BibSonomy publications for /concept/tag/featureFri Jun 20 12:09:32 CEST 2008Proc. Int'l Conf. Aspect-Oriented Software Development ({AOSD})Industry track paper at AOSD{Object-to-Aspect Refactorings for Feature Extraction}2004Me:ToRead AOP Thesis Feature SoftwareEvolution This report describes an experiment in using AspectJ to extract a feature from a Java code base in order to make it unpluggable. We describe issues and obstacles encountered while performing a series of code transformations and next present a collection of manual aspect-oriented refactorings, based on the experience gained in the process. These are described in detail and compounded with a self-contained example placing each refactoring in its proper context.Fri Jun 20 00:02:50 CEST 2008Random Features for Large-Scale Kernel Machines2007feature toread LargeScale kernel To accelerate the training of kernel machines, we propose to map the input data to a randomized low-dimensional feature space and then apply existing fast linear methods. The features are designed so that the inner products of the transformed data are approximately equal to those in the feature space of a user specified shift-invariant kernel. We explore two sets of random features, provide convergence bounds on their ability to approximate various radial basis kernels, and show that in large-scale classification and regression tasks linear machine learning algorithms applied to these features outperform state-of-the-art large-scale kernel machines.Thu Jun 19 17:46:40 CEST 2008Lake Como, Italy18-20 AprilLNCSGenetic Programming, Proceedings of Euro{GP}'200120382001Program Length model, Dynamic Parallel DNA, VHDL, Fitness, Function distributions, MAX Boolean Arm, Machine Layered Grammatical Symbolic Problem Active Evolvable Controller Developmental Computational Intrinsic control, Complexity, Contour individuals, distributed problem, Robotic representations, Process Reasoning, Feature Artificial Learning, Discovery, bias, Pattern Linear Neutral of Block-oriented genetic points, Recognition, function size, STGP, Hardware, Evolution, Polymorphism, Time Cellular Turing One-then-zeros structures, landscape, CAD, robot, modular Retina, design, Evolution Crossover Inverse VLSI Image Generator, Fixed Mapping, Genetic processing, machines, representation, Knowledge programming, Bloat, Systems, prediction, Problem, modelling, Character Genotype-Phenotype Multipopulation Series Extraction, BDD, Animat, detection, Regression, Trees, algorithms, robust, Evolvability, Causality, mutation, Multi-expression Iterated Kinematics, Thu Jun 19 17:46:40 CEST 2008Los Alamitos, CA, USAIEEE Transactions on Knowledge and Data EngineeringNovember111518--1528Evolutionary Constructive Induction172005programming, genetic construction, Feature classification algorithms, Feature construction in classification is a preprocessing step in which one or more new attributes are constructed from the original attribute set, the object being to construct features that are more predictive than the original feature set. Genetic programming allows the construction of nonlinear combinations of the original features. We present a comprehensive analysis of genetic programming (GP) used for feature construction, in which four different fitness functions are used by the GP and four different classification techniques are subsequently used to build the classifier. Comparisons are made of the error rates and the size and complexity of the resulting trees. We also compare the overall performance of GP in feature construction with that of GP used directly to evolve a decision tree classifier, with the former proving to be a more effective use of the evolutionary paradigm.Thu Jun 19 17:46:40 CEST 2008IEEE Transactions on Systems, Man and Cybernetics, Part BFebruary1106--117Genetic programming for simultaneous feature selection and classifier design362006classification, algorithm, feature pattern design algorithms, problem, intelligence), learning evolutionary online ranking c-class (artificial programming, selection extraction, classifier genetic classifier, Classification, scheme, This paper presents an online feature selection algorithm using genetic programming (GP). The proposed GP methodology simultaneously selects a good subset of features and constructs a classifier using the selected features. For a c-class problem, it provides a classifier having c trees. In this context, we introduce two new crossover operations to suit the feature selection process. As a byproduct, our algorithm produces a feature ranking scheme. We tested our method on several data sets having dimensions varying from 4 to 7129. We compared the performance of our method with results available in the literature and found that the proposed method produces consistently good results. To demonstrate the robustness of the scheme, we studied its effectiveness on data sets with known (synthetically added) redundant/bad features.Thu Jun 19 17:46:40 CEST 2008Properties and Applications of Dielectric Materials, 2003. Proceedings of the 7th International Conference on1-5 June258--261Genetic programming for partial discharge feature construction in large generator diagnosis12003programming, feature genetic extraction stator, algorithms, Thu Jun 19 17:46:40 CEST 2008LondonGECCO '07: Proceedings of the 9th annual conference on Genetic and evolutionary computation7-11 July1694--1701Improving the human readability of features constructed by genetic programming22007selection, knowledge post-processing programming, algorithms, discovery, construction, readability, factors, human parsimony, genetic feature The use of machine learning techniques to automatically analyse data for information is becoming increasingly widespread. In this paper we examine the use of Genetic Programming and a Genetic Algorithm to pre-process data before it is classified by an external classifier. Genetic Programming is combined with a Genetic Algorithm to construct and select new features from those available in the data, a potentially significant process for data mining since it gives consideration to hidden relationships between features. We then examine techniques to improve the human readability of these new features and extract more information about the domain.Thu Jun 19 17:46:40 CEST 2008Genetic Programming and Evolvable MachinesSeptemberPublished online: 17 August 20053265--281Genetic Programming with a Genetic Algorithm for Feature Construction and Selection62005classification, machine programming, genetic construction, selection, algorithms, learning feature The use of machine learning techniques to automatically analyse data for information is becoming increasingly widespread. In this paper we primarily examine the use of Genetic Programming and a Genetic Algorithm to pre-process data before it is classified using the C4.5 decision tree learning algorithm. Genetic Programming is used to construct new features from those available in the data, a potentially significant process for data mining since it gives consideration to hidden relationships between features. A Genetic Algorithm is used to determine which such features are the most predictive. Using ten well-known datasets we show that our approach, in comparison to C4.5 alone, provides marked improvement in a number of cases. We then examine its use with other well-known machine learning techniques.Thu Jun 19 17:46:40 CEST 2008IEEE Transactions on Systems, Man and Cybernetics, Part C: Applications and Reviews3287--300Fingerprint classification based on learned features35Augdiscovery, programming, Bayesian methods, extraction, visual primitive feature databases classifier, feature-learning algorithm, database, image identification, intelligence), algorithms, method, operations processing genetic operator composite classification NIST-4 classification, fingerprint Bayes (artificial learning In this paper, we present a fingerprint classification approach based on a novel feature-learning algorithm. Unlike current research for fingerprint classification that generally uses well defined meaningful features, our approach is based on Genetic Programming (GP), which learns to discover composite operators and features that are evolved from combinations of primitive image processing operations. Our experimental results show that our approach can find good composite operators to effectively extract useful features. Using a Bayesian classifier, without rejecting any fingerprints from the NIST-4 database, the correct rates for 4- and 5-class classification are 93.3percent and 91.6percent, respectively, which compare favourably with other published research and are one of the best results published to date.Thu Jun 19 17:46:40 CEST 2008Pattern RecognitionApril4505--512{GP}-based secondary classifiers382005programming, Classification, digit algorithms, Secondary selection, Handwritten classifiers Feature recognition, genetic Genetic programmingnext term (GP) is used to evolve secondary classifiers for disambiguating between pairs of handwritten digit images. The inherent property of feature selection accorded by GP is exploited to make sharper decision between conflicting classes. Classification can be done in several steps with an available feature set and a mixture of strategies. A two-step classification strategy is presented in this paper. After the first step of the classification using the full feature set, the high confidence recognition result will lead to an end of the recognition process. Otherwise a secondary classifier designed using a sub-set of the original feature set and the information available from the earlier classification step will help classify the input further. The feature selection mechanism employed by GP selects important features that provide maximum separability between classes under consideration. In this way, a sharper decision on fewer classes is obtained at the secondary classification stage. The full feature set is still available in both stages of classification to retain complete information. An intuitive motivation and detailed analysis using confusion matrices between digit classes is presented to describe how this strategy leads to improved recognition performance. In comparison with the existing methods, our method is aimed for increasing recognition accuracy and reliability. Results are reported for the BHA test-set and the NIST test-set of handwritten digits.Thu Jun 19 17:46:40 CEST 2008Seattle, Washington, USA{GECCO 2006:} Proceedings of the 8th annual conference on Genetic and evolutionary computation8-12 July887--894Synthesis of interest point detectors through genetic programming12006invariants, genetic synthesis, feature programming, algorithms, representation, theory program Thu Jun 19 17:46:40 CEST 2008GenomicsApril4471--479Applying genetic programming to the prediction of alternative m{RNA} splice variants892007exon, splicing, programming, Feature matrix, retention, genetic Alternative signals Cassette Splice algorithms, Intron Genetic programming (GP) can be used to classify a given gene sequence as either constitutively or alternatively spliced. We describe the principles of GP and apply it to a well-defined data set of alternatively spliced genes. A feature matrix of sequence properties, such as nucleotide composition or exon length, was passed to the GP system Discipulus To test its performance we concentrated on cassette exons (SCE) and retained introns (SIR). We analysed 27,519 constitutively spliced and 9641 cassette exons including their neighbouring introns; in addition we analysed 33316 constitutively spliced introns compared to 2712 retained introns. We find that the classifier yields highly accurate predictions on the SIR data with a sensitivity of 92.1percent and a specificity of 79.2percent. Prediction accuracies on the SCE data are lower, 47.3percent (sensitivity) and 70.9percent (specificity), indicating that alternative splicing of introns can be better captured by sequence properties than that of exons.Thu Jun 19 17:46:40 CEST 2008Mechanical Systems and Signal ProcessingMarch2271--289Fault detection using genetic programming192005Fault selection, Roller genetic bearing programming, Feature Condition algorithms, monitoring, detection, Genetic programming (GP) is a stochastic process for automatically generating computer programs. GP has been applied to a variety of problems which are too wide to reasonably enumerate. As far as the authors are aware, it has rarely been used in condition monitoring (CM). GP is used to detect faults in rotating machinery. Featuresets from two different machines are used to examine the performance of two-class normal/fault recognition. The results are compared with a few other methods for fault detection: Artificial neural networks (ANNs) have been used in this field for many years, while support vector machines (SVMs) also offer successful solutions. For ANNs and SVMs, genetic algorithms have been used to do feature selection, which is an inherent function of GP. In all cases, the GP demonstrates performance which equals or betters that of the previous best performing approaches on these data sets. The training times are also found to be considerably shorter than the other approaches, whilst the generated classification rules are easy to understand and independently validate.Thu Jun 19 17:46:40 CEST 2008Washington DC, USA{GECCO 2005}: Proceedings of the 2005 conference on Genetic and evolutionary computation25-29 June795--802Evolving optimal feature extraction using multi-objective genetic programming: a methodology and preliminary study on edge detection12005algorithms, edge Evolutionary Multiobjective detector, theory design, programming, extractor, multi-objective feature Optimisation, genetic Thu Jun 19 17:46:40 CEST 2008UKVIE 2006/001A Generic Optimal Feature Extraction Method using Multiobjective Genetic Programming2006Recognition Feature Pattern MOGP, algorithms, programming, genetic Optimisation, Extraction, Multiobjective In this paper, we present a generic, optimal feature extraction method using multiobjective genetic programming. We reexamine the feature extraction problem and argue that effective feature extraction can significantly enhance the performance of pattern recognition systems with simple classifiers. A framework is presented to evolve optimised feature extractors that transform an input pattern space into a decision space in which maximal class separability is obtained. We have applied this method to real world datasets from the UCI Machine Learning