Abstract: Concept drift is one of the most common phenomena in data stream mining, where the underlying knowledge patterns in the data stream change dynamically over time, leading to a decline in the accuracy of previously established classifiers. To address this issue, we propose a Concept Drift Data Stream Classification Algorithm Based on Incremental Weighting (SCIW). This algorithm employs a heuristic weight updating strategy combined with an adaptive method based on accuracy differences, and it also improves the Poisson distribution-based resampling strategy. The SCIW algorithm is capable of adapting to various types of concept drift, effectively mitigating the decline in classifier accuracy. Experimental results on 14 synthetic datasets and 6 real-world datasets demonstrate that SCIW and ARF outperform other algorithms in terms of accuracy. Additionally, SCIW significantly outperforms ARF in terms of time and memory consumption, with overall average time consumption at approximately 83% of ARF's and overall average memory consumption at approximately 13% of ARF's.

Key words: data stream, concept drift, classification, ensemble learning, adaptive