Material Detail

Prequential Statistics

Prequential Statistics

This video was recorded at Machine Learning seminars at the Cambridge University Engineering Department. Machine Learning is a multidisciplinary field which aims to understand and design algorithms that automatically extract useful information from data. Since real world data are typically noisy, ambiguous and occasionally erroneous, a central requirement of a learning system is that it must be able to handle uncertainty. Probability theory provides an ideal basis for representing and manipulating uncertain knowledge, so many successful algorithms in machine learning are based on probabilistic i.e. Bayesian inference. Bayesian inference provides a principled framework for machine learning, but exact inference is often intractable, so most algorithms rely on approximations such as variational methods or Markov chain Monte Carlo. More >


  • 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.