Nonnegative Matrix Factorization with Gaussian Process Priors

<table>Toy example:
root mean squared error (RMSE) with respect to the noisy data, the underlying
noise-free data, and the true underlying nonnegative factors. The CNMF solution
fits the noisy data slightly better, but the GPP-NMF solution fits the
underlying data much better.</table>

Computational Intelligence and Neuroscience

fig3

Figure 3

Figure 3: Nonnegative Matrix Factorization with Gaussian Process Priors 