The Experts below are selected from a list of 1281909 Experts worldwide ranked by ideXlab platform

Geoffrey I. Webb - One of the best experts on this subject based on the ideXlab platform.

  • SDM - Scaling log-Linear Analysis to datasets with thousands of variables
    Proceedings of the 2015 SIAM International Conference on Data Mining, 2015
    Co-Authors: Francois Petitjean, Geoffrey I. Webb
    Abstract:

    Association discovery is a fundamental data mining task. The primary statistical approach to association discovery between variables is log-Linear Analysis. Classical approaches to log-Linear Analysis do not scale beyond about ten variables. We have recently shown that, if we ensure that the graph supporting the log-Linear model is chordal, log-Linear Analysis can be applied to datasets with hundreds of variables without sacrificing the statistical soundness [21]. However, further scalability remained limited, because state-of-the-art techniques have to examine every edge at every step of the search. This paper makes the following contributions: 1) we prove that only a very small subset of edges has to be considered at each step of the search; 2) we demonstrate how to efficiently find this subset of edges and 3) we show how to efficiently keep track of the best edges to be subsequently added to the initial model. Our experiments, carried out on real datasets with up to 2000 variables, show that our contributions make it possible to gain about 4 orders of magnitude, making log-Linear Analysis of datasets with thousands of variables possible in seconds instead of days.

  • ICDM - A Statistically Efficient and Scalable Method for Log-Linear Analysis of High-Dimensional Data
    2014 IEEE International Conference on Data Mining, 2014
    Co-Authors: Francois Petitjean, Lloyd Allison, Geoffrey I. Webb
    Abstract:

    Log-Linear Analysis is the primary statistical approach to discovering conditional dependencies between the variables of a dataset. A good log-Linear Analysis method requires both high precision and statistical efficiency. High precision means that the risk of false discoveries should be kept very low. Statistical efficiency means that the method should discover actual associations with as few samples as possible. Classical approaches to log-Linear Analysis make use of a#x03C7;2 tests to control this balance between quality and complexity. We present an information-theoretic approach to log-Linear Analysis. We show that our approach 1) requires significantly fewer samples to discover the true associations than statistical approaches -- statistical efficiency -- 2) controls for the risk of false discoveries as well as statistical approaches -- high precision - and 3) can perform the discovery on datasets with hundreds of variables on a standard desktop computer -- computational efficiency.

  • ICDM - Scaling Log-Linear Analysis to High-Dimensional Data
    2013 IEEE 13th International Conference on Data Mining, 2013
    Co-Authors: Francois Petitjean, Geoffrey I. Webb, Ann E. Nicholson
    Abstract:

    Association discovery is a fundamental data mining task. The primary statistical approach to association discovery between variables is log-Linear Analysis. Classical approaches to log-Linear Analysis do not scale beyond about ten variables. We develop an efficient approach to log-Linear Analysis that scales to hundreds of variables by melding the classical statistical machinery of log-Linear Analysis with advanced data mining techniques from association discovery and graphical modeling.

Francois Petitjean - One of the best experts on this subject based on the ideXlab platform.

  • SDM - Scaling log-Linear Analysis to datasets with thousands of variables
    Proceedings of the 2015 SIAM International Conference on Data Mining, 2015
    Co-Authors: Francois Petitjean, Geoffrey I. Webb
    Abstract:

    Association discovery is a fundamental data mining task. The primary statistical approach to association discovery between variables is log-Linear Analysis. Classical approaches to log-Linear Analysis do not scale beyond about ten variables. We have recently shown that, if we ensure that the graph supporting the log-Linear model is chordal, log-Linear Analysis can be applied to datasets with hundreds of variables without sacrificing the statistical soundness [21]. However, further scalability remained limited, because state-of-the-art techniques have to examine every edge at every step of the search. This paper makes the following contributions: 1) we prove that only a very small subset of edges has to be considered at each step of the search; 2) we demonstrate how to efficiently find this subset of edges and 3) we show how to efficiently keep track of the best edges to be subsequently added to the initial model. Our experiments, carried out on real datasets with up to 2000 variables, show that our contributions make it possible to gain about 4 orders of magnitude, making log-Linear Analysis of datasets with thousands of variables possible in seconds instead of days.

