New objects/additions V 1.5

We have added small (usually one or two line) code examples to almost every algorithm's Matlab help page. We have also added references for algorithms, complete with web links. Hopefully it is a little easier to see how to use an algorithm now. We also added new sections to the tutorial: on multi-class classification, feature selection and working with unbalanced data to name a few.

We also have the following new objects implemented:

  • mksvm - multi kernel Linear Programming-SVM

  • mrank - manifold ranking

  • kvq - kernel vector quantization

  • bayessel - Bayesian model selection for SVM/SVR

  • toy2d - Simple two-dimensional toy problem generator

  • toyreg - - Linear Regression with o outputs and n inputs

  • spiral - Simple two-dimensional spiral problem generator

  • ppca - Probabilistic PCA

  • adaboost - Adaboost

  • lgzc - Local and Global Consistent Learner

  • rank_perceptron - general input / output mapping with joint kernels