Augmenting Matrix Factors

Hi,
I implemented the exercise provided at [spark mllib training](http://ampcamp.berkeley.edu/big-data-mini-course/movie-recommendation-with-mllib.html). But in final part, it is recommended to implement a matrix factorization to improve the algorithm. I could not find any example to do that. Is there anyone who can show me a way to handle this?
