Ūnus

Ūnus is a machine learning based patient matching engine built for the ONC Patient Matching Challenge.

Detailed write up are available here:

http://mindfulmachines.io/blog/2017/7/23/onc-patient-matching-challenge-part1

http://mindfulmachines.io/blog/2017/7/23/onc-patient-matching-challenge-part-2

Technology

The technologies used to build the matching engine are:

Scala
Play
Spark
PostgreSQL
Docker

Installing

Clone this repository:

git clone https://github.com/mindfulmachines/unus.git

Install Java 8

Install SBT

Install Docker

Running

Place the FInalDataset.csv file in the data/ folder.

Configure a docker volume for PostgreSQL:

docker volume create postgresql

Start the postgres server:

docker-compose up

Optionally configure the database url by editing conf/application.conf. Specifically if you're using Docker Machine (Windows 7) the database url is probably jdbc:postgresql://192.168.99.100/postgres.

Start the play application:

sbt "runMain PlayApp"

Go to localhost:9000 to view the UI. The first time it may take a while for the UI to be available as data is loaded into the DB. You can also check localhost:4040 for how Spark is progressing.