I want to train a machine learning model that uses the following parameters to predict the optimal treatment choice. Say I have data on patient characteristics, imaging data, and which treatment was given to the patient (X, Y, Z), the outcome is complete remission. i want to note that for example if patient was given X and he went to remission then X is the drug of choice for that patient. If patient was give Y and it didnt work he can be given Z and if Z works then Z is the drug of choice for that patient.

​

The outcome here is not to predict the outcome, it is to predict the ideal treatment. The other way is I can predict outcome for each of the treatment choice.

​

How can one go about this?

​

​

Thank you all