Virtual resource scheduling prediction based on a support vector machine in cloud computing

in this study, a virtual resource scheduling prediction algorithm based on a support vector machine (SVM) is proposed to handle the complex, dynamic, changing environment of the cloud platform. First, virtual resource sequences were reconstructed by reconstructing the phase space. Then, the reconstructed virtual resource sequences were used as inputs into an SVM for training and predicting. Finally, a prediction experiment was conducted using actual virtual resource data. The experimental results showed that SVM improved the prediction accuracy and stability of the virtual resource; in addition, the SVM could satisfy the real-time performance and high-accuracy requirements of virtual resource prediction.