Has there been work in featuretools (or an additional Python package) that would integrate it with a common ML library, such as sklearn? E.g., it would be nice to test a feature for its predictive power, and if it's high enough, generate more features like it (e.g., use the same initial variable). In other words, can the process of generating new features be guided by their predictive power?
Asked
Active
Viewed 137 times
0
-
There's been no work inside of Featuretools on this at this point in time. We would be very interested in seeing a proof of concept of this if you built it! – Max Kanter Aug 14 '18 at 16:11
-
I'm voting to close this question as off-topic because it is not about programming – desertnaut Aug 29 '18 at 11:54