www.pudn.com > Classification_toolbox.part01.rar > contents.m
% Classification GUI and toolbox % Version 1.0 % % GUI start commands % % classifier - Start the classification GUI % enter_distributions - Starts the parameter input screen (used by classifier) % multialgorithms - Start the algorithm comparison screen % show_algorithms - Show possible algorithms for classification, clustering and feature selection % % Preprocessing methods % % ADDC - Compute k clusters for the data using the agglomerative clustering method % AGHC - Compute k clusters for the data using the agglomerative hierarchical clustering method % BIMSEC - Compute k clusters for the data using the basic iterative MSE clustering method % Competitive_learning - Compute k clusters for the data using a competitive NN % Deterministic_annealing - Compute k patterns which typify the data using the Deterministic SA algorithm % Deterministic_SA - Compute k patterns which typify the data using the Deterministic SA algorithm (Another implementation) % DSLVQ - Distinction sensitive linear vector quantization % Fuzzy_k_means - Compute k means for the data using the fuzzy_k_means algorithm % FishersLinearDiscriminant - Fisher linear discriminant % k_means - Compute k means for the data using the k_means algorithm % Kohonen_SOFM - Reduce data points using a Kohonen self-orgenizing feature map % Leader_Follower - Compute clusters for the data using the basic leader-follower clustering method % LVQ1 - Linear vector quantization with one neighbor % LVQ3 - Linear vector quantization 3 algorithm % min_spanning_tree - Reduce data points using a minimum spanning tree % PCA - Principal component analysis % SOHC - Compute k clusters for the data using the stepwise optimal hierarchical clustering method % Stochastic_SA - Compute k patterns which typify the data using the Stochastic SA algorithm % % Parametric classification algorithms % % Balanced_Winnow - Balanced Winnow algorithm % Bayesian_Model_Comparison - Find a Gaussian model using Bayesian model comparison % EM - Expectation maximization algorithm % Gibbs - The Gibbs algorithm % Ho_Kashyap - The regular and modified Ho-Kashyap algorithm % LMS - Least-means square algorithm % LS - Least squares algorithm % Marginalization - Classify when a feature is missing using the marginal distribution % ML - Maximum likelihood algorithm % ML_diag - Maximum likelihood with diagonal covariance matrices % ML_II - Find a Gaussian model using maximum likelihood model comparison % NDDF - Normal density discriminant function % None - A dummy file % Perceptron - Single perceptron algorithm % Perceptron_Batch - Batch perceptron algorithm % Perceptron_BVI - Batch variable increment perceptron algorithm % Perceptron_FM - Perceptron which improves according to the example farthest from the margin % Perceptron_VIM - Variable increment perceptron with margin algorithm % Pocket - Pocket algorithm % RDA - Regularized descriminant analysis (Friedman shrinkage algorithm) % Relaxation_BM - Batch relaxation with margin % Relaxation_SSM - Single-sample relaxation with margin % Stumps - Simple stump classifier % % Non-parametric classification algorithms % % Ada_Boost - Ada Boost algorithm % Backpropagation_Batch - Neural network trained with the batch backpropagation algorithm % Backpropagation_CGD - Neural network trained with the batch backpropagation algorithm and conjugate gradient descent % Backpropagation_Quickprop - Neural network trained with the quickprop backpropagation algorithm % Backpropagation_Recurrent - A recurrent neural network trained with the batch backpropagation algorithm % Backpropagation_SM - A recurrent neural network trained with the stochastic backpropagation algorithm with momentum % Backpropagation_Stochastic - Neural network trained with the stochastic backpropagation algorithm % C4_5 - The C4.5 algorithm % CART - Classification and regression trees % CARTfunctions - Used by CART % Cascade_Correlation - Cascade-correlation type neural network % Components_with_DF - Component classifiers with descriminant functions % Components_without_DF - Component classifiers without descriminant functions % Deterministic_Boltzmann - Deterministic Boltzmann learning % Discrete_Bayes - Bayes classifier for discrete patterns % Genetic_algorithm - Basic genetic algorithm % Genetic_programming - Genetic programming of a solution % ID3 - Quinlan's ID3 classification tree algorithm % Interactive_Learning - Interactive learning (Learning with queries) % Local_Polynomial - Local polynomial fitting % loglikelihood - Used by Local polynomial fitting % LocBoost - Local boosting % LocBoostFunctions - Used by LocBoost % Minimum_Cost - Classify under a minimum cost strategy with histogram equalization % Multivariate_Splines - Multivariate adaptive regression splines % Nearest_Neighbor - Nearest neighbor algorithm % NearestNeighborEditing - Nearest neighbor editing algorithm % Optimal_Brain_Surgeon - Train a backprop. Neural net. and prune it using the optimal brain surgeon algorithm % Parzen - Parzen window algorithm % Perceptron_Voted - Voted perceptron algorithm. % PNN - Probabilistic neural network % Projection_pursuit - Projection pursuit regression for classification % RCE - Reduced coulomb energy algorithm % RBF_Network - Train a radial-basis function neural network % Store_Grabbag - An improvement on the nearest neighbor algorithm % SVM - Support vector machines % % Feature selection % % combinations - Return all combinations of indices (Used by sequential and exhaustive feature selection) % Exhaustive_Feature_Selection- Exhaustive feature selection % Genetic_Culling - A Culling type genetic algorithm for feature selection % HDR - Hierarchical dimensionality reduction % ICA - Independent component analysis % infomat - Generates the mutual information matrix. Used by Information_based_selection % Information_based_selection - Choose the most relevant features using the Koller-Sawami algorithm % MDS - Multidimensional scaling % MultipleDiscriminantAnalysis- Multiple descriminant analysis % NLPCA - Non-linear PCA % PCA - Principle component analysis % Sequential_Feature_Selection- Sequential (Forward/Backward) feature selection % % Error estimation % % calculate_error - Calculates the classification error given a decision surface % classification_error - Used by claculate_error % classify_paramteric - Builds a decision region for multi-Gaussian distributions % % Error bounds % % Bhattacharyya % Chernoff % Discriminability % % GUI housekeeping functions % % calculate_region - Finds the data scatter region % classifier_commands - Classifier screen commands % click_points - Graphically enter a distribution % enter_distribution_commands - Used by enter_distributions % feature_selection - The feature selection GUI open when data with more than 2D is loaded % feature_selection_commands - The commands file for the feature selection GUI % FindParameters - A GUI for finding the optimal parameters for a classifier % FindParametersFunctions - The commands file for FindParameters % GaussianParameters - Opens a GUI for displaying the gaussian parameters of a distribution % generate_data_set - Generate a data set given Gaussian parameters % high_histogram - Generate a histogram for high-dimensional data % load_file - Load data files % make_a_draw - Randomly find indices from a data set % multialgorithms_commands - Multialgorithms screen comands % plot_process - Plot partition centers during the algorithm execution % plot_scatter - Make a scatter plot of a data set % Predict_performance - Predict performance of algorithms from their learning curves % process_params - Read a parameter vector and return it's components % read_algorithms - Reads an algorithm file into a data structure % start_classify - Main function used by classifier % voronoi_regions - Plot Voronoi regions % % Data sets (Ending _data means that the file contains patterns, % _params means that the file contains the distribution parameters) % % chess - The parameters for a 4x4 chess board distribution % clouds - A data set composed of four Gaussians % seperable - A linearly seperable data set % spiral - Two interlocking spirals data set % XOR - XOR distribution % % %____________________________________________________________________________________ % Elad Yom-Tov (elad@ieee.org) and David Stork % Technion - Israel Institute of Technology % Haifa, Israel