wolffd@0: wolffd@0: wolffd@0: wolffd@0: NETLAB Reference Documentation wolffd@0: wolffd@0: wolffd@0: wolffd@0:

NETLAB Online Reference Documentation

wolffd@0: Welcome to the NETLAB online reference documentation. wolffd@0: The NETLAB simulation software is designed to provide all the tools necessary wolffd@0: for principled and theoretically well founded application development. The wolffd@0: NETLAB library is based on the approach and techniques described in Neural wolffd@0: Networks for Pattern Recognition (Bishop, 1995). The library includes software wolffd@0: implementations of a wide range of data analysis techniques, many of which are wolffd@0: not widely available, and are rarely, if ever, included in standard neural wolffd@0: network simulation packages. wolffd@0:

The online reference documentation provides direct hypertext links to specific Netlab function descriptions. wolffd@0:

If you have any comments or problems to report, please contact Ian Nabney (i.t.nabney@aston.ac.uk) or Christopher Bishop (c.m.bishop@aston.ac.uk).

Index wolffd@0:

wolffd@0: An alphabetic list of functions in Netlab.

wolffd@0:

wolffd@0:
wolffd@0: conffig
wolffd@0: Display a confusion matrix. wolffd@0:
wolffd@0: confmat
wolffd@0: Compute a confusion matrix. wolffd@0:
wolffd@0: conjgrad
wolffd@0: Conjugate gradients optimization. wolffd@0:
wolffd@0: consist
wolffd@0: Check that arguments are consistent. wolffd@0:
wolffd@0: convertoldnet
wolffd@0: Convert pre-2.3 release MLP and MDN nets to new format wolffd@0:
wolffd@0: datread
wolffd@0: Read data from an ascii file. wolffd@0:
wolffd@0: datwrite
wolffd@0: Write data to ascii file. wolffd@0:
wolffd@0: dem2ddat
wolffd@0: Generates two dimensional data for demos. wolffd@0:
wolffd@0: demard
wolffd@0: Automatic relevance determination using the MLP. wolffd@0:
wolffd@0: demev1
wolffd@0: Demonstrate Bayesian regression for the MLP. wolffd@0:
wolffd@0: demev2
wolffd@0: Demonstrate Bayesian classification for the MLP. wolffd@0:
wolffd@0: demev3
wolffd@0: Demonstrate Bayesian regression for the RBF. wolffd@0:
wolffd@0: demgauss
wolffd@0: Demonstrate sampling from Gaussian distributions. wolffd@0:
wolffd@0: demglm1
wolffd@0: Demonstrate simple classification using a generalized linear model. wolffd@0:
wolffd@0: demglm2
wolffd@0: Demonstrate simple classification using a generalized linear model. wolffd@0:
wolffd@0: demgmm1
wolffd@0: Demonstrate density modelling with a Gaussian mixture model. wolffd@0:
wolffd@0: demgmm3
wolffd@0: Demonstrate density modelling with a Gaussian mixture model. wolffd@0:
wolffd@0: demgmm4
wolffd@0: Demonstrate density modelling with a Gaussian mixture model. wolffd@0:
wolffd@0: demgmm5
wolffd@0: Demonstrate density modelling with a PPCA mixture model. wolffd@0:
wolffd@0: demgp
wolffd@0: Demonstrate simple regression using a Gaussian Process. wolffd@0:
wolffd@0: demgpard
wolffd@0: Demonstrate ARD using a Gaussian Process. wolffd@0:
wolffd@0: demgpot
wolffd@0: Computes the gradient of the negative log likelihood for a mixture model. wolffd@0:
wolffd@0: demgtm1
wolffd@0: Demonstrate EM for GTM. wolffd@0:
wolffd@0: demgtm2
wolffd@0: Demonstrate GTM for visualisation. wolffd@0:
wolffd@0: demhint
wolffd@0: Demonstration of Hinton diagram for 2-layer feed-forward network. wolffd@0:
wolffd@0: demhmc1
wolffd@0: Demonstrate Hybrid Monte Carlo sampling on mixture of two Gaussians. wolffd@0:
wolffd@0: demhmc2
wolffd@0: Demonstrate Bayesian regression with Hybrid Monte Carlo sampling. wolffd@0:
wolffd@0: demhmc3
wolffd@0: Demonstrate Bayesian regression with Hybrid Monte Carlo sampling. wolffd@0:
wolffd@0: demkmean
wolffd@0: Demonstrate simple clustering model trained with K-means. wolffd@0:
wolffd@0: demknn1
wolffd@0: Demonstrate nearest neighbour classifier. wolffd@0:
wolffd@0: demmdn1
wolffd@0: Demonstrate fitting a multi-valued function using a Mixture Density Network. wolffd@0:
wolffd@0: demmet1
wolffd@0: Demonstrate Markov Chain Monte Carlo sampling on a Gaussian. wolffd@0:
wolffd@0: demmlp1
wolffd@0: Demonstrate simple regression using a multi-layer perceptron wolffd@0:
wolffd@0: demmlp2
wolffd@0: Demonstrate simple classification using a multi-layer perceptron wolffd@0:
wolffd@0: demnlab
wolffd@0: A front-end Graphical User Interface to the demos wolffd@0:
wolffd@0: demns1
wolffd@0: Demonstrate Neuroscale for visualisation. wolffd@0:
wolffd@0: demolgd1
wolffd@0: Demonstrate simple MLP optimisation with on-line gradient descent wolffd@0:
wolffd@0: demopt1
wolffd@0: Demonstrate different optimisers on Rosenbrock's function. wolffd@0:
wolffd@0: dempot
wolffd@0: Computes the negative log likelihood for a mixture model. wolffd@0:
wolffd@0: demprgp
wolffd@0: Demonstrate sampling from a Gaussian Process prior. wolffd@0:
wolffd@0: demprior
wolffd@0: Demonstrate sampling from a multi-parameter Gaussian prior. wolffd@0:
wolffd@0: demrbf1
wolffd@0: Demonstrate simple regression using a radial basis function network. wolffd@0:
wolffd@0: demsom1
wolffd@0: Demonstrate SOM for visualisation. wolffd@0:
wolffd@0: demtrain
wolffd@0: Demonstrate training of MLP network. wolffd@0:
wolffd@0: dist2
wolffd@0: Calculates squared distance between two sets of points. wolffd@0:
wolffd@0: eigdec
wolffd@0: Sorted eigendecomposition wolffd@0:
wolffd@0: errbayes
wolffd@0: Evaluate Bayesian error function for network. wolffd@0:
wolffd@0: evidence
wolffd@0: Re-estimate hyperparameters using evidence approximation. wolffd@0:
wolffd@0: fevbayes
wolffd@0: Evaluate Bayesian regularisation for network forward propagation. wolffd@0:
wolffd@0: gauss
wolffd@0: Evaluate a Gaussian distribution. wolffd@0:
wolffd@0: gbayes
wolffd@0: Evaluate gradient of Bayesian error function for network. wolffd@0:
wolffd@0: glm
wolffd@0: Create a generalized linear model. wolffd@0:
wolffd@0: glmderiv
wolffd@0: Evaluate derivatives of GLM outputs with respect to weights. wolffd@0:
wolffd@0: glmerr
wolffd@0: Evaluate error function for generalized linear model. wolffd@0:
wolffd@0: glmevfwd
wolffd@0: Forward propagation with evidence for GLM wolffd@0:
wolffd@0: glmfwd
wolffd@0: Forward propagation through generalized linear model. wolffd@0:
wolffd@0: glmgrad
wolffd@0: Evaluate gradient of error function for generalized linear model. wolffd@0:
wolffd@0: glmhess
wolffd@0: Evaluate the Hessian matrix for a generalised linear model. wolffd@0:
wolffd@0: glminit
wolffd@0: Initialise the weights in a generalized linear model. wolffd@0:
wolffd@0: glmpak
