attribute implications (CAI) and attributional condition implications (ACI) in triadic contexts, using context augmentation and quasi-features to achieve minimal, complete bases.

ABSTRACT

This article studies implications in triadic contexts. Specifically, we focus on those introduced by Ganter and Obiedkov, namely conditional attribute and attributional condition implications. Our aim is to construct an optimal base for these implications.

Motivation & Objective

• Motivate the study of triadic implications and their expressive power in triadic contexts.
• Propose a framework to construct optimal bases for CAI and ACI implications.
• Introduce augmentation, quasi-features, and unit pseudo-features as tools for basis construction.
• Develop an algorithm for basis construction and analyze its computational complexity.

Proposed method

• Formalize triadic contexts and the types of triadic implications (BCAI, BACI, CAI, ACI).
• Introduce augmentation of triadic contexts and relate augmented-context concepts to original ones.
• Define quasi-features and unit pseudo-features as building blocks for bases.
• Propose complete bases for CAI and ACI based on quasi-features and unit pseudo-features.
• Provide an algorithm for constructing these bases and study its complexity.

Experimental results

Research questions

• RQ1 How can CAI and ACI implications be represented by complete, minimal bases in triadic contexts?
• RQ2 What role do quasi-features and unit pseudo-features play in generating optimal bases?
• RQ3 How does triadic context augmentation affect the derivation of implications and their bases?
• RQ4 What is the computational complexity of constructing these optimal bases?

Key findings

• An optimal base for CAI and ACI is constructed via unit pseudo-features and context augmentation.
• Propositions relate augmented contexts to original contexts and preserve feature sets.
• A complete base for CAI and a complete base for ACI are obtained using quasi-features and unit pseudo-features.
• An algorithm is proposed for constructing these bases and its complexity is analyzed.
• The work extends prior results on BCAI and BACI to CAI and ACI, with an emphasis on compact representation of implications.