This deck will discuss application of Matrix Factorization in Machine Learning. It will discuss Least Square Matrix Factorization, Poisson Matrix Factorization.
6. Poisson Matrix Factorization (NMF)
• Suitable if you are looking for non-negative factors
𝑣 = 𝑒/01
𝑙𝑟 𝑛/𝑛!
• Leads to the well known Generalized KL-Divergence
max∑ (𝑛𝑖𝑗log𝑙𝑖′𝑟𝑗 − 𝑙𝑖′𝑟𝑗)$%
• Well known update equations exist*
* “Generalized Nonnegative Matrix Approximations with Bregman Divergences” by Dhillon and Sra in NIPS 2005.
* “Distributed Nonnegative Matrix Factorization for Web-Scale Dyadic Data Analysis on MapReduce” by Liu et al in WWW 2010.