Material Detail

Modern Nonparametric Statistics on Modern Big Data

Modern Nonparametric Statistics on Modern Big Data

This video was recorded at NIPS Workshops, Lake Tahoe 2012. Modern data is increasing very large in terms of both the number of objects and the number of dimensions. While in a statistical sense massive amounts of data make nonparametric methods entirely appropriate, their computational cost has made practitioners typically conclude that they are not possible in such scenarios. I will review a few common conceptual classes of nonparametric methods, including both classical and modern variants, then review recent algorithmic advances which can make nonparametric methods tractable


  • User Rating
  • Comments
  • Learning Exercises
  • Bookmark Collections
  • Course ePortfolios
  • Accessibility Info

More about this material


Log in to participate in the discussions or sign up if you are not already a MERLOT member.