Prediction challenges#

The MNIST dataset is a library of handwritten digits#

  • A lot of the coolest applications of statistical and machine learning are prediction challenges.

http://www.stanford.edu/class/stats202/figs/mnist.jpg

Fig. 1 MNIST dataset#

  • In a prediction challenge, you are given a training set of images of handwritten digits, which are labeled from 0 to 9.

  • You are also given a test set of handwritten digits, which are not identified.

  • Your job is to assign a digit to each image in the test set.


The Netflix prize#

Netflix popularized prediction challenges by organizing an open, blind contest to improve its recommendation system.

http://www.stanford.edu/class/stats202/figs/tikz/netflix1.png

http://www.stanford.edu/class/stats202/figs/tikz/netflix2.png

http://www.stanford.edu/class/stats202/figs/tikz/netflix3.png

The prize was $1 million.

(Cue Dr. Evil jokes if anyone knows Austin Powers movies…)