Vagus HRV Code
Supplementary data for the manuscript “Decoding vagal contributions to fetal heart rate variability” by C.L. Herry et al.
Batch Computation of HRV Metrics using NeuroKit API
Navigate to the Jupyter notebook to see a demonstration of batch processing for HRV metrics.
Predicting Maternal Morbidity
A project considering the impact of ethnicity and socioeconomic status on rehospitalization and postpartum depression. You can obtain the nuMoM2b dataset here.
Introduction
The advent of machine learning (ML) in medicine has opened opportunities for harnessing its power to predict patient outcomes based on data in electronic medical records (EMRs). While singular features may be weak learners, combining them in an ML framework can create highly predictive models of health outcomes.
The information-rich nuMoM2b dataset allows us to tackle two important hypotheses:
- Demographic and socioeconomic characteristics influence rehospitalization outcomes.
- Demographic, socioeconomic, and psychiatric history influence postpartum depression outcomes.
Methods
We used R Studio to create a reproducible coding notebook that ingests the nuMoM2b dataset. The software is optimized to run on standard desktop computers, requiring no special computing resources to encourage wide utilization. All steps can be easily reproduced, and the models can be updated as new data become available.
Results
View the results by opening the following HTML files:
Credit & References
Credit for the decision tree visualization code goes to Gregory Kanevsky.
Updated HRV Pipeline
Supplementary data for the MethodsX manuscript “Comprehensive HRV estimation pipeline in Python using Neurokit2” by M.G. Frasch.
ChatGPT Demo for ML Code Generation
View a sample conversation demonstrating how to use ChatGPT to generate machine learning models in Python and R/H2O.