[coeff, score, pcvar] = ppca(Y, K)
[coeff, score, pcvar, mu, v, S] = ppca(Y, K, Name, Value)
ppca computes a probabilistic principal component model for a real data matrix. Missing values encoded as NaN are estimated iteratively.
Supported name-value options are W0, v0, and Options. The Options structure can be created with statset and supports Display, MaxIter, TolFun, and TolX.
The structure S contains W, Xexp, Recon, v, NumIter, RMSResid, and nloglk fields.
Y = [1 2 3; 2 NaN 5; 4 5 8; 5 7 NaN; 7 8 13];
[coeff, score, pcvar, mu, v, S] = ppca(Y, 2)| Version | Description |
|---|---|
| 2.0.0 | initial version |