Monthly load prediction method based on XGBoost algorithm

The invention provides a monthly load prediction method based on an XGBoost algorithm, and the method comprises the steps: carrying out the conversion of a load index, and carrying out the thermal coding of index data influencing a load factor; taking the monthly maximum load data of the user as the output of the model, and selecting an influence factor variable with strong correlation with the monthly maximum load as an input variable; converting the influence load prediction factor variable of the selected model into a sparse matrix to form XGboost modeling data; defining the monthly maximumload as an XGboost model and outputting the XGboost model; defining a model learning objective function, regression tree generation parameters and the like to construct an XGboost model for load prediction; and performing cross validation test on each parameter of the XGboost to obtain a parameter combination with the highest model precision, and performing load prediction based on the obtained parameter combination. Ver