How to run

To use ProMS for selecting features and training a full model, run proms_train. The required arguments are -f run_config_file and -d data_config_file. proms_train will perform training and validation using the training data. After training/validation, a final full model will be trained with all train data with hyper-parameters that achieves the best average cross validation performance. If test data is provided in the data configuration file, proms_train will also apply the full model to the test data.

To see all available arguments for the command:

$ proms_train -h
usage: proms_train [-h] -f FILE -d FILE [-s SEED] [-o OUTPUT_ROOT]
                [-r RUN_VERSION] [-p]

optional arguments:
-h, --help            show this help message and exit
-f FILE, --file FILE  configuration file for the run
-d FILE, --data FILE  configuration file for data set
-s SEED, --seed SEED  random seed
-o OUTPUT_ROOT, --output OUTPUT_ROOT
                        output directory
-r RUN_VERSION, --run_version RUN_VERSION
                        name of the run, default to current date/time
-p, --include_pca     include supervised PCA method in the results