3 ms·
This is my explanation as an armchair observant of the science of prediction. Nice high-level overview of the Netflix Prize: the problem, the competitors, the e
by wallflower 17y ago
This is my explanation as an armchair observant of the science of prediction. Nice high-level overview of the Netflix Prize: the problem, the competitors, the evolving science of CF - http://www.nytimes.com/2008/11/23/magazine/23Netflix-t.html?_r=2&pagewanted=all http://www.nytimes.com/2008/11/23/magazine/23Netflix-t.html?...
CF = Collaborative Filtering (e.g. "Customers who liked X also liked Y")
KNN = k-Nearest Neighbor and a very popular CF algorithm (the winners of the Netflix Prize used KNN in part of their program)
Pearson correlation = There were 17,770 movies in the Netflix Prize dataset. Necessary input to KNN model. Which movies might be considered neighbors (e.g. similar) to each other.