自适应数据库中基于特征向量的聚类算法的研究与改进

The Research and Improvement of the Clustering Algorithm based on Feature Vectors in Autonomic Database

在负载自适应数据库系统中,负载特征化部件是关键部分,首先要对负载分类,然后根据分类的情况预测负载性能。负载的分类一般采用聚类算法,聚类算法中比较典型的就是K-means算法。但在K-means算法中,k值必须提前设定而且不能根据负载的实际情况改变,就是对算法的一个改进,使得k值动态的、能够根据负载的实际情况改变。实验结果表明,使用该算法的分类结果预测负载运行时间的准确性有明显提高。

In autonomic database system, workload characterization is a key part. In workload characterization, workload should be class, then anticipate workload performance. Workload classification use clustering algorithm, and in clustering algorithm, the typical is the K-means algorithm. But in the K-means clustering algorithm, k should be defined and not changed. This paper makes a improvement in the algorithm, so the k is changed if needed. The result of the experiment shows that the veracity used cluster based on feature vectors makes of forecast workload runtime is improved.