wolffd@0: Combines weights and biases into one weights vector. wolffd@0:
wolffd@0: glmtrain
wolffd@0: Specialised training of generalized linear model wolffd@0:
wolffd@0: glmunpak
wolffd@0: Separates weights vector into weight and bias matrices. wolffd@0:
wolffd@0: gmm
wolffd@0: Creates a Gaussian mixture model with specified architecture. wolffd@0:
wolffd@0: gmmactiv
wolffd@0: Computes the activations of a Gaussian mixture model. wolffd@0:
wolffd@0: gmmem
wolffd@0: EM algorithm for Gaussian mixture model. wolffd@0:
wolffd@0: gmminit
wolffd@0: Initialises Gaussian mixture model from data wolffd@0:
wolffd@0: gmmpak
wolffd@0: Combines all the parameters in a Gaussian mixture model into one vector. wolffd@0:
wolffd@0: gmmpost
wolffd@0: Computes the class posterior probabilities of a Gaussian mixture model. wolffd@0:
wolffd@0: gmmprob
wolffd@0: Computes the data probability for a Gaussian mixture model. wolffd@0:
wolffd@0: gmmsamp
wolffd@0: Sample from a Gaussian mixture distribution. wolffd@0:
wolffd@0: gmmunpak
wolffd@0: Separates a vector of Gaussian mixture model parameters into its components. wolffd@0:
wolffd@0: gp
wolffd@0: Create a Gaussian Process. wolffd@0:
wolffd@0: gpcovar
wolffd@0: Calculate the covariance for a Gaussian Process. wolffd@0:
wolffd@0: gpcovarf
wolffd@0: Calculate the covariance function for a Gaussian Process. wolffd@0:
wolffd@0: gpcovarp
wolffd@0: Calculate the prior covariance for a Gaussian Process. wolffd@0:
wolffd@0: gperr
wolffd@0: Evaluate error function for Gaussian Process. wolffd@0:
wolffd@0: gpfwd
wolffd@0: Forward propagation through Gaussian Process. wolffd@0:
wolffd@0: gpgrad
wolffd@0: Evaluate error gradient for Gaussian Process. wolffd@0:
wolffd@0: gpinit
wolffd@0: Initialise Gaussian Process model. wolffd@0:
wolffd@0: gppak
wolffd@0: Combines GP hyperparameters into one vector. wolffd@0:
wolffd@0: gpunpak
wolffd@0: Separates hyperparameter vector into components. wolffd@0:
wolffd@0: gradchek
wolffd@0: Checks a user-defined gradient function using finite differences. wolffd@0:
wolffd@0: graddesc
wolffd@0: Gradient descent optimization. wolffd@0:
wolffd@0: gsamp
wolffd@0: Sample from a Gaussian distribution. wolffd@0:
wolffd@0: gtm
wolffd@0: Create a Generative Topographic Map. wolffd@0:
wolffd@0: gtmem
wolffd@0: EM algorithm for Generative Topographic Mapping. wolffd@0:
wolffd@0: gtmfwd
wolffd@0: Forward propagation through GTM. wolffd@0:
wolffd@0: gtminit
wolffd@0: Initialise the weights and latent sample in a GTM. wolffd@0:
wolffd@0: gtmlmean
wolffd@0: Mean responsibility for data in a GTM. wolffd@0:
wolffd@0: gtmlmode
wolffd@0: Mode responsibility for data in a GTM. wolffd@0:
wolffd@0: gtmmag
wolffd@0: Magnification factors for a GTM wolffd@0:
wolffd@0: gtmpost
wolffd@0: Latent space responsibility for data in a GTM. wolffd@0:
wolffd@0: gtmprob
wolffd@0: Probability for data under a GTM. wolffd@0:
wolffd@0: hbayes
wolffd@0: Evaluate Hessian of Bayesian error function for network. wolffd@0:
wolffd@0: hesschek
wolffd@0: Use central differences to confirm correct evaluation of Hessian matrix. wolffd@0:
wolffd@0: hintmat
wolffd@0: Evaluates the coordinates of the patches for a Hinton diagram. wolffd@0:
wolffd@0: hinton
wolffd@0: Plot Hinton diagram for a weight matrix. wolffd@0:
