Prediction of Workloads in Cloud using ARIMA-ANN

This study introduces an innovative hybrid ARIMA-ANN model personalized for cloud workload prediction. Unlike existing models that focus solely on linear or nonlinear patterns, the approach combines the strengths of ARIMA for time-series linear trends and ANN for nonlinear data complexities. This integration ensures higher accuracy, as validated using the MIT Supercloud dataset. The methodology leverages data pre-processing, sensitivity analysis, and advanced validation techniques, demonstrating improved accuracy in scenarios of high workload variability. This model supports cloud providers in resource optimization and dynamic load management.