Irradiance data interpolation method and system based on decision tree classification

Abstract

The invention provides an irradiance missing data interpolation method and system based on decision tree classification. The method comprises the following steps: detecting a missing irradiance value;classifying the data according to weather by using a decision tree classification method; calculating a solar altitude angle according to the date, time, longitude and latitude; under different weather categories, fitting an irradiance sequence by using different neural network parameters according to the current time, power and environment temperature; and interpolating the missing irradiance value. According to the method, the information of the weather phenomenon is reserved through the classification of the decision-making tree, the input of the solar altitude angle is added into the neural network, and the information of date and time is reserved so that the data interpolation effect is obviously superior to that of other algorithms, and the interpolation accuracy is greatly improved. According to the method, missing irradiance data can be filled, the problem that irradiance, temperature and power timestamps are inconsistent is solved on the premise that the performance is not reduced, and the number of samples can be increased.

Description

technical field [0001] The invention belongs to the technical field of data processing, and relates to sequence prediction technology, in particular to a method and system for interpolating missing irradiance data based on decision tree classification. Background technique [0002] Generally, during data preprocessing, data with complete sequences will be selected for training. If there are some missing data, the entire data record will be deleted, resulting in the loss of a lot of data information. Therefore, the interpolation of missing data has attracted much attention. It can not only improve the integrity of the sequence and facilitate subsequent data processing, but also effectively solve the problem of inconsistent timestamps of different types of data through interpolation. However, most of the existing missing data interpolation methods only start from the perspective of a single sequence, mining the regular information of the sequence to interpolate, resulting in l...

AI Technical Summary

Problems solved by technology

However, most of the existing missing data interpolation methods only start from the perspective of a single sequence, mining the regular information of the sequence to interpolate, resulting in low accuracy of data interpolation, but will bring more interference to the original data, which is difficult to implement. application

Images