Dynamic machine learning approach for workload prediction in cloud environments.

If the data exhibit a clear trend, the framework performs differencing until the series becomes stationary, starting with first-order differencing (d = 1) and proceeding to higher orders if necessary.

also in 19,460 other papers