Stream Concept-cognitive Computing System for Streaming Data Learning

• DOI: 10.36227/techrxiv.24189984
• Publication Date: 2023-10-02T16:10:07Z
• Authors: Yunlong Mi
• Keywords: Concept Drift; Streaming Data; Streaming algorithm

<p>People can often acquire knowledge dynamically and rapidly from different types of data, yet existing incremental learning algorithms are still computationally time consuming most stream methods mainly designed for streaming data while ignoring other data. Hence, this paper proposes a novel dynamic concept (CL) algorithm by imitating human cognitive processes the perspective brain logical cognition, which is named concept-cognitive computing system (streamC3S). For streamC3S, it consists three aspects: space, CL process, model update process. Moreover, considering drift frequently occurs in over time, an extended version streamC3S (namely, streamC3S<sub>E</sub>) also proposed work. Specifically, we first show related theories streamC3S<sub>E</sub> on basis space. Then overall framework its corresponding shown. Finally, experimental results various datasets, including standard machine image two traffic streams, validate effectiveness our compared to state-of-the-art algorithms. </p>