| AdpBP90a | Offers the gradient-based optimization algorithm with individual adaptive learning rates by Silva and Almeida (Adaptive BackPropagation) |
| AdpBP90b | Offers the second version of the gradient descent based optimization algorithm with individual adaptive learning rates by Silva and Almeida (Adaptive BackPropagation). This optimization algorithm introduced by Silva and Almeida adds individual adaptive learning rates and weight-backtracking to standard steepest descent |
| AffineLinearFunction | The LinearFunction class represents a simple linear function  where the vector v and the offset b make up the parameters of the model |
| AffineLinearMap | The LinearMap class represents a simple affine linear model  where the matrix A and the vector b make up the parameters of the model |
| AllInOneMcSVM | Meta Model for SVM training |
| BalancedClassificationError | The ClassificationError class returns the number of classification errors, rescaled by the class magnitudes |
| BFGS | Offers methods to use the Broyden-Fletcher-Goldfarb-Shanno algorithm for the optimization of models |
| BinaryCriterion | Use thresholds for transforming output and target vectors into binary vectors and then calculates the classification error |
| C_SVM | Meta Model for SVM training |
| CachedMatrix | Efficient quadratic matrix cache |
| CachedMatrix::tCacheEntry | |
| CG | Offers methods to use the Conjugate Gradients algorithm for the optimization of models |
| Chessboard | Distribution of the chessboard classification problem |
| ClassificationError | Returns the number of classification errors |
| CMAOptimizer | The CMA-ES as a ReClaM Optimizer |
| CMAOptimizer::ModelFitness | |
| ComponentWiseModel | The ComponentWiseModel encapsulates the component wise application of a base model |
| CoverTree | A cover tree is a data structure forstering the fast computation of nearest neighbor queries |
| CoverTree::Node | |
| CoverTree::tNode | |
| CrammerSingerMcSVM | Meta Model for SVM training |
| CrossEntropy | Error measure for classication tasks that can be used as the objective function for training |
| CrossEntropyIndependent | Error measure for classification tasks of non exclusive attributes that can be used for model training |
| CVError | ErrorFunction based on a cross validation procedure |
| CVModel | Collection of sub-models for cross validation |
| DataFile | The DataFile class is a DataSource based upon a file |
| Dataset | The Dataset class encapsulates a realization of data from a DataSource |
| DataSource | Abstract description of a source of a dataset |
| DF_CrossEntropy | Error measure for classication tasks that can be used as the objective function for training |
| DF_CrossEntropyIndependent | Error measure for classification tasks of non exclusive attributes that can be used for model training |
| DF_MeanSquaredError | Mean Squared Error, using the GeneralDerivative interface |
| DiagGaussKernel | Guassian Kernel with independent scaling of every axis |
| EarlyStopping | Used for monitoring purposes, to avoid overfitting |
| Epsilon_SVM | Meta Model for SVM training |
| ErrorFunction | Base class of all error measures |
| ErrorPercentage | Calculates the error percentage based on the mean squared error |
| ExpNorm | Normalizes the output of a base model to sum to one using softmax |
| FFNet | Offers the functions to create and to work with a feed-forward network |
| GaussianProcess | Gaussian Process |
| GaussianProcessEvidence | Negative evidence of a Gaussian process |
| GaussianProcessVariance | Variance of a Gaussian process |
| GeneralGaussKernel | General Guassian Kernel |
| GridSearch | Optimize by trying out a grid of configurations |
| InputLabelMatrix | Kernel Gram matrix with input and label kernel |
| IntWiseDouble | |
| InverseClassSeparability | Inverse of the Class Separability Measure J by Huilin Xiong and M |
| IRpropMinus | This class offers methods for the usage of the improved Resilient-Backpropagation-algorithm without weight-backtracking |
| IRpropPlus | This class offers methods for the usage of the improved Resilient-Backpropagation-algorithm with weight-backtracking |
| JaakkolaHeuristic | Jaakkola's heuristic and related quantities for Gaussian kernel selection |
