Surprisingly popular-based multivariate conceptual knowledge acquisition method

Abstract

Formal Concept Analysis (FCA) plays a crucial role in uncertain artificial intelligence by revealing specialization and generalization relationships among concepts. Unlike artificial neural network methods, concepts are fundamental to human cognition and learning, making knowledge acquisition and decision-making methods highly significant, especially when influenced by data and cognition. However, existing FCA models and extensions predominantly emphasize data-driven approaches, often disregarding the cognitive attributes of individual learners. This paper introduces a novel method for acquiring diverse conceptual knowledge and a cognition-enhanced decision-making approach based on the Surprisingly Popular (SP) method. Initially, it defines a new multivariate relational formal context and its corresponding decision-making method to facilitate a more sophisticated exploration of uncertain information. Additionally, it presents a method for quantifying the similarity of multivariate concepts. The SP approach is then integrated to identify the core and peripheral multivariate concepts within the multivariate concept lattice. Furthermore, the paper develops a multivariate cognitive decision-making method and presents the corresponding algorithm. Finally, instance analysis is conducted on the UCI dataset to compare the proposed method with state-of-the-art models. The results indicate that the proposed model effectively uncovers core and peripheral concepts within uncertain information by incorporating human cognitive decision processes.