Optimising the Attribute Order in Fuzzy Rough Rule Induction

2025-06-03

IJCRS

AI Summary

This paper investigates the impact of attribute ordering on interpretable rule learning in Fuzzy Rough Rule Induction (FRRI). We find that optimizing the attribute removal sequence alone yields no significant improvement in classification accuracy or rule conciseness. However, integrating a lightweight feature screening step—grounded in fuzzy rough set theory and combining fuzzy indiscernibility relations with a greedy strategy—prior to rule induction effectively balances accuracy and average rule length. By pre-selecting only a small subset of critical attributes, this approach avoids computationally expensive search procedures. Empirical evaluation across multiple metrics demonstrates that our method substantially outperforms both the original FRRI and classical feature selection baselines. The key contribution is twofold: (1) revealing that attribute ordering is not a dominant factor in FRRI performance, and (2) proposing an efficient “pre-screening + rule induction” paradigm that synergistically decouples feature selection from rule generation.

Technology Category

• Data Mining & Knowledge Management: Rule Mining & Pattern Mining
• Machine Learning: Dimensionality Reduction/Feature Selection
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

Interpretability is the next pivotal frontier in machine learning research. In the pursuit of glass box models - as opposed to black box models, like random forests or neural networks - rule induction algorithms are a logical and promising avenue, as the rules can easily be understood by humans. In our previous work, we introduced FRRI, a novel rule induction algorithm based on fuzzy rough set theory. We demonstrated experimentally that FRRI outperformed other rule induction methods with regards to accuracy and number of rules. FRRI leverages a fuzzy indiscernibility relation to partition the data space into fuzzy granules, which are then combined into a minimal covering set of rules. This indiscernibility relation is constructed by removing attributes from rules in a greedy way. This raises the question: does the order of the attributes matter? In this paper, we show that optimising only the order of attributes using known methods from fuzzy rough set theory and classical machine learning does not improve the performance of FRRI on multiple metrics. However, removing a small number of attributes using fuzzy rough feature selection during this step positively affects balanced accuracy and the average rule length.

Problem

• Optimizing attribute order in fuzzy rough rule induction
• Improving interpretability of machine learning models
• Enhancing rule induction performance via feature selection

Innovation

• Fuzzy rough rule induction for interpretability
• Optimizing attribute order with fuzzy rough
• Feature selection improves accuracy and rule length