基于二分K-均值的SVM决策树自适应分类方法

分析和研究了自适应降维算法在高维数据挖掘中的应用。针对已有数据挖掘算法因维灾难导致的在处理高维数据时准确率和聚类质量都较低的情况,将二分K-均值聚类和SVM决策树算法结合在一起,提出了一种适用于高维数据聚类的自适应方法 BKM-SVMDT。该算法能保证二分K-均值聚类是在低维数据空间中进行,其结果再反过来帮助SVM在高维空间中的执行,这样反复执行以取得较好的分类精度和效率。标准数据集的实验结果证明了该方法的有效性。

Adaptive SVM decision tree classification algorithm based on bisecting K-means

This paper analyzed and researched the applications of adaptive dimension reduction algorithm in high-dimensional data mining. To improve the situation of low accuracy and low clustering quality caused by existing data mining algorithms dealing with high dimensional data, it proposed an adaptively classification algorithm, combining bisecting K-means clustering and support vector machine decision tree, for high dimensional data classification. The BKM-SVMDT algorithm transformed the high dimensional dataset into low dimensional one to ensure data mining in the low-dimensional space, and its results could in turn help SVMDT in high-dimensional space. Adaptively executed the algorithm in order to obtain better classification accuracy and efficiency. Extensive experimental results on standard datasets show the effectiveness of the algorithm.