基于连续细节特征分解的数据并行聚类挖掘

Data Parallel Clustering Mining Algorithm Based on Continuous Detail Feature Decomposition

为了提高对物联网多媒体音视频数据的检测识别能力,提出基于分段聚类的物联网多媒体音视频数据并行聚类挖掘方法。对采集的物联网多媒体音视频数据进行分段检测和连续谱密度特征分解,采用子空间匹配降噪方法进行多媒体音视频数据的滤波提纯处理,建立物联网多媒体音视频数据的多载波分析模型,结合频谱融合和连续细节特征分解方法进行物联网多媒体音视频数据特征提取,对提取的多媒体音视频数据谱密度特征量进行连续细节特征融合分解分析,根据特征聚类结果,进行物联网多媒体音视频数据并行聚类挖掘。仿真结果表明,采用该方法进行物联网多媒体音视频数据并行聚类挖掘的准确性较高,对多媒体音视频数据检测的抗干扰能力较强,提高了对物联网中多媒体音视频数据的准确识别和特征辨识能力。

In order to improve the detection and recognition ability of Internet of Things multimedia audio and video data,a parallel clustering mining method of Internet of Things multimedia audio and video data based on subsection clustering is proposed.Segmented detection and continuous spectral density feature decomposition are carried out on the collected multimedia audio and video data of Internet of Things.Subspace matching noise reduction method is adopted to filter and purify the multimedia audio and video data.Multi-carrier analysis model of multimedia audio and video data of Internet of Things is established.Combining spectrum fusion and continuous detail feature decomposition method,features of extracted multimedia audio and video data are extracted,and continuous detail feature fusion decomposition analysis is carried out.According to feature clustering results,parallel clustering mining of multimedia audio and video data of Internet of Things is carried out.Simulation results show that the proposed method has high accuracy in parallel clustering mining of multimedia audio and video data in Internet of Things,and has strong anti-interference ability in multimedia audio and video data detection,which improves the ability of accurate recognition and feature recognition of multimedia audio and video data in Internet of Things.