基于深度置信网络的监测数据分类方法研究

Research on Classification Method of Monitoring Data Based on Depth Confidence Network

深度置信网络数据分布具有随机性,分类难度较高,为了提高网络数据的监测性能,提出基于模糊K均值聚类的深度置信网络监测数据分类算法。构建深度置信网络监测数据的相空间分布结构模型,结合模糊聚类方法,建立深度置信网络监测数据的模糊关联规则分布集,提取深度置信网络监测数据的统计特征量,采用模糊K均值聚类分析法方法进行深度置信网络数据的特征融合聚类,根据聚类结果在信息融合中心进行深度置信网络数据的自适应分类和特征辨识。仿真结果表明,采用该方法监测数据分类的准确性较高,数据监测性能较好,提高了深度置信网络数据监测和信息分类能力。

The data distribution of deep confidence network is random and the classification is difficult.In order to improve the monitoring performance of network data,a deep confidence network monitoring data classification algorithm based on fuzzy K-means clustering is proposed.The phase space distribution structure model of depth confidence network monitoring data is constructed.Combined with fuzzy clustering method,the fuzzy association rule distribution set of depth confidence network monitoring data is established,the statistical characteristic quantity of depth confidence network monitoring data is extracted,and the fuzzy K-means clustering method is used to carry out feature fusion clustering of deep confidence network data.According to the clustering results,the adaptive classification and feature identification of deep confidence network data are carried out in the information fusion center.The simulation results show that the accuracy of monitoring data classification is high,the performance of data monitoring is good,and the ability of deep confidence network data monitoring and information classification is improved.