Optimisation of classification algorithm of associated data features of large-scale network system. Issue 2 (28th February 2020)

• Record Type: Journal Article
• Authors: Cao, Yu
• Is Part Of: International journal of internet protocol technology. Volume 13:Issue 2(2020)
• Page Start: 55
• Page End: 60
• Publication Date: 2020-02-28
• Subjects: large-scale network system -- deep learning -- data classification -- feature extraction; File Transfer Protocol (Computer network protocol) -- Periodicals; Multicasting (Computer networks) -- Periodicals; 004.678
• Languages: English
• ISSNs: 1743-8209

Abstract

Data feature extraction is vulnerable to external interferences during traditional classification of associated data of large-scale network system, resulting in low classification accuracy and large time consumption, so an optimised classification algorithm of large-scale network system associated data features based on deep learning is proposed. In this algorithm, associated data of large-scale network system is acquired through the multi-sensor quantisation fusion method, and the acquired data is done with spectral decomposition to obtain the convergence conditions for feature components of associated data of large-scale network system; then the feature components are processed based on a spectrum analysis model to extract features of associated data of large-scale network system; the features of the associated data are done with piecewise regression analysis, and data samples are output in classification based on the deep learning algorithm. Simulation results show that the proposed method can accurately classify associated data with relatively short learning steps with accuracy 24.7% higher than that of the ARMA classification method and 22.6% higher than that of the decision tree classification method. It is verified that the method proposed in this paper is obviously better than the traditional methods with higher classification accuracy and less time consumption.