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.workflow.predict.benchmark_pred_exp(dataset, datasetdir, datadir, outdir)[source]¶

Compare the learned latent space of different models using prediction analysis.

Parameters

dataset : str

the dataset name: euaims or hbn.

datasetdir : str

the path to the dataset associated data.

datadir : str

the path containing the embedding data.

outdir : str

the destination folder.

Notes

The samples are generated with the ‘bench-latent’ sub-command and are stored in the ‘outdir’ in two files named ‘latent_vecs_<dataset>.npz’ and ‘latent_vecs_train_<dataset>.npz’ for the test and train sets, respectively..
The associated metadata are stored in two others files named ‘latent_meta_<dataset>.tsv’ and ‘latent_meta_train_<dataset>.tsv’.
The samples shape is (n_samples, n_subjects, latent_dim). All samples must have the same number of samples and subjects, but possibly different latent dimensions.
The metadata columns must be the same.