Modeling over MEDS data

Modeling over MEDS data can be as simple as deploying the executable file for an existing modeling package, or as complex as writing your own neural network, pre-training/fine-tuning paradigm for MEDS datasets. To see a few of these options in action, check out the following tutorials:

1. Make predictions using MEDS-Tab
2. Make predictions using MEDS-Torch
3. Make predictions using CEHR-BERT

You can also leverage individual model pipeline components to help simplify the building of your models, with tools like MEDS-Torch; check those functionalities out in action via their documentation.