Prediction method of power market bidding data based on least square method

Abstract

The invention discloses a method for forecasting transaction data of electric power market bidding based on the least square method, which comprises the following steps: acquiring historical data of average transaction price, establishing a prediction model of time and average transaction price, calculating the residual sum of squares, and determining the value of the prediction equation Parameters, calculate the partial derivative of the residual sum of squares to the parameters of the prediction equation, set the partial derivative to zero, find the value of the parameter, set the multiple powers of the independent variables of the prediction equation, and train the data to obtain different residual sums of squares; Calculate the sum of squares of the residuals in turn for comparison, and take the minimum value of the sum of squares of the residuals to obtain an accurate prediction equation; train more real data based on a variety of prediction equations to obtain the optimal solution; set a variety of The prediction equation is used to match the data, and finally the prediction equation with the highest goodness of fit is obtained.

Description

technical field [0001] The invention belongs to the field of electric power market, and relates to a method for predicting bidding transaction data in the electric power market based on the least square method. Background technique [0002] Due to the introduction of the competition mechanism in the electricity market, there are many uncertain factors in the electricity market. Enterprises must accurately and timely understand the current production cost situation, and at the same time fully understand the market supply and demand information, and adopt the most effective bidding strategy based on this information, and the bidding strategy generates It is necessary to derive different bidding strategies based on some basic data such as marginal cost of power plants, bidding price, transaction volume, and market clearing price. Therefore, modern machine learning research is used to realize the prediction of power market bidding transaction data, and let the real data Reaching...

AI Technical Summary

Problems solved by technology

[0006] In recent years, the research on load forecasting in this field based on the least squares method has become popular, but compared with electricity price forecasting, the difficulty is different. The electricity price forecast of bidding transaction has many uncertain factors, it has a non-growth trend and some related human factors , it is difficult to express mathematically. If only one prediction equation is limited to train the data and then fit, it is difficult to predict the data of the change curve

Images