【摘要】 在数据流分类大环境中,数据量级不断增大,数据样本对应的概念也在不断发生变化,这不但产生"概念漂移",数据类别分布不平衡的现象也出现愈发频繁。面对这些问题,为了快速察觉到数据分布的变化,及时调整分类模型以适应新的数据分布,针对在类别不平衡环境中的不同类型的概念漂移,设计利用部分标记数据给出不同判定方式以及分类模型再构建。实验结果显示新模型有较好的性能。更多还原

【Abstract】 In the data stream classification environment, the magnitude of data keeps increasing, and the concepts corresponding to data samples keep changing, which not only results in concept drift, but also results in more and more frequent unbalanced distribution of data categories. In the face of these problems, in order to quickly detect the change of data distribution and adjust the classification model in time to adapt to the new data distribution, in view of the concept drift of different types in the category unbalanced environment, different judgment methods are designed to use partial labeled data and the classification model is rebuilt. Experimental results show that the new model has better performance.更多还原