  • ICDM - A Statistically Efficient and Scalable Method for Log-Linear Analysis of High-Dimensional Data
    2014 IEEE International Conference on Data Mining, 2014
    Co-Authors: Francois Petitjean, Lloyd Allison, Geoffrey I. Webb
    Abstract:

    Log-Linear Analysis is the primary statistical approach to discovering conditional dependencies between the variables of a dataset. A good log-Linear Analysis method requires both high precision and statistical efficiency. High precision means that the risk of false discoveries should be kept very low. Statistical efficiency means that the method should discover actual associations with as few samples as possible. Classical approaches to log-Linear Analysis make use of a#x03C7;2 tests to control this balance between quality and complexity. We present an information-theoretic approach to log-Linear Analysis. We show that our approach 1) requires significantly fewer samples to discover the true associations than statistical approaches -- statistical efficiency -- 2) controls for the risk of false discoveries as well as statistical approaches -- high precision - and 3) can perform the discovery on datasets with hundreds of variables on a standard desktop computer -- computational efficiency.

  • ICDM - Scaling Log-Linear Analysis to High-Dimensional Data
    2013 IEEE 13th International Conference on Data Mining, 2013
    Co-Authors: Francois Petitjean, Geoffrey I. Webb, Ann E. Nicholson
    Abstract:

    Association discovery is a fundamental data mining task. The primary statistical approach to association discovery between variables is log-Linear Analysis. Classical approaches to log-Linear Analysis do not scale beyond about ten variables. We develop an efficient approach to log-Linear Analysis that scales to hundreds of variables by melding the classical statistical machinery of log-Linear Analysis with advanced data mining techniques from association discovery and graphical modeling.

Jan G. Rots - One of the best experts on this subject based on the ideXlab platform.

  • Shell elements for sequentially Linear Analysis: Lateral failure of masonry structures
    Engineering Structures, 2009
    Co-Authors: Matthew J. Dejong, Beatrice Belletti, Max A.n. Hendriks, Jan G. Rots
    Abstract:

    Shell elements are incorporated into the sequentially Linear Analysis method allowing the modeling of three-dimensional structures under non-proportional loading. The shell element implementation is first presented and then applied to simulate two previous experimental tests on full-scale unreinforced masonry structures under cyclic loading. Experimental cyclic envelopes and three-dimensional failure mechanisms are effectively predicted, providing further evidence to support sequentially Linear Analysis as an alternative to non-Linear Analysis in the finite element framework.

  • Sequentially Linear Analysis of fracture under non-proportional loading
    Engineering Fracture Mechanics, 2008
    Co-Authors: Matthew J. Dejong, Max A.n. Hendriks, Jan G. Rots
    Abstract:

    Sequentially Linear Analysis avoids convergence problems typical of non-Linear finite element Analysis by directly specifying a damage increment instead of a load or displacement increment. A series of Linear analyses is used to model highly non-Linear behavior without an iterative solution algorithm. The primary contribution herein is the extension of sequentially Linear Analysis to non-proportional loading, through an algorithm which selects the critical integration point to which a damage increment is applied. An orthotropic cracking model is presented which parallels the traditional fixed smeared crack concept. The non-proportional modeling framework is demonstrated through simulation of experimental results and settlement damage to a masonry facade.

J. Alwan - One of the best experts on this subject based on the ideXlab platform.

  • Non-Linear Analysis of RC beams using a SIFCON matrix
    Materials and Structures, 1993
    Co-Authors: Antoine E. Naaman, Hans W. Reinhardt, C. Fritz, J. Alwan
    Abstract:

    The results of a non-Linear Analysis model of reinforced concrete beams using a high-performance fibre-reinforced cement (FRC) based matrix are described and compared with experimental observations. The constitutive relationship of the FRC materials used in both compression and tension and the logical flow chart for the non-Linear Analysis model are described. Analytical predictions of moment versus curvature and moment versus deflection curves are compared with experimentally observed results of reinforced concrete beams using a SIFCON (slurry infiltrated fibre concrete) matrix. The use of a fibre-reinforced matrix in the compression zone of a reinforced concrete beam allows for other over-reinforced sections to achieve significant increases in structureal ductility while resisting loads close to their ultimate load. Ductility and energy ratios, normalized by the control beam without fibres, ranged from 2.47 to 3.60 and 3.28 to 5.69, respectively; that is several hundred per cent in improvement. The analytical model is shown to predict, with reasonably good agreement, experimental observations of moment-curvature and moment-deflection curves. Thus it can be used to evaluate the effects of various parameters in order to optimize structural performance. On décrit les résultats d’analyse non linéaire de poutres de béton armé utilisant une matrice de haute performance à base de composites de ciment-fibres (CCF), et on les compare aux résultats expérimentaux. On décrit les relations constitutives des matériaux CCF utilisés aussi bien en compression qu’en traction, ainsi que le diagramme d’écoulement logique pour le modèle d’analyse non linéaire. On compare les prédictions analytiques moment-courbure et moment-flèche aux résultats expérimentaux observés sur des poutres de béton armé utilisant une matrix SIFCON (slurry-infiltrated fibre concrete). L’utilisation d’une matrice renforcée de fibres dans des zones de compression d’une poutre en béton armé permet à des sections qui, sinon, seraient surchargées en armature, d’atteindre des améliorations appréciables de ductilité structurelle quand elles sont soumises à des charges proches de la charge de rupture. Les pourcentages de ductilité et d’énergie, normalisés par la poutre de contrôle dépourvue de fibres, se situaient respectivement entre 2,47 et 3,60, et 3,28 et 5,69, soit plusieurs centaines de pourcents d’amélioration. On constate que le modèle analytique prédit, avec une concordance assez satisfaisante, les observations expérimentales moment-courbure et moment-flèche. Il peut donc être utilisé pour évaluer les effets de différents paramètres afin d’optimiser la performance structurelle.

