基于概念漂移检测算法的数据流分类模型

Concept drift detection method based data stream classification model

为了克服数据流概念漂移现象对分类模型的影响,提高数据流分类准确率,提出了一种基于概念漂移检测算法的数据流分类模型。针对不同概念漂移类型使用不同的方法进行检测,该模型通过对概念漂移进行监控,从而有效控制分类模型的更新频率,做到有的放矢地更新分类器模型,提高分类模型的分类性能。通过使用两种不同的数据集进行实验,并与传统分类模型进行比较,验证了该模型的有效性和正确性。

To overcome the effect of the data stream concept drift phenomenon for data stream classification, improve the accura- cy of classification, a concept drift detection method based data stream classification model is presented. Firstly, the related in- formation of concept drift is introduced. For different concept drift types, there are different kinds of concept drift detection methods for recognizing. Using the proposed concept drift detection method, the updating process of the data stream classifica- tion model can be controlled in reason, and the accuracy of classification can also be improved. The experiment employs two kinds of data stream for testing the validation and correctness of the proposed model. In addition, some traditional classification models are also used for comparing test.