| KalmanFilter | The class provides an implementation of a standard linear KalmanFilter |
| KernelCoverTree | Cover tree with kernel distances |
| KernelFunction | Definition of a kernel function as a ReClaM model |
| KernelKMeans | Kernel k-means clustering |
| KernelMatrix | Kernel Gram matrix |
| KernelMeanClassifier | The kernel mean classifier is parameter free, that is, it does not require training |
| KernelNearestNeighbor | The kernel nearest neighbor classifier is parameter free, that is, it does not require training |
| LDA | Linear Discriminant Analysis (LDA) |
| LinearClassifier | Multi class classifier model suited for linear discriminant analysis |
| LinearEquation | Model and Error Function for the iterative approximate solution of a linear system |
| LinearFunction | Simple linear function  where the vector v makes up the parameters of the model |
| LinearKernel | Linear Kernel, parameter free |
| LinearMap | Simple linear model  where the matrix A makes up the parameters of the model |
| LinearRegression | Linear Regression |
| LinNorm | Normalizes the (non-negative) output of a base model by dividing by the overall sum |
| LinOutFFNet | Offers the functions to create and to work with (F)eed-(F)orward (Net)works with (Lin)ear (Out)put |
| LinOutMSEBFFNet | Offers the functions to create and to work with a a (F)eed-(F)orward (Net)works with (Lin)ear (Out)put and with explicit defined neuron-to-(B)ias connections. The network is combined with the (M)ean (S)quared (E)rror measure. This combination is created due to computational efficiency |
| LMSEFFNet | Offers the functions to create and to work with a feed-forward network with explicit defined neuron-to-bias connections. The network is combined with the mean squared error measure. This combination is created due to computational efficiency |
| LOO | Leave One Out (LOO) Error for Support Vector Machines |
| MeanSquaredError | Calculates the mean squared error |
| MetaSVM | Base class of all meta models for SVM training |
| Model | Base class of all models |
| ModelInterface | [DEPRECATED] Realizes the communication between the different modules |
| ModelWithErrorFunction | |
| MSEFFNet | Offers the functions to create and to work with a feed-forward network combined with the mean squared error measure. This combination is created due to computational efficiency |
| MSERBFNet | Offers the functions to create and to work with radial basis function networks and to train it with the mean squared error. This combination provides more computational efficiency compared to using the class MeanSquaredError |
| MSERNNet | A recurrent neural network regression model that learns with Back Propagation Through Time and assumes the MSE error functional |
| MultiClassSVM | Multi Class Support Vector Machine Model |
| MultiClassTestProblem | |
| NegativeBKTA | Balanced version of the NegativeKTA |
| NegativeKTA | Implementation of the negative Kernel Target Alignment (KTA) as proposed by Nello Cristianini |
| NegativeLogLikelihood | Negative logarithm of the likelihood of a probabilistic binary classification model |
| NegativePolarization | Implementation of the negative Kernel Polarization Measure, that is, Kernel Target Alignment without normalization |
| NestedGridSearch | Nested grid search |
| NetParams | Offers functions for easily reading information about a network, an error measure and an optimization algorithm from a configuration file |
| NoisyChessboard | Noisy version of the chessboard classification problem |
| NoisyInterval | Distribution of the noisy interval problem |
| NoisyRprop | Rprop-like algorithm for noisy function evaluations |
| NoisySvmLikelihood | Model selection objective for SVMs |
| NormalizedKernel | Normalized version of a kernel function |
| NormalizedRBFKernel | Gaussian rbf kernel, with density normalization |
| OCCMcSVM | Meta Model for SVM training |
| OneClassSVM | Meta Model for SVM training |
| Optimizer | Base class of all optimizers |
| OVAMcSVM | Meta Model for SVM training |
| Paraboloid | Convex quadratic model and error function |
| Partitioning | Defined a partitioning of a set of training points and labels as it is required for a cross validation procedure |
