Vincent Granville makes some provocative statements about the interaction of statistics and data science. I often worked with professional statisticians for problems that had regulatory dimensions, and they would have disagreed here. Yet analytics applications have always picked and chosen methods that provided the best results, with the data available and the kind of decision being modeled. Faster methods and more data have allowed us to broaden our choices. What we used was certainly in his list of methods. If these are old or new statistics is a matter of opinion, often dependent on your training and experience. Useful piece to read.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment