Hybrid Machine L earning Model based Continuous wind power forecasting

One significant renewable energy source with long-term electrical potential is wind energy. Notwithstanding the benefits, issues such high initial costs, turbine immobility, and challenges finding wind-producing sites persist. An accurate wind power projection is necessary to minimize the effect on the grid and operational expenses. This paper proposes an Invasive Weed Optimization (IWO) approach inside a mixed machine learning framework to produce accurate wind power estimates. The first analysis is carried out to identify significant features from unprocessed wind power time series datasets using mutual information. The forecast from each sequential sequence is then combined to get the overall forecast. Compared to previous models, this hybrid machine learning technique produces lower Mean Absolute Errors (MAEs) at 1.59% and higher Mean Absolute Percentage Errors (MAPEs) at 8.13%. An additional example demonstrates the model's adaptability and robustness and shows how it may be used to enhance wind power projections for day-ahead energy spot trading markets.