| PCA | "Principal Component Analysis" class for data compression |
| Perceptron | Perceptron online learning algorith |
| PointSearch | Optimize by trying out predefined configurations |
| PolynomialKernel | Polynomial Kernel |
| ProbenBNet | This special network is optimal for benchmark tests with the "proben 1" set |
| ProbenNet | This special network is optimal for benchmark tests with the "proben 1" set |
| QpBoxAllInOneDecomp | Quadratic program solver for box constrained multi class problems |
| QpBoxAllInOneDecomp::tExample | Data structure describing one training example |
| QpBoxAllInOneDecomp::tVariable | Data structure describing one variable of the problem |
| QpBoxAndEqCG | Quadratic program solver for box constrained problems using conjugate gradients |
| QpBoxAndEqDecomp | Very simple quadratic program solver for box constrained problems |
| QpBoxDecomp | Quadratic program solver for box constrained problems |
| QpCrammerSingerDecomp | Quadratic program solver for the multi class SVM as proposed by Crammer and Singer |
| QpCrammerSingerDecomp::tExample | Data structure describing one training example |
| QpCrammerSingerDecomp::tVariable | Data structure describing one variable of the problem |
| QPMatrix | Encapsulation of a quadratic matrix for quadratic programming |
| QPMatrix2 | SVM regression matrix |
| QPSolver | Abstract base class of all quadratic program solvers |
| QpSvmCG | Quadratic program solver for SVMs |
| QpSvmDecomp | Quadratic program solver for SVMs |
| Quickprop | This class offers methods for using the popular heuristic "Quickprop" optimization algorithm |
| QuickpropOriginal | This class offers methods for using the popular heuristic "Quickprop" optimization algorithm |
| RadiusMargin | Squared Radius-Margin-Quotient |
| RBFKernel | Definition of the RBF Gaussian kernel |
| RBFNet | Offers the functions to create and to work with radial basis function network |
| RegularizationNetwork | Meta Model for SVM training |
| RegularizedKernelMatrix | Kernel Gram matrix |
| ROC | ROC-Curve - false negatives over false positives |
| RpropMinus | This class offers methods for the usage of the Resilient-Backpropagation-algorithm without weight-backtracking |
| RpropPlus | This class offers methods for the usage of the Resilient-Backpropagation-algorithm with weight-backtracking |
| SigmoidFit | Optimize a sigmoid after SVM outputs to turn them into probability estimates |
| SigmoidModel | Standard sigmoid function with two parameters |
| SimpleSigmoidModel | Simple sigmoid function with one parameter |
| SpanBound | SpanBound for the 2-norm SVM |
| SpanBound1 | SpanBound for the 1-norm SVM |
| SparseDistribution | Sparse Distribution |
| SphereDistribution1 | Spherical Distribution |
| SquaredError | Calculates the sum-of-squares error |
| SteepestDescent | Standard steepest descent |
| StochasticGradientDescent | Learning strategies based on SteepestDescent, but with randomly chosen patterns |
| SVM | Support Vector Machine (SVM) as a ReClaM Model |
| SVM_Optimizer | Optimizer for SVM training by quadratic programming |
| SvmApproximation | Approximation of Support Vector Machines (SVMs) |
| SvmApproximationErrorFunction | Approximation of Support Vector Machines (SVMs) |
| SvmApproximationErrorFunctionGlobal | Approximation of Support Vector Machines (SVMs) |
| SvmApproximationModel | Approximation of Support Vector Machines (SVMs) |
| TanhNet | Offers a predefined feed-forward neural network, which uses the "tanh" activation function for hidden and output neurons |
| TransformedProblem | Linear transformation of data |
| ValidationError | Error on a hold out set |
| Variance | Cariance of data in each column of a data set |
| VarianceEstimator | Offers the methods to deal with active learning of a neural network |
| WeightedSumKernel | Weighted sum of kernel functions |
| WeightedSumKernel2 | Weighted sum of kernel functions |
| WTA | Winner-takes-all error function |
| ZeroOneLoss | The 0-1-loss counts the number of errors, where any deviation of the prediction from the target counts as one error, while only an exact match counts as a correct prediction |