Algorithm to use for below problem



Hi Expert,
I am learning algorithm through self tutorial. I am not able to decide which algorithm can be used for below scenario. Could you please help me.
Suppose a patient took paracetamol drug and develop several adverse event (say 5). I would like to find out the worst event among all based on certain categorical parameters.
It means if parameter is present , then ‘1’ else ‘0’.

The algorithm should check the cond. for first parameter. if yes, then it gives some score, then move to second parameter, if yes, again give some score. In this way, I need to traverse till end for the single patient. But in my dataset, more than 10000 patients are there, what could be the solution.

I am thinking of decision tree. But it is not working as unique Pat_id is there. I want to find the adverse event for every patients available in the database.


@abraham30 you can go ahead with the decision tree model, use Param1 to Param7 as independent variables and Adv.Event as the target/dependent variable. It is ok not use Pat_ID in the model.


Thanks Pjoshi for your response. If I will not use PatID in the program, then how can I differentiate the same event when it occurs in two or more patients. If the event are unique, then it will be easier to retrieve which patient is showing this event.
If same type of events (suppose vomiting in pat_001 and pat_003) are happening in more than one patid, how can I find those ID as there are thousand of records.


seems to me like a case of data mining - use Apriori algorithmn


Thanks narayanee for the suggestion. To run apriori algorithm, I think I need to exclude pat_id and Adv. Event. . If I exclude these two important column, then how can I find which event is having association with the particular patient. Its very difficult to find it from dataframe. In that scenario, how can I proceed. I am sorry for asking silly question


Abraham, what do u exactly want to predict?