Hi all,

I’m hoping to get some guidance and inputs on what kind of model/architecture/techniques can be applied to model the following data.

This is how 1 data point looks like -

My output variable(`y`

), that I am trying to predict is a `contiguous`

variable - price

My input data(`x`

) comes from 2 DB tables, say `A`

and `B`

where there is a *one-to-many* relationship from `A`

→ `B`

.

So, 1 record from table `A`

and a number of records from table `B`

, serve as 1 input sample(`x`

). Number of records picked from table `B`

can be different for different input samples(`x`

).

Also, most of the data in these tables are of `continuous`

and `categorical`

nature. There is 1 column in table `B`

that is a text field but it seems to be strongly correlated with other input columns.

Any help is much appreciated!!