Multiclass Model for Quality of Service Using Machine Learning and Cloud Computing

Abstract

Many web services on the internet have sprung up as a result of the rapid advancement of technology and the deployment of computing. Since users need many web services to achieve their requests, there are many web services that share the same functionality at different qualities. This requires the identification of the quality of web services. Using cloud computing, this paper proposes a model of multi-classification to predict the quality of web services by using machine learning techniques. There are four algorithms of machine learning applied in this work: Multiclass Logistic Regression, Multiclass Decision Forest (DF), Multiclass Decision Jungle (DJ), and Multiclass Neural Network (NN), After comparing the results, it has been found that the Multiclass Neural Network obtained the highest overall accuracy and average accuracy. By using features selection and normalization in this work and compare the algorithms. The selection of the best model is followed by the creation of a web service for the prediction of quality using Azure ML studio.