This random forest model predicts a patient's longitudinal K-means cluster assignment based on absolute neutrophil count (ANC) data, defined as follows:

- Very good: high ANC nadir and rapid recovery
- Good: low ANC nadir and rapid recovery
- Poor: low ANC nadir and delayed recovery
- Very poor: aplastic phenotype

These 5 ANC values demonstrated the highest accuracy to predict cluster assignments in our test set. If ANC values from these specific days are not available, you can enter ANC values from +/- 3 days, though the substitutions may affect the performance of the model.

Additional information on the development and performance of these models can be found in the manuscript: TBD.