There are 3 tables:
Information about investiment funds; Information about customers who come to us requesting credit from those investiment funds; Transactions between customers and investiment funds (from the two tables explained above). Here's an image to illustrate the scenario:
There are 3 tables:
Information about investiment funds; Information about customers who come to us requesting credit from those investiment funds; Transactions between customers and investiment funds (from the two tables explained above). Here's an image to illustrate the scenario:
enter image description here
My goals is to relate customers to funds in order to recommend possible transactions to future customers based on the pattern found between funds and customers.
My first idea is to cluster the Funds table and Customers table (possibly with kmeans) and try to find relation between these clusters, but I'm not sure how I'll do it yet, cause I would have to test every possible number of centroids of each model and find the ones that fit best.
Any suggestion?