Linear Statistical Features for the Purposes of Computer Network Automation
Linear Statistical Features for the Purposes of Computer Network Automation

Author(s): Milan Milivojević, Milan Pavlović, Marija Zajeganović
Subject(s): Social Sciences
Published by: Udruženje ekonomista i menadžera Balkana
Keywords: Computer networks; Automation; Python; Statistical analysis
Summary/Abstract: The development of artificial intelligence finds applications in various practical areas, such as computer networks. The primary goal of introducing automated methods is the efficient optimization of computer network operations. As input data for each algorithm, parameters that best describe the computer network are defined. Therefore, it is essential to define which parameters are relevant to ensure the significant use of automation methods. Parameters like Round-Trip Time delay resulting from ping commands can be used as the basis for defining input parameters for the automation system. This approach can help detect anomalies in network operation due to topology disruptions and increased load on specific links within the network. These anomalies can be mitigated by adjusting routing protocol parameters and activating redundant links. The authors describe basic features that can be extracted from time series data containing information about delay times. Special emphasis is placed on the characteristics that result from linear statistical analysis using the Python programming language.

• Details
• Contents