Given the product of the first purchase, what will the customer buy on her next purchase?
Say for example I have the following table of historical purchases:
CustomerID OrderID ProductType MonthsBetweenPurch OrderNumber
----------------------------------------------------------------------
130293699 1013448571 Womenswear 0 1
130293699 1013448571 Tops 0 1
130293699 1013574573 Tops 0 1
130293699 1013448577 Sweaters 1.5 2
130194668 1013735788 Tops 0 1
130194668 1013445564 Accessories 1 2
130194668 1013448575 Sweaters 0 1
130675885 1013426777 Tops 2 2
130675885 1013441869 Underwear 0 1
130675885 1013448444 Sweaters 0 1
130675885 1013448444 Accessories 0 1
130675885 1013448444 Accessories 0 1
where the MonthsBetweenPurch
is the number of months from purchase 1 to purchase 2 and OrderNumber
is whether its the first order or the second order. The other fields are self-explanatory.
The question is: Given that a customer buys certain product/products in the first purchase, what product/products will the customer most likely buy in the second purchase?
I understand that this is an association problem but I'm not sure how I should structure my data and and what algorithm would be the best to use. Any suggestions on how I could attack this problem? I'm a very junior DS so I'm still learning new stuff.
Topic association-rules
Category Data Science