The core function of the PBCT algorithm is included in the file utils/PBCT.py. Given the labeled and unlabeled training data as well as the test data, it triggers the training of the complete-view model and parital-view models, save the model parameters in the desired paths, and return the test error measured using RMSE. An example for utilizing the PBCT algorithm is provided in the main section of this file.

This repository is a reproduction work of the PBCT algorithm on the field data, which can be found in dataset. The main changes are made in the folder: /utils.