Build clinical deep-learning pipelines for healthcare datasets in PyHealth.
npx skills add https://github.com/jasrajtulsi/GRAD-SCOPE --skill pyhealth-jasrajtulsi
Please help me install this Agent Skill. Skill: pyhealth Source: https://github.com/jasrajtulsi/GRAD-SCOPE/tree/main/.claude/skills/pyhealth Command: npx skills add https://github.com/jasrajtulsi/GRAD-SCOPE --skill pyhealth-jasrajtulsi
PyHealth helps you turn messy healthcare data into reusable deep-learning pipelines without hand-wiring the entire workflow. It reduces the friction of loading EHR, signal, or imaging datasets, matching them to the right prediction task, and training a clinically meaningful model with the right evaluation metrics.
Ask me to build, troubleshoot, or explain a PyHealth pipeline for your healthcare dataset, task, model, and metric, and I will produce the correct end-to-end setup.
High-intent search queries and answers about installing and using this skill.
To build a MIMIC-III mortality prediction pipeline, you align the Dataset, Task, Model, Trainer, and Metrics components. You load MIMIC-III, define the mortality task, train a model like Transformer or RETAIN, and evaluate with clinically appropriate metrics.
Yes, you can use PyHealth for sleep staging and EEG event detection. The framework supports signal datasets like SleepEDF, allowing you to set up EEG event detection tasks, pair them with appropriate models, and train through the unified trainer.
Yes, the workflow supports medical code mapping for ICD, ATC, NDC, RxNorm, and CCS codes. You can look up these medical codes or cross-map between different coding systems for feature engineering and clinical interpretation.
For multilabel ICD coding tasks, the best way is to align your Dataset, Task, Model, and Trainer, then select the correct monitor for multilabel training. This ensures the model is evaluated with the appropriate metrics for multiple output labels.
Patient-level splitting is required for healthcare machine learning datasets to prevent data leakage. By splitting data at the patient level, you ensure that records from the same patient do not appear in both training and validation sets, maintaining valid evaluation.
Yes, you can train a readmission prediction model using eICU and OMOP datasets. PyHealth provides dataset-specific constructors to load these EHR sources, define the readmission task, and train the model using the unified trainer.
crazymsnBuild clinical ML pipelines with datasets, tasks, models, and trainers.
Estrella-231Build clinical deep-learning pipelines with Dataset, Task, Model, Trainer, and Metrics.
dralkhBuild clinical machine learning pipelines from raw patient data to evaluated predictions.
shushuznBuild clinical deep learning pipelines using PyHealth’s unified workflow.
