Nearest Neighbor and Decision Tree Based Cloud Service QoS Classification

Cloud services can be categorized based on their Quality of Service (QoS) parameters. Service recommendations can help users to select the best services among other alternatives. Mostly the recommendations are done either manually or rule-based. However, such recommendations also can be done using Machine Learning (ML) based approaches based on past collected data and user experience. In this work, we have performed classification to categorize cloud services based on the QoS parameters using a public data set. A comparison between widely accepted classifiers such as 1) Decision Tree (DT) and K Nearest Neighborhood (KNN) is carried out with standard performance measures such as Overall Accuracy (OA), Average Accuracy (AA), and Kappa coefficient. The preliminary result suggests that the KNN classifier is strongly suggested as compared to the DT approach for the cloud service QoS classification problem.