Forscasting the power distribution by the machine learning method - randomforest
It's to pload the model-needed data(table) to the R global env, and to update(add new) data to the table.
This is designed for re-training the rf model or re-using the exist model(380 channel have 380 models,every channel has a own model!)
This part is the hardcore part,the forecasting result has been store as a csv file,so we can also download the result matrix.