WIMMC:加权的增量式极大边界准则算法

WIMMC:weighted incremental maximum margin criterion

一些经典降维算法并不是最优的降维策略,它们不再适用于流形式且大尺度的Web文本数据,因此提出了一种加权的增量式有监督的降维算法,称为加权的增量式极大边界准则(Weighted Incremental Maximum Margin Criterion,WIMMC)。WIMMC通过加权得到比传统算法更好的结果,而且可以增量地有监督地处理大尺度的Web文本数据。给出了算法的收敛性证明和一些实验,并从实验结果可以看出,通过WIMMC降维之后的分类效果比其他降维算法更有效。

Some traditional algorithms can no longer deal with the streaming text data,so this paper proposes a weighted incre mental supervised algorithm,called Weighted Incremental Maximum Margin Criterion（WIMMC） to reduce the dimensionality.WIMMC can supervised incrementally handle large-scale text data and improve the performance of following classification procedure.The paper proves the convergence of the algorithm and gives experiments to show that WIMMC can more effectively improve following classification than other dimensionality reduction algorithms.