I am trying to build a Linear Regression model using a not so big dataset. I'm more comfortable doing classification and I am not really an expert in regression.
In classification, I was used to build train/test sets containing, more or less, the same number of elements for each class. I was wondering whether there is something similar for regression: my dataset is composed of around 100 rows (elements, like animals), but some elements are much more representative than others. Is it possible to create (for instance, with SKlearn) train/test sets which are balanced? I was thinking about the possibility to pass a vector containing the labels of each element to the function doing the split.
