I am very pleased to announce that the Readmission Demo has been released as open source. Many thanks to the Solution Factory team that worked hard on making this possible.
Here are the changes:
- It is now running on top of IRIS Community 2019.3.0.308.0.
- It has synthetic data based on Synthea instead of the real data we were using before. But the machine learning models are 100% real
- Image RRLACESrv has been renamed to “image-riskengine”
- Image “readmissionsrv” has been renamed to “image-risksrv"
- We refactored a lot of the code on image-riskengine to look better. Expect to see:
- Nicer cuber definitions and support utility methods.
- Nicer data lake data model
- The code to load synthea FHIR JSON bundles into IRIS data lake
- There is a new Zeppelin notebook with the ML model built for the synthetic data. We left the old zeppelin mode there as well as evidence that this works with real data as well.
The demo can be found found on GitHub:
And it will soon be available on Open Exchange (waiting for approval):
Both pages will show instructions about how to run it. There is no need for building it on your PCs anymore. You can do it, if you want to change it. But you don’t need to.
A video about this demo is on the works!
PS: If you want to build it on your PC, try doing it with fewer Synthea patients first. Generate the 5000 patients takes time. Load them into the database takes time too.
Register or login to poll