Application of Machine Learning Techniques to Classify Web Services

Abstract

The architects and designers of some of the most recent web applications are resorting to the use of a variety of different languages and platforms. Each of these applications are required to communicate among one another to make a common unifying service. Communication may involve the sending and receiving information or the transformation and simplification of it. Each of these heterogeneous applications serves a different intent and are therefore taken to share a different architecture. Different languages and descriptions are so used to leverage the same. It therefore becomes difficult for these applications to communicate with and exchange information amongst one another. Web service technologies are independent of the design and architecture of the underlying applications. Thus, these technologies can tackle this issue by standardizing the way these applications communicate with one another. This paper proposes machine learning models that could be used to classify these web services viz., K-Nearest Neighbors (KNN), Naïve Bayes (Gaussian Naïve Bayes Classifier), Kernel Support Vector Machine Classifier (Kernel SVM), linear SVM, Decision Trees and Random Forests. The QWS (Quality Web Services) dataset has been used for classification and analysis