wolffd@0: histp
wolffd@0: Histogram estimate of 1-dimensional probability distribution. wolffd@0:
wolffd@0: hmc
wolffd@0: Hybrid Monte Carlo sampling. wolffd@0:
wolffd@0: kmeans
wolffd@0: Trains a k means cluster model. wolffd@0:
wolffd@0: knn
wolffd@0: Creates a K-nearest-neighbour classifier. wolffd@0:
wolffd@0: knnfwd
wolffd@0: Forward propagation through a K-nearest-neighbour classifier. wolffd@0:
wolffd@0: linef
wolffd@0: Calculate function value along a line. wolffd@0:
wolffd@0: linemin
wolffd@0: One dimensional minimization. wolffd@0:
wolffd@0: maxitmess
wolffd@0: Create a standard error message when training reaches max. iterations. wolffd@0:
wolffd@0: mdn
wolffd@0: Creates a Mixture Density Network with specified architecture. wolffd@0:
wolffd@0: mdn2gmm
wolffd@0: Converts an MDN mixture data structure to array of GMMs. wolffd@0:
wolffd@0: mdndist2
wolffd@0: Calculates squared distance between centres of Gaussian kernels and data wolffd@0:
wolffd@0: mdnerr
wolffd@0: Evaluate error function for Mixture Density Network. wolffd@0:
wolffd@0: mdnfwd
wolffd@0: Forward propagation through Mixture Density Network. wolffd@0:
wolffd@0: mdngrad
wolffd@0: Evaluate gradient of error function for Mixture Density Network. wolffd@0:
wolffd@0: mdninit
wolffd@0: Initialise the weights in a Mixture Density Network. wolffd@0:
wolffd@0: mdnpak
wolffd@0: Combines weights and biases into one weights vector. wolffd@0:
wolffd@0: mdnpost
wolffd@0: Computes the posterior probability for each MDN mixture component. wolffd@0:
wolffd@0: mdnprob
wolffd@0: Computes the data probability likelihood for an MDN mixture structure. wolffd@0:
wolffd@0: mdnunpak
wolffd@0: Separates weights vector into weight and bias matrices. wolffd@0:
wolffd@0: metrop
wolffd@0: Markov Chain Monte Carlo sampling with Metropolis algorithm. wolffd@0:
wolffd@0: minbrack
wolffd@0: Bracket a minimum of a function of one variable. wolffd@0:
wolffd@0: mlp
wolffd@0: Create a 2-layer feedforward network. wolffd@0:
wolffd@0: mlpbkp
wolffd@0: Backpropagate gradient of error function for 2-layer network. wolffd@0:
wolffd@0: mlpderiv
wolffd@0: Evaluate derivatives of network outputs with respect to weights. wolffd@0:
wolffd@0: mlperr
wolffd@0: Evaluate error function for 2-layer network. wolffd@0:
wolffd@0: mlpevfwd
wolffd@0: Forward propagation with evidence for MLP wolffd@0:
wolffd@0: mlpfwd
wolffd@0: Forward propagation through 2-layer network. wolffd@0:
wolffd@0: mlpgrad
wolffd@0: Evaluate gradient of error function for 2-layer network. wolffd@0:
wolffd@0: mlphdotv
wolffd@0: Evaluate the product of the data Hessian with a vector. wolffd@0:
wolffd@0: mlphess
wolffd@0: Evaluate the Hessian matrix for a multi-layer perceptron network. wolffd@0:
wolffd@0: mlphint
wolffd@0: Plot Hinton diagram for 2-layer feed-forward network. wolffd@0:
wolffd@0: mlpinit
wolffd@0: Initialise the weights in a 2-layer feedforward network. wolffd@0:
wolffd@0: mlppak
wolffd@0: Combines weights and biases into one weights vector. wolffd@0:
wolffd@0: mlpprior
wolffd@0: Create Gaussian prior for mlp. wolffd@0:
wolffd@0: mlptrain
wolffd@0: Utility to train an MLP network for demtrain wolffd@0:
wolffd@0: mlpunpak
