Thatcher, Levi
2016-12-30
<p>This is the first full release of healthcare.ai for R. Note that</p>
<p>This release encompasses basic healthcare ML functionality:</p>
<ul>
<li>Model comparison between random forest, lasso, and mixed model algorithms</li>
<li>Feature selection via lasso and random forest feature importance</li>
<li>Model deployment to SQL Server, providing top-three most important features</li>
<li>Imputation (column mean for numeric and column mode for categorical)</li>
<li>Hyperparameter tuning using mtry and number of trees for random forest</li>
<li>ROC and PR Curves plotted</li>
<li>Model performance evaluated via AU_ROC and AU_PR</li>
<li>To assist, these functions are available:
<ul>
<li>groupedLOCF (for longitudinal imputation)</li>
<li>findTrends (for Nelson rule 3)</li>
<li>convertDateTimeColToDummies to create data-based features from datetime stamp</li>
<li>calculateAllCorrelations for correlations across all numeric cols in data frame</li>
<li>calculateTargetedCorrelations for correlations across numeric cols and specific column</li>
</ul>
</li>
</ul>
<p>For this release the following infrastructure: is in place:</p>
<p>AppveyorCI<br>
Roxygen2<br>
mkdocs for website (which files reside in documentation repo)</p>
https://doi.org/10.5281/zenodo.999329
oai:zenodo.org:999329
Zenodo
https://github.com/HealthCatalyst/healthcareai-r/releases/tag/v0.1.10
https://doi.org/10.5281/zenodo.999328
info:eu-repo/semantics/openAccess
MIT License
https://opensource.org/licenses/MIT
healthcare
machine learning
data science
R
data
healthcareai-r v0.1.10
info:eu-repo/semantics/other