Hyperspectral imaging predicts yield and nitrogen content in grass-legume polyculturesem

  

predict_nutrient.py demonstrates PLSR modelling using Bootstrap validation and it was tested in Python 3.6. 

The folder "organised_data" includes all of the pre-processed data, including reflectance data and laboratory-measured data. 

The program conducts the following parts: 

1. Trains a PLSR model using the original data and then validates the model using the original data. The validation results will be 

saved in a .xlsx file with column names of 'xxx_full'. 

2. Trains a PLSR model using the Bootstrap data (re-sampling with replacement) and then validates the model using the original data. The validation results will be 

saved in the .xlsx file with column names of 'xxx_bs. 

3. Trains a PLSR model using the Bootstrap data and then validates the model using the Bootstrap data. The validation results will be saved 

in the .xlsx file with column names 'xxx_a'.