No BSD License  

Highlights from
Netlab

• demnlab(action); DEMNLAB A front-end Graphical User Interface to the demos
• demprgp(action); DEMPRGP Demonstrate sampling from a Gaussian Process prior.
• demprior(action); DEMPRIOR Demonstrate sampling from a multi-parameter Gaussian prior.
• demtrain(action); DEMTRAIN Demonstrate training of MLP network.
• [C,rate]=confmat(Y,T) CONFMAT Compute a confusion matrix.
• conjgrad(f, x, options, gradf... CONJGRAD Conjugate gradients optimization.
• consist(model, type, inputs, ... CONSIST Check that arguments are consistent.
• datread(filename) DATREAD Read data from an ascii file.
• datwrite(filename, x, t) DATWRITE Write data to ascii file.
• dem2ddat(ndata) DEM2DDAT Generates two dimensional data for demos.
• demgpot(x, mix) DEMGPOT Computes the gradient of the negative log likelihood for a mixture model.
• demhint(nin, nhidden, nout) DEMHINT Demonstration of Hinton diagram for 2-layer feed-forward network.
• demmet1(plot_wait) DEMMET1 Demonstrate Markov Chain Monte Carlo sampling on a Gaussian.
• demopt1(xinit) DEMOPT1 Demonstrate different optimisers on Rosenbrock's function.
• dempot(x, mix) DEMPOT Computes the negative log likelihood for a mixture model.
• dist2(x, c) DIST2 Calculates squared distance between two sets of points.
• eigdec(x, N) EIGDEC Sorted eigendecomposition
• errbayes(net, edata) ERRBAYES Evaluate Bayesian error function for network.
• evidence(net, x, t, num) EVIDENCE Re-estimate hyperparameters using evidence approximation.
• fevbayes(net, y, a, x, t, x_t... FEVBAYES Evaluate Bayesian regularisation for network forward propagation.
• fh=conffig(y, t) CONFFIG Display a confusion matrix.
• gauss(mu, covar, x) GAUSS Evaluate a Gaussian distribution.
• gbayes(net, gdata) GBAYES Evaluate gradient of Bayesian error function for network.
• glm(nin, nout, outfunc, prior... GLM Create a generalized linear model.
• glmderiv(net, x) GLMDERIV Evaluate derivatives of GLM outputs with respect to weights.
• glmerr(net, x, t) GLMERR Evaluate error function for generalized linear model.
• glmevfwd(net, x, t, x_test, i... GLMEVFWD Forward propagation with evidence for GLM
• glmfwd(net, x) GLMFWD Forward propagation through generalized linear model.
• glmgrad(net, x, t) GLMGRAD Evaluate gradient of error function for generalized linear model.
• glmhess(net, x, t, hdata) GLMHESS Evaluate the Hessian matrix for a generalised linear model.
• glminit(net, prior) GLMINIT Initialise the weights in a generalized linear model.
• glmpak(net) GLMPAK Combines weights and biases into one weights vector.
• glmtrain(net, options, x, t) GLMTRAIN Specialised training of generalized linear model
• glmunpak(net, w) GLMUNPAK Separates weights vector into weight and bias matrices.
• gmm(dim, ncentres, covar_type... GMM Creates a Gaussian mixture model with specified architecture.
• gmmactiv(mix, x) GMMACTIV Computes the activations of a Gaussian mixture model.
• gmmem(mix, x, options) GMMEM EM algorithm for Gaussian mixture model.
• gmminit(mix, x, options) GMMINIT Initialises Gaussian mixture model from data
• gmmpak(mix) GMMPAK Combines all the parameters in a Gaussian mixture model into one vector.
• gmmpost(mix, x) GMMPOST Computes the class posterior probabilities of a Gaussian mixture model.
• gmmprob(mix, x) GMMPROB Computes the data probability for a Gaussian mixture model.
• gmmsamp(mix, n) GMMSAMP Sample from a Gaussian mixture distribution.
• gmmunpak(mix, p) GMMUNPAK Separates a vector of Gaussian mixture model parameters into its components.
• gp(nin, covar_fn, prior) GP Create a Gaussian Process.
• gpcovar(net, x) GPCOVAR Calculate the covariance for a Gaussian Process.
• gpcovarf(net, x1, x2) GPCOVARF Calculate the covariance function for a Gaussian Process.
• gpcovarp(net, x1, x2) GPCOVARP Calculate the prior covariance for a Gaussian Process.
• gperr(net, x, t) GPERR Evaluate error function for Gaussian Process.
• gpfwd(net, x, cninv) GPFWD Forward propagation through Gaussian Process.
