Making product recommendations
The practical application of machine learning we will implement is making product recommendations on an ecommerce website with the goal being to increase the value of a customer order.
The problem is how to decide what products to recommend to the visitor.
Problem analysis
On 2nd October 2006, Netflix started an open prize challenge for the best algorithm for predicting customer ratings for films based only on previous ratings. The dataset that Netflix provided had 17,000 movies, 500,000 users, and 100 million ratings. For example, user 437822 gives movie 12934 a rating of 4 out of 5.
Singular Value Decomposition (SVD) is a decomposition method for reducing a matrix to make later calculations simpler and Simon Funk shared with the community how he and his team used it to get near the top of the rankings during the competition.
More Information: You can read more about Simon Funk's use of SVD at the following link: https...