Evaluating Machine Learning Algorithms for Load Forecasting in Smart Grid

Because it makes computing much simpler and eliminates the need to purchase the actual hardware required for computations, cloud computing is quickly replacing on-premise computing in the information technology sector. These businesses depend on the availability of a reliable and affordable electrical power supply because they house numerous computers and servers whose primary power source is electricity. Cloud centers use a lot of energy. With recent increases in electricity prices, one of the biggest obstacles in designing and efficiently placing data and scheduling nodes to unload or transfer storage is one of the upkeep of such centers. Another difficulty is to reduce the amount of electricity that data centers use and conserve energy. In this project, we suggest using an Extreme Gradient Boosting (XGBoost) model to offload or transfer storage, forecast electricity prices, and as a result cut data center energy expenses. On a real-world dataset provided by the Independent Electricity System Operator (IESO) in Ontario, Canada, the effectiveness of this strategy is assessed in order to offload data storage in data centers and effectively reduce energy consumption. 70% of the data is used for training and 30% for testing.