Two-way Concept-Cognitive Learning with Multi-source Fuzzy Context

Concepts learning is the most fundamental unit in the process of human cognition in philosophy. Granularity is one of the fundamental concepts of human cognition. The combination of granular computing and concept learning is critical in the cognitive process. Meanwhile, efficiently and accurately using the information collected from different sources is the focus of data mining in the contemporary. Hence, how to sufficiently learn concepts under a multi-sources context is an essential concern in the field of cognition. This paper offers a new thought for two-way concept-cognitive learning based on granular computing in multi-source fuzzy decision tables. Firstly, based on the best possible guarantee of the classification ability, original information from different sources is fused by conditional entropy, which is the kind of multi-source fusion method (i.e., CE-fusion). Secondly, we learn concepts from a given object set, attribute set, or pair of object and attribute sets in the fused information table, and these three types of concept learning algorithms are designed. This analysis shows that two-way concept learning based on multi-source information fusion is a suitable method of multi-source concept learning. Some examples are valuable for applying these theories to deal with practical issues. Our work will provide a convenient novel tool for researching concept-cognitive learning methods with multi-source fuzzy context.

Automated summary

Concepts learning is crucial in human cognition, and granularity is a fundamental concept. The combination of granular computing and concept learning is critical. Utilizing information from different sources effectively and accurately is a focus in data mining. Learning concepts under a multi-source context is essential in cognition.