• gpgrad(net, x, t) GPGRAD Evaluate error gradient for Gaussian Process.
• gpinit(net, tr_in, tr_targets... GPINIT Initialise Gaussian Process model.
• gppak(net) GPPAK Combines GP hyperparameters into one vector.
• gpunpak(net, hp) GPUNPAK Separates hyperparameter vector into components.
• gradchek(w, func, grad, varar... GRADCHEK Checks a user-defined gradient function using finite differences.
• graddesc(f, x, options, gradf... GRADDESC Gradient descent optimization.
• gsamp(mu, covar, nsamp) GSAMP Sample from a Gaussian distribution.
• gtm(dim_latent, nlatent, dim_... GTM Create a Generative Topographic Map.
• gtmem(net, t, options) GTMEM EM algorithm for Generative Topographic Mapping.
• gtmfwd(net) GTMFWD Forward propagation through GTM.
• gtminit(net, options, data, s... GTMINIT Initialise the weights and latent sample in a GTM.
• gtmlmean(net, data) GTMLMEAN Mean responsibility for data in a GTM.
• gtmlmode(net, data) GTMLMODE Mode responsibility for data in a GTM.
• gtmmag(net, latent_data) GTMMAG Magnification factors for a GTM
• gtmpost(net, data) GTMPOST Latent space responsibility for data in a GTM.
• gtmprob(net, data) GTMPROB Probability for data under a GTM.
• hbayes(net, hdata) HBAYES Evaluate Hessian of Bayesian error function for network.
• hesschek(net, x, t) HESSCHEK Use central differences to confirm correct evaluation of Hessian matrix.
• hintmat(w); HINTMAT Evaluates the coordinates of the patches for a Hinton diagram.
• hinton(w); HINTON Plot Hinton diagram for a weight matrix.
• histp(x, xmin, xmax, nbins) HISTP Histogram estimate of 1-dimensional probability distribution.
• hmc(f, x, options, gradf, var... HMC Hybrid Monte Carlo sampling.
• kmeans(centres, data, options) KMEANS Trains a k means cluster model.
• knn(nin, nout, k, tr_in, tr_t... KNN Creates a K-nearest-neighbour classifier.
• knnfwd(net, x) KNNFWD Forward propagation through a K-nearest-neighbour classifier.
• linef(lambda, fn, x, d, varar... LINEF Calculate function value along a line.
• linemin(f, pt, dir, fpt, opti... LINEMIN One dimensional minimization.
• mdn(nin, nhidden, ncentres, d... MDN Creates a Mixture Density Network with specified architecture.
• mdn2gmm(mdnmixes) MDN2GMM Converts an MDN mixture data structure to array of GMMs.
• mdndist2(mixparams, t) MDNDIST2 Calculates squared distance between centres of Gaussian kernels and data
• mdnerr(net, x, t) MDNERR Evaluate error function for Mixture Density Network.
• mdnfwd(net, x) MDNFWD Forward propagation through Mixture Density Network.
• mdngrad(net, x, t) MDNGRAD Evaluate gradient of error function for Mixture Density Network.
• mdninit(net, prior, t, option... MDNINIT Initialise the weights in a Mixture Density Network.
• mdnpak(net) MDNPAK Combines weights and biases into one weights vector.
• mdnpost(mixparams, t) MDNPOST Computes the posterior probability for each MDN mixture component.
• mdnprob(mixparams, t) MDNPROB Computes the data probability likelihood for an MDN mixture structure.
• mdnunpak(net, w) MDNUNPAK Separates weights vector into weight and bias matrices.
• metrop(f, x, options, gradf, ... METROP Markov Chain Monte Carlo sampling with Metropolis algorithm.
• minbrack(f, a, b, fa, ... MINBRACK Bracket a minimum of a function of one variable.
• mlp(nin, nhidden, nout, outfu... MLP Create a 2-layer feedforward network.
• mlpbkp(net, x, z, deltas) MLPBKP Backpropagate gradient of error function for 2-layer network.
• mlpderiv(net, x) MLPDERIV Evaluate derivatives of network outputs with respect to weights.
• mlperr(net, x, t) MLPERR Evaluate error function for 2-layer network.
• mlpevfwd(net, x, t, x_test, i... MLPEVFWD Forward propagation with evidence for MLP
• mlpfwd(net, x) MLPFWD Forward propagation through 2-layer network.
• mlpgrad(net, x, t) MLPGRAD Evaluate gradient of error function for 2-layer network.
• mlphdotv(net, x, t, v) MLPHDOTV Evaluate the product of the data Hessian with a vector.
