Dimensionality reduction for clustered data sets
IEEE Trans. Pattern Analysis Machine Intelligence
We present a novel probabilistic latent variable model to perform linear
reduction on data sets which contain clusters. We prove that the maximum
likelihood solution of the model is an unsupervised generalisation of
linear discriminant analysis. This provides a completely new approach to
one of the most established and widely used classification algorithms.
The performance of the model is then demonstrated on a
number of real and artificial data sets.