moved random forest code to models/random_forest

genophenoenvo / neon-datasets

Repository for code and derived data from NEON data products

MIT License

0 stars 4 forks source link

moved random forest code to models/random_forest #26

Closed dlebauer closed 3 years ago

dlebauer commented 3 years ago

@DebasmitaPal1206

we have two teams, PEG_RFR0 and PEG_RFR for the forecasts based on raw and interpolated observations, respectively
can the ipynb be separated out into a notebook for data and model analysis and a python script that runs the model?
you can forecast out as far as you want but the minimum is 35 days. Since the forecast should start on day n+1, this should be adjusted to make sure max(time) >= n +36

DebasmitaPal1206 commented 3 years ago

@dlebauer I am working on separating .ipynb files. I'll also adjust the number of predicted days.

DebasmitaPal1206 commented 3 years ago

@dlebauer I have prepared the automated scripts for daily predictions. The code is attached herewith. Each of the folders contains trained models for each of the sites and a python script which will generate csv file with date. The modification to 36 days is incorporated in these scripts.

Generated csv file name for PEG_RFR0 : prediction_mm-dd-yy.csv Generated csv file name for PEG_RFR :prediction_interpolation_mm-dd-yy.csv

PEG_RFR.zip PEG_RFR0.zip