• mlphess(net, x, t, hdata) MLPHESS Evaluate the Hessian matrix for a multi-layer perceptron network.
• mlphint(net); MLPHINT Plot Hinton diagram for 2-layer feed-forward network.
• mlpinit(net, prior) MLPINIT Initialise the weights in a 2-layer feedforward network.
• mlppak(net) MLPPAK Combines weights and biases into one weights vector.
• mlpprior(nin, nhidden, nout, ... MLPPRIOR Create Gaussian prior for mlp.
• mlptrain(net, x, t, its); MLPTRAIN Utility to train an MLP network for demtrain
• mlpunpak(net, w) MLPUNPAK Separates weights vector into weight and bias matrices.
• netderiv(w, net, x) NETDERIV Evaluate derivatives of network outputs by weights generically.
• neterr(w, net, x, t) NETERR Evaluate network error function for generic optimizers
• netevfwd(w, net, x, t, x_test... NETEVFWD Generic forward propagation with evidence for network
• netgrad(w, net, x, t) NETGRAD Evaluate network error gradient for generic optimizers
• nethess(w, net, x, t, varargi... NETHESS Evaluate network Hessian
• netinit(net, prior) NETINIT Initialise the weights in a network.
• netopt(net, options, x, t, al... NETOPT Optimize the weights in a network model.
• netpak(net) NETPAK Combines weights and biases into one weights vector.
• netunpak(net, w) NETUNPAK Separates weights vector into weight and bias matrices.
• olgd(net, options, x, t) OLGD On-line gradient descent optimization.
• pca(data, N) PCA Principal Components Analysis
• plotmat(matrix, textcolour, g... PLOTMAT Display a matrix.
• ppca(x, ppca_dim) PPCA Probabilistic Principal Components Analysis
• quasinew(f, x, options, gradf... QUASINEW Quasi-Newton optimization.
• rbf(nin, nhidden, nout, rbfun... RBF Creates an RBF network with specified architecture
• rbfbkp(net, x, z, n2, deltas) RBFBKP Backpropagate gradient of error function for RBF network.
• rbfderiv(net, x) RBFDERIV Evaluate derivatives of RBF network outputs with respect to weights.
• rbferr(net, x, t) RBFERR Evaluate error function for RBF network.
• rbfevfwd(net, x, t, x_test, i... RBFEVFWD Forward propagation with evidence for RBF
• rbffwd(net, x) RBFFWD Forward propagation through RBF network with linear outputs.
• rbfgrad(net, x, t) RBFGRAD Evaluate gradient of error function for RBF network.
• rbfhess(net, x, t, hdata) RBFHESS Evaluate the Hessian matrix for RBF network.
• rbfjacob(net, x) RBFJACOB Evaluate derivatives of RBF network outputs with respect to inputs.
• rbfpak(net) RBFPAK Combines all the parameters in an RBF network into one weights vector.
• rbfprior(rbfunc, nin, nhidden... RBFPRIOR Create Gaussian prior and output layer mask for RBF.
• rbfsetbf(net, options, x) RBFSETBF Set basis functions of RBF from data.
• rbfsetfw(net, scale) RBFSETFW Set basis function widths of RBF.
• rbftrain(net, options, x, t) RBFTRAIN Two stage training of RBF network.
• rbfunpak(net, w) RBFUNPAK Separates a vector of RBF weights into its components.
• rosegrad(x) ROSEGRAD Calculate gradient of Rosenbrock's function.
• rosen(x) ROSEN Calculate Rosenbrock's function.
• scg(f, x, options, gradf, var... SCG Scaled conjugate gradient optimization.
• som(nin, map_size) SOM Creates a Self-Organising Map.
• somfwd(net, x) SOMFWD Forward propagation through a Self-Organising Map.
• sompak(net) SOMPAK Combines node weights into one weights matrix.
• somtrain(net, options, x) SOMTRAIN Kohonen training algorithm for SOM.
• somunpak(net, w) SOMUNPAK Replaces node weights in SOM.
• Contents.m
• demard.m
• demev1.m
• demev2.m
• demev3.m
• demgauss.m
• demglm1.m
• demglm2.m
• demgmm1.m
• demgmm2.m
• demgmm3.m
• demgmm4.m
• demgmm5.m
• demgp.m
• demgpard.m
• demgtm1.m
• demgtm2.m
• demhmc1.m
• demhmc2.m
• demhmc3.m
• demkmn1.m
• demknn1.m
• demmdn1.m
• demmlp1.m
• demmlp2.m
• demns1.m
• demolgd1.m
• demrbf1.m
• demsom1.m
• View all files