healthcareai
The aim of healthcareai
is to make machine learning easy on healthcare data and make you look very smart. The package has two main goals:
-
Allow one to easily develop and compare models based on tabular data, and deploy a best model that pushes predictions to either databases or flat files.
-
Provide tools related to data cleaning, manipulation, and imputation.
For those starting out
Note: if you're setting up R on an ETL server, don't download RStudio--simply open up RGui
Install the latest release on Windows
Open RStudio and work in the console
install.packages('healthcareai')
If you don't have admin rights on the machine you are working on, and
library(healthcareai)
throws an error about packages not being available, you can likely solve the problem by defining a custom location in which to store R packages. To do this, open the Control Panel and click through User Accounts -> User Accounts -> Change my environment variables, and add a variable calledR_LIBS_USER
with the value being a path to a folder where you want to keep R packages. For example, you might create a new directory:C:\Users\your.name\Documents\R\my_library
and use that to store your R packages. Then restart R Studio, runinstall.packages("healthcareai")
andlibrary(healthcareai)
again and all should be well.
How to install the latest version on macOS
Open RStudio and work in the console
install.packages('healthcareai')
How to install latest version on Ubuntu (Linux)
- An Ubuntu 14.04 Droplet with at least 1 GB of RAM is required for the installation.
- Follow steps 1 and 2 here to install R
- Run
sudo apt-get install libiodbc2-dev
- Run
sudo apt-get install unixodbc unixodbc-dev
- After typing
R
runinstall.packages('healthcareai')
Install the bleeding edge version (for folks providing contributions)
Open RStudio and work in the console
library(devtools)
devtools::install_github(repo='HealthCatalyst/healthcareai-r')
Tips on getting started
Built-in examples
Load the package you just installed and read the built-in docs
library(healthcareai)
?healthcareai
Website examples
See our docs website
Join the community
Read the blog and join the slack channel at healthcare.ai
What's new?
The CRAN 1.0.0 release features:
- Added:
- Kmeans clustering
- XGBoost multiclass support
- findingVariation family of functions
- Changed:
- Develop step trains and saves models
- Deploy no longer trains. Loads and predicts on all rows.
- SQL uses a DBI back end
- Removed:
testWindowCol
is no longer a param.- SQL reading/writing is outside model deployment.
For issues
- Double check that the code follows the examples in the built-in docs
library(healthcareai)
?healthcareai
-
Make sure you've thoroughly read the descriptions found here
-
If you're still seeing an error, file an issue on Stack Overflow using the healthcare-ai tag. Please provide
- Details on your environment (OS, database type, R vs Py)
- Goals (ie, what are you trying to accomplish)
- Crystal clear steps for reproducing the error
Contributing
You want to help? Woohoo! We welcome that and are willing to help newbies get started.
First, see here for instructions on setting up your development environment and how to contribute.