pharma_research_model

Contains the pkl file of the state dict for the model described in our associated publication.

This model was created using our vulcan framework and the hyperparameters listed in the paper.

The data scaler is a sklearn object https://scikit-learn.org/stable/modules/generated/sklearn.preprocessing.StandardScaler.html for which the variable ordering is

var_ordering_scaling = ['insomnia_early_onset', 'hamd15', 'psychomotor_agitation', 
    'insomnia_late_night', 'suicidal_ideation_overall_recent', 'weight_loss_two_weeks', 
    'hamd14', 'guilt_thoughts_beliefs', 'appetite_decreased', 'insomnia_middle_night', 
    'feelings_of_worthlessness', 'hamd11', 'hamd17', 'somatic_gastro_intestinal', 'hamd10', 
    'hopeless_future', 'sociodem2', 'anxiety_mood', 'somatic_leaden_paralysis']

The equating object is a python dictionary that contains the equipercentile transformations needed as described in the associated publication.