A Distributed Renewable Energy Output Prediction Method Based on Multi-source Time-varying Data Optimal Multi-kernel Function

Abstract

The invention discloses a distributed new energy output prediction method based on a multi-source time-varying data optimal multinuclear function. Through using a multi-source time-varying data distribution characteristic method, a single-source data support vector machine prediction or classification method, and a nuclear matrix rank space difference method, an optimal multinuclear function which reflects and influences a multi-source time-varying data distribution characteristic of distributed new energy output is constructed. During distributed new energy output prediction, the optimal multinuclear function is used to fuse the multi-source time-varying data so as to increase distributed new energy output prediction precision.

Description

technical field [0001] The invention relates to a distributed new energy output prediction method based on multi-source time-varying data optimal multi-core function, which belongs to the field of data mining. Background technique [0002] In recent years, with the rapid development and popularization of computer and information technology, the scale of industrial application systems has expanded rapidly, and the data generated by industrial applications has grown explosively. This explosive growth of data volume poses challenges to multi-source, heterogeneous, high-dimensional, distributed, and non-deterministic data. Big data thinking is a product of this environment. One of the salient features of big data is multi-source heterogeneity. characteristic. [0003] There are many data that affect distributed new energy forecasting, such as distributed new energy access and operation information, power grid production management information, geographic meteorological informat...

AI Technical Summary

Problems solved by technology

[0005] 1. Directly select a global kernel function and a local kernel function, and form a mixed kernel function after simple weighting. The disadvantages of this method are: ① limit the composition of multi-kernel functions to only two single-kernel functions; The characteristics of multi-source heterogeneous data; ③ commonly used global kernel functions include polynomial kernel function Polynomial and Sigmoid kernel function, this method usually chooses one of the global kernel functions, which is blind

This method ignores different types of features from different data sources, and does not make full use of multi-kernel functions to effectively process multi-source heterogeneous data.

Images