A Gaussian Process Latent Variable Model for Subspace Clustering

Algorithm 1

The optimization process of C-GPLVM.

	Input: training set , dimension of latent variable, hyperparameters , , .
	Output: parameter and the clustering result.
(1)	Pretrain the following model to initialize the latent variable
(2)	while ()
(3)	With fixed , minimize (18) to obtain , , and .
(4)	With fixed , , and , minimize (20) to obtain .
(5)	Calculate the difference of the objective function between the last two iterations
(6)	end while
(7)	Compute the latent variable based on the learned parameter .
(8)	Run k-means algorithm on to get the final clustering result.