wolffd@0: Separates weights vector into weight and bias matrices. wolffd@0:
wolffd@0: netderiv
wolffd@0: Evaluate derivatives of network outputs by weights generically. wolffd@0:
wolffd@0: neterr
wolffd@0: Evaluate network error function for generic optimizers wolffd@0:
wolffd@0: netevfwd
wolffd@0: Generic forward propagation with evidence for network wolffd@0:
wolffd@0: netgrad
wolffd@0: Evaluate network error gradient for generic optimizers wolffd@0:
wolffd@0: nethess
wolffd@0: Evaluate network Hessian wolffd@0:
wolffd@0: netinit
wolffd@0: Initialise the weights in a network. wolffd@0:
wolffd@0: netopt
wolffd@0: Optimize the weights in a network model. wolffd@0:
wolffd@0: netpak
wolffd@0: Combines weights and biases into one weights vector. wolffd@0:
wolffd@0: netunpak
wolffd@0: Separates weights vector into weight and bias matrices. wolffd@0:
wolffd@0: olgd
wolffd@0: On-line gradient descent optimization. wolffd@0:
wolffd@0: pca
wolffd@0: Principal Components Analysis wolffd@0:
wolffd@0: plotmat
wolffd@0: Display a matrix. wolffd@0:
wolffd@0: ppca
wolffd@0: Probabilistic Principal Components Analysis wolffd@0:
wolffd@0: quasinew
wolffd@0: Quasi-Newton optimization. wolffd@0:
wolffd@0: rbf
wolffd@0: Creates an RBF network with specified architecture wolffd@0:
wolffd@0: rbfbkp
wolffd@0: Backpropagate gradient of error function for RBF network. wolffd@0:
wolffd@0: rbfderiv
wolffd@0: Evaluate derivatives of RBF network outputs with respect to weights. wolffd@0:
wolffd@0: rbferr
wolffd@0: Evaluate error function for RBF network. wolffd@0:
wolffd@0: rbfevfwd
wolffd@0: Forward propagation with evidence for RBF wolffd@0:
wolffd@0: rbffwd
wolffd@0: Forward propagation through RBF network with linear outputs. wolffd@0:
wolffd@0: rbfgrad
wolffd@0: Evaluate gradient of error function for RBF network. wolffd@0:
wolffd@0: rbfhess
wolffd@0: Evaluate the Hessian matrix for RBF network. wolffd@0:
wolffd@0: rbfjacob
wolffd@0: Evaluate derivatives of RBF network outputs with respect to inputs. wolffd@0:
wolffd@0: rbfpak
wolffd@0: Combines all the parameters in an RBF network into one weights vector. wolffd@0:
wolffd@0: rbfprior
wolffd@0: Create Gaussian prior and output layer mask for RBF. wolffd@0:
wolffd@0: rbfsetbf
wolffd@0: Set basis functions of RBF from data. wolffd@0:
wolffd@0: rbfsetfw
wolffd@0: Set basis function widths of RBF. wolffd@0:
wolffd@0: rbftrain
wolffd@0: Two stage training of RBF network. wolffd@0:
wolffd@0: rbfunpak
wolffd@0: Separates a vector of RBF weights into its components. wolffd@0:
wolffd@0: rosegrad
wolffd@0: Calculate gradient of Rosenbrock's function. wolffd@0:
wolffd@0: rosen
wolffd@0: Calculate Rosenbrock's function. wolffd@0:
wolffd@0: scg
wolffd@0: Scaled conjugate gradient optimization. wolffd@0:
wolffd@0: som
wolffd@0: Creates a Self-Organising Map. wolffd@0:
wolffd@0: somfwd
wolffd@0: Forward propagation through a Self-Organising Map. wolffd@0:
wolffd@0: sompak
wolffd@0: Combines node weights into one weights matrix. wolffd@0:
wolffd@0: somtrain
wolffd@0: Kohonen training algorithm for SOM. wolffd@0:
wolffd@0: somunpak
wolffd@0: Replaces node weights in SOM. wolffd@0:
wolffd@0: wolffd@0:
wolffd@0:

Copyright (c) Christopher M Bishop, Ian T Nabney (1996, 1997) wolffd@0: wolffd@0: