查看文章

Logical Analysis of Data (LAD) is a well-known classification technique that generates interpretable patterns with competitive accuracy. The challenge encountered in applying LAD comes from its long computational time, which makes it unsuitable for handling a large volume of data. In this paper, we propose a novel mechanism for developing an ensemble system for LAD (LAD-ENS) to improve its computational efficiency, while preserving its interpretability and promising accuracy. This new mechanism aims to maintain the explanatory power of classical LAD by combining the individual classifiers at the level of patterns. The developed ensemble system enables LAD to be run in parallel computing environments. Using datasets obtained from the UCI Machine Learning Repository, computational experiments are conducted to demonstrate the performance of LAD-ENS in terms of computational time, classification accuracy, and interpretability. Furthermore, we introduce the concept of the comprehensibility index in order to study the change in the explanatory power of LAD. In addition to achieving a statistically significant reduction in computational time, the developed LAD-ENS achieves competitive classification accuracies compared to two classical LAD approaches and five common machine learning algorithms.