Abstract:

This paper established time, cost, service capability, credibility integrated optimal monitoring model and the improved particle swarm algorithm was used to optimize parameters of support vector machine,the monitoring model was predicted by time series prediction.When the actual and predicted values of the model error ranges satisfied monitoring requirements, the resource service was normal. Finally, through an example, using the root mean square error(RMSE) as prediction accuracy of evaluation model,the results and comparative analysis show that the method is effective and feasible.

Key words:  , cross-organizational resources chain, particle swarm algorithm, support vector machine, time series prediction

摘要：

建立了以时间、成本、服务能力、信誉度综合最优的监控模型，并利用改进的粒子群算法优化支持向量机参数对监控模型进行时间序列预测，当监控模型的实际值与预测值在规定的误差范围内时，该资源服务是正常运行的。最后通过一个算例进行监控预测研究，以均方根误差(RMSE)作为评价监控模型的预测精度，研究结果及分析对比表明，该方法有效、可行。

关键词: 跨组织资源链, 粒子群算法, 支持向量机, 时间序列预测