A Classification and Novel Class Detection Algorithm for Concept Drift Data Stream Based on the Cohesiveness and Separation Index of Mahalanobis Distance

<div>Comparison of Kappa Statistic on datasets. (a) The chunk size in the Covertype dataset is set to 58102, and (b) the chunk size in the ArtificialCDS dataset is set to 10000.</div>

Journal of Electrical and Computer Engineering

fig1

Figure 1

Figure 1: A Classification and Novel Class Detection Algorithm for Concept Drift Data Stream Based on the Cohesiveness and Separation Index of Mahalanobis Distance 

Figure 1 | A Classification and Novel Class Detection Algorithm for Concept Drift Data Stream Based on the Cohesiveness and Separation Index of Mahalanobis Distance 