Multi-Class Semi-Supervised Kernel Spectral Clustering

- Only few labelled but large amount of unlabelled instances are available!

- Guiding an unsupervised model with additional expert knowledge (labels)!

- Discovering hidden clusters where expert didn't label!

- Learning from partially labelled datasets!

- Out-of-sample predication!

- Low dimensional embedding!

- Easy to solve: a linear system of equations.

* Nice properties ha! Read more here:

S. Mehrkanoon, C. Alzate, R. Mall, R. Langone, J. A. K. Suykens, "Multi-class semi-supervised learning based upon Kernel Spectral Clustering", IEEE Transactions on Neural Networks and Learning Systems, Vol. 26, Mar.2015, pp. 720-733.[PDF]