Note: tested with tensorflow 2.0 (pip install tensorflow==2.0)
This LSTM network serves as a basis for a MACHINE-LEARNING-BASED FORECASTING OF PV-BASED MICROSYSTEM POWER OUTPUT paper i made during april 2019.
The data was downloaded from PVDAQ database. Data was collected from Andre Agassi Academy for the last 7 or so years. https://maps.nrel.gov/pvdaq/?aL=nrSQLM%255Bv%255D%3Dt%26zNVfnk%255Bv%255D%3Dt%26zNVfnk%255Bd%255D%3D1&bL=clight&cE=0&lR=0&mC=35.93354064249312%2C-112.8076171875&zL=7