The problem for short term lenders, or new challenger lending start ups whether loan companies or challenger banks, is that most often than not “retro” data for past historic cohorts of a population don’t necessarily represent the same population as who you are lending to. For example if you are setting up a loan company in Mexico, Africa orLatvia then you want a model built on data from that specific population, don’t you? Or a loan model just for females? Traditionally modelling techniques are biased and fail to be diverse and inclusive.
Find out more in our white-paper here https://gallery.mailchimp.com/20a3b80bd19628a20c3dbf7a9/files/f36c37b2-5b58-4525-bb6d-af1e881e0d98/Heimdall_how_Explainable_XAI_Smart_Data_is_better_for_creating_Loan_Propensity_Models_that_Embrace_Inclusivity_and_Diversity.pdf