基于DBN深度学习算法的一站式诉求响应预测方法

One-stop Demand Response Prediction Method Based on DBN Deep Learning Algorithm

为了提高诉求响应的速度,提出了基于机器学习的一站式诉求响应技术。在物理架构中采用事故数据记录器(ADR)服务器和数字化X线摄影术(DR)运行管理,实现一站式诉求响应;利用建模工具来构建例图进行描述诉求响应的运行细节,通过逻辑架构的感知层、网络层和应用层,实现了对一站式诉求响应的逻辑分析;利用机器学习预测方式和深度置信网络(DBN),实现一站式诉求响应的预测。实验表明,在进行对响应的速度进行测试时,所提出的系统响应所需时间最少为1.1 s,在进行对响应预测的准确性测试时,响应预测的准确性最高为97%。

In order to improve the speed of request response,this research proposes a one-stop request response technology based on machine learning.ADR server and DR operation management are used in the physical architecture to achieve one-stop request response.Modeling tools are used to build example diagrams and describe the operation details of the request response.It realizes the logical analysis of the one-stop request response through the perception layer,network layer and application layer of the logical architecture.The machine learning prediction method and the deep belief network(DBN)are used to realize the one-stop request response prediction.Experiments show that the response time of the system proposed is at least 1.1 s in testing the speed of response,and when testing the accuracy of response prediction,the accuracy of response prediction is up to 97%.