MLHO (pronounced as melo) is a thinkin’ Machine Learning framework that implements iterative sequential representation mining, and feature and model selection to predict health outcomes.
Installation
You can install the released version of mlho from Github with:
devtools::install_github("hestiri/mlho")Data model
To implement MLHO you’ll need 2 tables, which can be extracted from any clinical CMD. The current examples are based on the i2b2 star schema.
1- a table with outcome labels (called labeldt) and patient numbers
| patient_num | label |
| character | factor |
2- a patient clinical data table (called dbmart) with 3 columns. Concepts are used as features by MLHO.
| patient_num | start_date | phenx |
| character | date | character |
The column phenx contains the entire feature space. In an i2b2 data model, for instance, this column is the equivalent of concept_cd.
3- a demographic table is optional, but recommended.
| patient_num | age | gender | … |
| character | character | character | character |
see the instructions on how to use the MLHO package on the articles page