摘要： 为提高分布式多媒体服务组合系统中路径的健壮性和可靠性，提出一种基于Markov链和加权朴素贝叶斯分类器(WNBC)的异常预测算法。该算法利用Markov模型预测系统节点的资源状态信息，使用WNBC对预测的节点状态进行分类，以判断节点是否可能发生异常。实验结果表明，该算法能根据节点的状态信息预测系统节点的不同异常状态，性能较同类算法有较大的改善。

关键词: 分布式多媒体, 服务组合, 异常预测, Markov链, 加权朴素贝叶斯分类器

Abstract: In order to improve robustness and reliability of the composite service path in distributed multimedia service composition system, this paper presents an anomaly prediction algorithm based on Markov chain and Weighted Na?ve Bayesian Classification(WNBC). It adopts Markov chain model to predict the resource states of the node. Based on the prediction, a WNBC is introduced to determine whether an exception occurs in the node. Simulation results demonstrate that the proposed algorithm can predict the anomalies effectively according to the states of node, and its performance is improved compared with previous algorithms.

Key words: distributed multimedia, service composition, anomaly prediction, Markov chain, Weighted Na?ve Bayesian Classifier(WNBC)

中图分类号: 

• TP301.6