I would like to ask for a proposal for a machine learning model that would be suitable for the following problem:
I have a training set where each element of type A corresponds to a certain number of elements of type B (both type A and type B elements are described by certain specific columns), i.e.:
A_1 -> B_1, A_1 -> B_2, A_1 -> B_3
A_2 -> B_4, A_2 -> B_5
A_3 -> B_6 ...
For a new element of type A, I want to select the most matching elements of type B from a certain set and indicate how many percent the elements match each other.
Can the problem formulated in this way be solved by machine learning or should I reformulate it in some way?