  • Non-Linear Analysis of RC beams using a SIFCON matrix
    Materials and Structures, 1993
    Co-Authors: Antoine E. Naaman, Hans W. Reinhardt, C. Fritz, J. Alwan
    Abstract:

    The results of a non-Linear Analysis model of reinforced concrete beams using a high-performance fibre-reinforced cement (FRC) based matrix are described and compared with experimental observations. The constitutive relationship of the FRC materials used in both compression and tension and the logical flow chart for the non-Linear Analysis model are described. Analytical predictions of moment versus curvature and moment versus deflection curves are compared with experimentally observed results of reinforced concrete beams using a SIFCON (slurry infiltrated fibre concrete) matrix. The use of a fibre-reinforced matrix in the compression zone of a reinforced concrete beam allows for other over-reinforced sections to achieve significant increases in structureal ductility while resisting loads close to their ultimate load. Ductility and energy ratios, normalized by the control beam without fibres, ranged from 2.47 to 3.60 and 3.28 to 5.69, respectively; that is several hundred per cent in improvement. The analytical model is shown to predict, with reasonably good agreement, experimental observations of moment-curvature and moment-deflection curves. Thus it can be used to evaluate the effects of various parameters in order to optimize structural performance.

R. Kari Thangarathanam - One of the best experts on this subject based on the ideXlab platform.

  • Geometric Non Linear Analysis of Composite Shells Under Thermomechanical Loading
    Volume 2: Computer Technology, 2005
    Co-Authors: B. Sutharson, R. Sarala, R. Kari Thangarathanam
    Abstract:

    A Geometric non-Linear Analysis of composite shells under thermomechanical loading has been discussed here. From the literature, it may be seen that the thermal stress Analysis of structural elements has continued to remain a research topic for a couple of decades. No one computationally verified the geometric non-Linear buckling of composite shells under thermomechanical loading using semiloof element. In this work, Linear buckling Analysis of Kari Thangaratnam (2) is extended to geometric non-Linear Analysis of composite shells under thermomechanical loading. A general shell element called the semiloof shell element has been extended to thermal stress Analysis of laminated shells. The formulation is based on nonLinear theory and the finite element method using semiloof element. The validation checks on the program are carried out using results on homogeneous isotropic shells available in the literature. The parameters considered in Analysis are (1) number of layers in the laminate, (2) Lay-up sequence (symmetry, antisymmetry, cross-ply etc.), (3) Fibre orientation angle, (4) Different aspect ratios, (5) Orthotrophy ratio, (6) Boundary conditions (simply supported, clamped and combination of boundary conditions).Copyright © 2005 by ASME

  • Geometric Non Linear Analysis of Composite Shells Under Thermomechanical Load
    Applied Mechanics, 2004
    Co-Authors: B. Sutharson, A. Elaya Perumal, R. Kari Thangarathanam
    Abstract:

    Geometric nonLinear Analysis of composite shells under thermomechanical load is reported here. From the literature, it may be seen that the thermal stress Analysis of structural elements has continued to remain a research topic for a couple of decades. No one computationally verified the geometric non-Linear buckling of composite shells under thermomechanical load using semiloof element. In this work, Linear buckling Analysis of Kari Thangaratnam (2) is extended to geometric non-Linear Analysis of composite shells under thermomechanical load. A general shell element called the semiloof shell element has been extended to thermal stress Analysis of laminated shells. The formulation is based on nonLinear theory and the finite element method using semiloof element. The validation checks on the program are carried out using results on homogeneous isotropic shells available in the literature. The parameters considered in Analysis are (1) number of layers in the laminate, (2) Lay-up sequence (symmetry, antisymmetry, cross-ply etc.), (3) Fibre orientation angle, (4) Different aspect ratios, (5) Orthotrophy ratio, (6) Boundary conditions (simply supported, clamped and combination of boundary conditions).Copyright © 2004 by ASME