Research on Curve Decomposition Method of Unit Contracts for Differences Based on Data Mining

Aiming at the problem that the decomposition results of curve CFDs based on optimization and empirical methods are not very convincing, this paper proposes a data mining-based method for unit CFD curve decomposition. Establish a historical case database based on historical data. The neural network is used to build a model case library to make up for the gaps in historical scenes. Based on the establishment of the case library, the monthly contract electricity amount is broken down into the settlement period by case-based reasoning. This method simplifies the decomposition process, improves the calculation speed, and has certain popularization significance.