Module connectome.models

helper functions, classes and wrappers to create and compute the different models

Sub-modules

connectome.models.E2E_conv

the custom Edge-to-Edge layer developed by Kawahara et al.

connectome.models.brainnet_cnn

prepares the data for the CNN and trains the CNN

connectome.models.ebm

wrapper for the explainable boosting machine with integrated feature selection based on the mutual information

connectome.models.evaluation

function to evaluate the fitted model on test data

connectome.models.framework

framework to open up saved models as well as train and save new models

connectome.models.lgb

wrapper for gradient boosting classification and regression models

connectome.models.pipeline_RF

function to create and compute a random forest classification or regression model

connectome.models.pipeline_elastic_net

functions to prepare data for elastic net (e.g. calculate squared or absolute values or interactions) and run elastic net model