Learners

 % Data splitting and re-ordering operations
 [X Y]             = reorderData(X,Y);                   % reorder (permute) data set
 [X Y]             = bootstrapData(X,Y,nBoot);           % resample (bootstrap) data set
 [Xtr Xte Ytr Yte] = crossValidate(X,Y,nFolds,iFold);    % split data for n-fold cross validation
 [Xtr Xte Ytr Yte] = splitData(X, Y, trainFraction);     % split data into training and test sets

 % Data pre-processing and feature operations
 [X scale]  = rescale(X [,scale])                        % scale data to unit variance
 [X mu sig] = whiten(X  [,mu,sig])                       % whiten (decorrelate) data
 [X feat] = fsubset(X, K [,feat])                        % extract (random) subset of features
 [X proj] = fproject(X, K [,proj])                       % (random) linear projection of features
 [X hash] = fhash(X, K [,hash])                          % create random hash projection of F into K features
 X = fpoly(X, degree)                                    % Xout = polynomial combinations of Xin
 X = fkitchenSink(X, K [,opt])                           % create random transform of F into dimension K
 [X T] = fsvd(X, K [,T])                                 % PCA-based (SVD) projection of X into K dimensions
 % [X L] = fLDA(X, Y, K [,L])                            % Fisher's LDA proj of X into K dimensions

 % Plotting
 plotClassify2D(learner, X,Y ,pre???);                   % plot data and classifier outputs on 2D data
 h = plotGauss2D( gMean, gCov, colorString, varargin)    % plot a Gaussian ellipse shape in 2D

 % Basic Learners
 knnRegress(xTrain, yTrain, K)                           % k-nearest-neighbors regression using given training data
 linearRegress(xTrain,yTrain,l2reg)                      % linear regression, with optional L2-regularization
 treeRegress(xTrain,yTrain [,options{:}])                % decision-tree least-squares regression
 %neuralNet???

 knnClassify(xTrain, yTrain, K)                          % k-nearest-neighbors classifier using given training data
 gaussBayesClassify(xTrain, yTrain, equalCovar)          % learn Gaussian class-conditional distributions
 logisticClassify(xTrain,yTrain [,...])                  % learn (mse loss) logistic regression classifier 
 %perceptron???

 %decisionTree???
 %treeClassify???
 %simpleStump
 %svmPrimal???  (or liblinear wrapper?)
 %svmDual???

 %Ensembling methods
 baggedClassif(baseLearner,N [,Xtr,Ytr,...])             % learn N bagged classifiers of type baseLearner
 gradBoostRegress(baseLearner, N, [,Xtr,Ytr,...])        % learn N gradient boosted regressors of type baseLearner
 adaboost(baseLearner [,Xtr,Ytr,N])                      % learn N boosted classifiers
 %ALSO: weighted combination?  (1) Add classifiers, choose weight by hold-out; (2) Add with MSE, Netflix method?