Fig. 9
From: Efficient resource provisioning for elastic Cloud services based on machine learning techniques
Percentage of unserved requests for forecasting models #3, #5, and #8, based on the resource allocation of Table 4 This figure shows the percentage of unserved requests on each hourly period for forecasting models #3, #5, and #8 (in order to avoid a mesh of points in the graph, we have chosen these three cases in representation of the basic models, the polynomial SVM-based models, and the RBF SVM-based models, respectively)