Sign up to take part
Registered users can ask their own questions, contribute to discussions, and be part of the Community!
Registered users can ask their own questions, contribute to discussions, and be part of the Community!
Hi everyone,
I'm working on a problem where basically I have one categorical variable and two numerical variables. For each modality of the qualitative variable, the relation between the two quantitative variables is almost linear.
From this observation, I'd like to model one linear regression per modality and ultimately obtain the N sets of linear regression parameters corresponding to the N modalities of the categorical variable.
Do you know if there's a way to implement such a model in DSS without using custom code ?
Thanks in advance.
Hi,
What you describe sounds similar to model partitioning, in which you train different models for each value of a partition column. Have a look:
https://doc.dataiku.com/dss/latest/machine-learning/partitioned.html
I hope this helps.