Wind power prediction error segmented fitting method based on Gaussian mixture model

The invention discloses a wind power prediction error segmented fitting method based on a Gaussian mixture model, and the method specifically comprises the steps: carrying out the normalization processing of a wind power prediction value and a wind power prediction error of the historical data of a wind power plant by taking the installed capacity of the wind power plant as a reference value; making a distribution scatter diagram of the wind power prediction value and the prediction error; dividing the wind power prediction value into a plurality of intervals from small to large according to the difference of all parts of the distributed scatter diagram, and respectively making frequency histograms of wind power prediction errors in all the intervals; applying a Gaussian mixture model to fit the frequency histograms of the wind power prediction errors in the intervals respectively, and adopting an expectation maximization algorithm to perform parameter estimation on the Gaussian mixture model in the intervals