Predicting Hospital Length of Stay

Implemented with Microsoft Machine Learning Services

SQL Database Tables


Below are the different data sets that you will find in your database after deployment.

         
Table Description
LengthOfStay Raw data about each hospital admission with missing values replaced by the mode or mean
LoS Analytical data set: cleaned data with new features
Train_Id List of encounter IDs that will go to the training set
Models Table storing the trained models
Boosted_PredictionPrediction results when testing the trained boosted trees (rxFastTrees)
Metrics Performance metrics of the models tested
LoS_Predictions Testing set with the predicted dates of the two trained models for discharge