Neural Network Predicted Data of Solar Images

Description

This data was generated by a neural network. It includes mean irradiance, weighed irradiance, cloud coverage, atmospheric indicator, irradiance obtained by equation, sun's position (real and obtained via picture), and their predicted values. They were obtained by using a convolutional neural network and a recurrent neural network hybrid. LOSS: 0.0131 // MAE: 0.0607 This dataset is meant to be used as a foundation of what neural networks can do and the accuracy of the data they can predict. Also, this dataset may be used to describe the movement of the sun and the climate conditions of the location in the photographs.

Steps to reproduce

The neural network was coded in Python 3.4. The data was extracted using various formulas of Solar Geometry and Climate.

Institutions

• Instituto Tecnológico de Ciudad Madero

  Tamaulipas, Ciudad Madero

Categories

Licence