Research On Data Compression Classification Mechanism For Edge Computing

With the development of 5G era,cloud centers generate massive data.Deployment of micro data processing equipment on the edge of the network has become one of the methods to solve the problem of data cache and improve the efficiency of data transmission.The popularity of cloud services promotes the development of edge computing.A large amount of data is transmitted from the cloud center to the edge nodes,which leads to disadvantages such as node overload and transmission delay increase.In order to reduce the massive data generated in the transmission process of edge network,the compression technology can be used to compress the original data.Classifying data when the edge of the network node,testing operation,transmission equipment in a state of compression of data information,the need to compress the data decoding process first,and so on the edge of the compressed data produced in the process of decoding network data transmission delay,the edge nodes cannot real-time data processing,thus causing edge of computing devices,data processing work lag.Therefore,this paper proposes a data compression classification mechanism oriented to edge computing.The main work content and innovation points of this paper are summarized as follows.1.In order to reduce the delay of data processing,a self-encoding data compression method is presented in this paper.The self-encoding network is used to achieve efficient homomorphic data compression,which provides an effective guarantee for the edge network nodes to implement accurate data compression.2.In view of the problem that the introduction of autocoded data compression method leads to the decline of data classification accuracy,this paper proposes a feature reconstruction data compression classification mechanism oriented to edge computing.Firstly,data is collected in the transmission segment of edge computing node,and the original input data is reconstructed with features.Then the self-encoding data compression method is used to construct the feature compression on the reconstructed input data,reduce the data scale,and finally build the data classification model oriented to the compression results.3.In order to accelerate the data transmission speed,effectively process the edge node data,and further build a more efficient data compression classification mechanism,this paper presents a data compression classification method called Manifold data compression classification,which realizes homomorphic feature mapping by constructing data feature compression method,and builds a classification model oriented to compressed data.Accurate and efficient data compression and classification in the edge network.Experimental results show that the proposed data compression classification mechanism not only reduces the data scale and improves the data processing efficiency,but also effectively ensures the classification accuracy of the compressed data.By reducing the data scale,the purpose of processing the mass data of the edge network quickly and efficiently is achieved.