Goal: apply the GP-LVM to time-series data; GP-LVM is a generalization of PCA (PCA a special case when the covariance kernel is linear). Can we define an analogous generatlization of tICA?


In [ ]: