wolffd@0: wolffd@0: wolffd@0: wolffd@0: Netlab Reference Manual demev2 wolffd@0: wolffd@0: wolffd@0: wolffd@0:

demev2 wolffd@0:

wolffd@0:

wolffd@0: Purpose wolffd@0:

wolffd@0: Demonstrate Bayesian classification for the MLP. wolffd@0: wolffd@0:

wolffd@0: Synopsis wolffd@0:

wolffd@0:
wolffd@0: demev2
wolffd@0: wolffd@0: wolffd@0:

wolffd@0: Description wolffd@0:

wolffd@0: A synthetic two class two-dimensional dataset x is sampled wolffd@0: from a mixture of four Gaussians. Each class is wolffd@0: associated with two of the Gaussians so that the optimal decision wolffd@0: boundary is non-linear. wolffd@0: A 2-layer wolffd@0: network with logistic outputs is trained by minimizing the cross-entropy wolffd@0: error function with isotroipc Gaussian regularizer (one hyperparameter for wolffd@0: each of the four standard weight groups), using the scaled wolffd@0: conjugate gradient optimizer. The hyperparameter vectors alpha and wolffd@0: beta are re-estimated using the function evidence. A graph wolffd@0: is plotted of the optimal, regularised, and unregularised decision wolffd@0: boundaries. A further plot of the moderated versus unmoderated contours wolffd@0: is generated. wolffd@0: wolffd@0:

wolffd@0: See Also wolffd@0:

wolffd@0: evidence, mlp, scg, demard, demmlp2
wolffd@0: Pages: wolffd@0: Index wolffd@0:
wolffd@0:

Copyright (c) Ian T Nabney (1996-9) wolffd@0: wolffd@0: wolffd@0: wolffd@0: