A platform for analysis & development of machine learning models using large de-identified healthcare datasets.
Switch branches/tags
Nothing to show
Clone or download
Latest commit ab6981b Mar 4, 2017

README.md

Computational Healthcare Library

The goal of Computational Healthcare Library is to help computer scientists perform high impact healthcare research by providing a simple interface to large publicly available healthcare datasets.

Please visit Computational Healthcare Library for more information and documentations

With Computational Healthcare library you can:

  • Build Outcomes prediction engine
  • Load & analyze data from up to 200 Million visits & 70 Million patients
  • Specify aggregation strategies and compute aggregate statistics in a privacy preserving manner
  • Build embedding models, perform transfer learning, predict rehospitalizations/revisits using TensorFlow
  • Benchmark results against baseline algorithms trained on publicly available datasets
  • In future it can be used for testing Differential Privacy algorithms for computing aggregate statistics & privacy preserving Machine Learning

Please note that this repository does not contains any data, nor do we provide any data. You should acquire the datasets on your own from AHRQ or other state agencies.

Copyright

Copyright Cornell University 2017; All rights reserved; Please contact us for more information.