In the real-world scenario of data stream classification, label scarcity is very common. More challenges are data streams always include concept drifts. To handle these challenges, an algorithm of semi-supervised classification of data streams based on adaptive density peak clustering (SSCADP) is proposed. In SSCADP, to generate concept clusters at leaves in a Hoeffding tree, a density peak clustering method and a change detection technique are combined to adaptively locate the clustering centers. Concerning concept drift detection, we argue th...
