The main notebook for evaluating downstream metrices is downstream_matrices_evaluations/downstream_matrices_evaluation/calculate_matrices.ipynb

Please download all the files in the downstream_matrices_evaluations/downstream_matrices_evaluation, including the three empty files first_trial_results, CSVs and pngs. We need these three files to store the model results, CSV files and plots separately.

Under downstream_matrices_evaluations, there are two other files, TRB_label_assist and Error_bars.
TRB_label_assist contains the model for mode prediction.
Error_bars contains the methodology for calculating energy consumption. 


Changes
1. TRB_label_assist
(1) performance_eval.py
(2) models.py
(3) classification_performance.ipynb (You can ignore all the changes in this jupyter notebook. I initially worked directly on this jupyter notebook, but I have moved all my work to calculate_matrices.ipynb and added a lot of new work there.)

2. Error_bars
(1) get_EC.py
(2) confusion_matrix_handling.py


Initial code came from:
1. TRB_label_assist:
https://github.com/e-mission/e-mission-eval-private-data

2. Error_bars: 
https://github.com/allenmichael099/e-mission-eval-private-data/tree/error_bars/Error_bars