Benchmark multi-model/multi-view models.

Note

This page is a reference documentation. It only explains the function signature, and not how to use it. Please refer to the gallery for the big picture.

mmbench.residualize.residualize(train_df, train_data, test_df, test_data, formula_res=None, formula_full=None, site_name=None, discrete_vars=None, continuous_vars=None)[source]¶

Custom linear and/or site Combat residualization.

Parameters

formula_res : str

what we want to residualize for ‘age + sex + site’ or ‘age + sex’ using linear regression.

formula_full : str

what we want to adjusted for e.g ‘age + sex + diagnosis’ or ‘age + sex + site + diagnosis’ using linear regression.

train_df : DataFrame

table defining the terms used in formula_full.

train_data : array (n_samples, n_features)

the training data.

test_df : DataFrame

table defining the terms used in formula_full.

test_data : array (n_samples, n_features)

the test data to be transformed.

site_name : str, default None

the name of the column containing the site information.

discrete_vars : list of str, default None

the name of the covariates which are categorical.

continuous_vars : list of str, default None

the name of the covariates which are continuous.

Returns

train_data : array (n_samples, n_features)

the residualize training data.

test_data : array (n_samples, n_features)

the residualize test data.