Precision index construction method for evaluating an electric power prediction model

Abstract

The invention relates to the technical field of precision index correlation, in particular to a precision index construction method for evaluating an electric power prediction model, which comprises the following steps of: establishing an identification framework of a relative error index; Constructing basic reliability distribution on the identification framework; Obtaining a credibility functionon the identification framework; And fusing multiple credibility functions to generate a new precision index. Adopting The relative error RE as the most common and typical precision index, and in order to comply with the principles of comprehensiveness, typicality and the like, a precision index system comprising the RE for evaluating the electric power prediction model needs to be established. Mining More valuable information of the RE is mined. meanwhile, the problem that RE is inconsistent with indexes such as MAE, MAPE, RMSE, SDE, FEM and r in time span and cannot be adopted at the same time is solved, all-around precision index selection is carried out, and therefore the precision of the electric power prediction model is evaluated more accurately and comprehensively, model optimization is achieved, and then the prediction risk is reduced.

Description

technical field [0001] The invention relates to the technical field of electric power prediction model evaluation, in particular to a method for constructing an accuracy index for electric power prediction model evaluation. Background technique [0002] Electric power forecasting, including but not limited to wind power forecasting, photovoltaic power forecasting, and load forecasting. Accurate electric power forecasting is the basis and premise of electric power planning and design, and it is also an important guarantee for the safe and economical operation of the power grid. At present, there are many electric power prediction models, such as gray prediction model, neural network prediction model, support vector machine prediction model, etc. These models use data information from different angles to predict electric power. How to achieve model optimization needs to evaluate the model first. When choosing to use a single index to evaluate the electric power prediction mo...

AI Technical Summary

Problems solved by technology

[0004] Based on this, in order to establish an accuracy evaluation index system including the most commonly used and typical RE indexes, more comprehensively and comprehensively evaluate electric power prediction models, realize model optimization, and reduce prediction risks, it is necessary to target the time between RE and other accuracy indexes In order to solve the problem that the spans are inconsistent and cannot be used at the same time, a method for constructing accuracy indicators for the evaluation of electric power prediction models is proposed, which is used to construct the evaluation index system of electric power prediction models, to comprehensively evaluate the accuracy of electric power prediction models, and to guide predictions Reasonable application of the results to realize the optimization of the electric power prediction model

Images