and StatLog databases to verify our approach and compare our proposed method with other reported results. We conclude that our algorithm is able to produce classifiers of superior (or equivalent) performance to the conventional classifiers examined, suggesting removal of the need to exhaustively evaluate a large family of conventional classifiers on any new problem.Thu Jun 19 17:35:00 CEST 2008Journal of Systems ArchitectureJuly7573--585Coevolving functions in genetic programming472001Hierarchical EDF, ADF, genetic programs, Feature selection/extraction, Speciation Knn, programming, algorithms, Classification, In this paper we introduce a new approach to the use of automatically defined functions (ADFs) within genetic programming. The technique consists of evolving a number of separate sub-populations of functions which can be used by a population of evolving main programs. We present and refine a set of mechanisms by which the number and constitution of the function sub-populations can be defined and compare their performance on two well-known classification tasks. A final version of the general approach, for use explicitly on classification tasks, is then presented. It is shown that in all cases the coevolutionary approach performs better than traditional genetic programming with and without ADFs.Thu Jun 19 17:35:00 CEST 2008Seattle, Washington, USA{GECCO 2006:} Proceedings of the 8th annual conference on Genetic and evolutionary computation8-12 July239--246Using genetic programming to classify node positive patients in bladder cancer12006Nodal Applications, feature Biological bladder synthesis classifier genetic learning rules, synthesis, measures, induction, selection, classification programming, algorithms, program design analysis, and pattern machine concept algorithms learning, evaluation, cancer, similarity staging, Thu Jun 19 17:35:00 CEST 2008Seattle, Washington, USA{GECCO 2006:} Proceedings of the 8th annual conference on Genetic and evolutionary computation8-12 July255--262Genetic programming for human oral bioavailability of drugs12006performance SMILES algorithms, bioinformatics, SVM, measures, feature complexity Applications, CFS, bioavailability, AIC, descriptors, LLSR, molecular selection, PCA, Biological genetic programming, ANN, Automatically assessing the value of bioavailability from the chemical structure of a molecule is a very important issue in biomedicine and pharmacology. In this paper, we present an empirical study of some well known Machine Learning techniques, including various versions of Genetic Programming, which have been trained to this aim using a dataset of molecules with known bioavailability. Genetic Programming has proven the most promising technique among the ones that have been considered both from the point of view of the accurateness of the solutions proposed, of the generalisation capabilities and of the correlation between predicted data and correct ones. Our work represents a first answer to the demand for quantitative bioavailability estimation methods proposed in literature, since the previous contributions focus on the classification of molecules into classes with similar bioavailability. Categories and Subject DescriptorsThu Jun 19 17:35:00 CEST 2008New YorkMonographs in Computer ScienceEvolutionary Synthesis of Pattern Recognition Systems2005genetic Image detection, Object visual algorithms, programming, synthesis, vision, processing, learning, Computer Pattern feature recognition Thu Jun 19 17:35:00 CEST 2008Lake Como, ItalyGenetic Programming, Proceedings of EuroGP'200118-20 April256--267LNCSFeature Extraction for the k-Nearest Neighbour Classifier with Genetic Programming20382001genetic Feature Extraction, programming, algorithms, Machine Learning In pattern recognition the curse of dimensionality can be handled either by reducing the number of features, e.g. with decision trees or by extraction of new features. We propose a genetic programming (GP) framework for automatic extraction of features with the express aim of dimension reduction and the additional aim of improving accuracy of the k-nearest neighbour (k-NN) classifier. We will show that our system is capable of reducing most datasets to one or two features while k-NN accuracy improves or stays the same. Such a small number of features has the great advantage of allowing visual inspection of the dataset in a two-dimensional plot. Since k-NN is a non-linear classification algorithm, we compare several linear fitness measures. We will show the a very simple one, the accuracy of the minimal distance to means (mdm) classifier outperforms all other fitness measures. We introduce a stopping criterion gleaned from numeric mathematics. New features are only added if the relative increase in training accuracy is more than a constant d, for the mdm classifier estimated to be 3.3%.