A Survey on Selection and Tuning of Machine Learning Models Dynamically for Cloud Network Analysis

Abstract

In every possible area, machine learning has recently been used to harness its incredible strength. Networking and distributed computing systems have long been the main infrastructure for providing machine learning with powerful computational resources. This promising technology can also benefit from networking itself. This paper Focus on the introduction of MLN, which will not only help to address old, intractable network problems, but would also support new network applications. In thispaper, we summarise the basic workflow to disclose how to apply AI innovation to the systems administration space. AI (ML) models tuned to the informational index would effortlessly get deficient.

The model might be unimaginably exact at one point in time, however it might lose its precision sometime in the future because of changes in input information due to its functionality. Dynamic model learning choice is therenfore regularly required. In this post, we suggest a new approach for cloud automated selection and tuning of ML models that automates the design and selection of models and competes with current methods. To further explore data space before, we use unsupervised learning automated development of targeted supervised learning models.