[ Main Page ]
# Demonstrations T-61.5040

Below are some of the demos shown in class. When the
demos are not written by me, the link points to
the author's page. The demos are very short illustrations
shown in the lectures. They are not intended as comprehensive
examples of the corresponding topics. They are offered as-is,
there are no (or few) comments, and most demos require
explanation that was given in class :-)

All demos are running in R, the statistical computing package available from R Project Webpage.

All demos are running in R, the statistical computing package available from R Project Webpage.

- Projection of high dimensional data
- Overfitting
- Learning with local assumptions
- Linear Regression with Nonlinear Basis Functions
- Linear classifier in feature space is nonlinear in data space
- Posterior of average error c, given fixed training data
- Learning Kernel Classifiers Toolbox
- Inferring Mean of a Normal
- Simple Normal Classifier
- T Distribution vs. Normal Distribution
- Robustness of Random Variance
- Hierarchical Mixture of Nonlinear and Linear Models (requires tgb library for R, installation instructions available here)
- Normalization is nontrivial
- Metropolis sampling of a Normal distribution
- Metropolis on a bimodal distribution
- Gibbs sampling of a Normal distribution
- Laplace approximation
- Variational approximation
- Free-form variational approximation
- Iterating the Means of Two-Normal Mixture
- Gibbs Sampler for Two-Normal Mixture
- EM algorithm for Normal mixtures
- Variational Bayes for Normal mixtures
- Missing Normal data, Gibbs Sampler vs. heuristics (needs an R package "norm" available from CRAN, under Contributed extension packages)
- Gaussian Process regression
- GP regression with covariance parameter estimation
- Mean-Field algorithm for GP classification (Opper and Winther). The demo requires the "time" package, available from CRAN, under Contributed extension packages, and the example dataset from Learning Kernel Classifiers Software Resources (search for "Example dataset")

You are at: CIS → /Opinnot/T-61.5040/Demos/demos.shtml

Page maintained by webmaster at cis.hut.fi, last updated Wednesday, 18-Apr-2007 14:08:30 EEST