Search

Support vector machines (SVMs) are very powerful tools for data classification and pattern recognition problems. They have been proven to have very good generalization performance in practice. In many real-life situations, the data to be trained are available on-line and batch training methods are not suitable. Here, we propose a new algorithm to efficiently train SVMs in such situations. The training problem is expressed as a general nonlinear optimization problem with special decomposition properties. The idea of incremental gradient technique is used and applied to interior-point methods and more precisely to a primal–dual technique. Computational results are given for various training data sets and the results are compared with those of the other state-of-the-art on-line training algorithms as well as of the batch training method. It is shown that the proposed algorithm achieves good results in terms of prediction accuracy as well as CPU time performance. [ABSTRACT FROM AUTHOR]

This paper applies learning models, such as support vector machines (SVM), neural networks, and mixed-integer programming kernel classifiers (MIPKC) to classify the flow pattern of a non-Newtonian fluid in an annulus/pipe. Classification of flow patterns is characterized by six attributes that represent the parameters that determine the fluid flow in the annulus/pipe. The SVM and MIPKC learning models construct a separating hyperplane in the feature space. The weights of the hyperplane represent a scaled level of importance for each of the parameters. Preliminary results show that the most efficient model with respect to computation time favours the SVM model with a polynomial kernel of degree 2. However, with respect to low error rates and sparseness of solution, one of the MIPKC models with a polynomial kernel of degree 2 outperforms the other methods. [ABSTRACT FROM AUTHOR]

Support vector machines (SVMs) have become one of the most powerful methods in machine learning for solving classification and regression problems. Finding the SVM solution can be regarded as estimating the centre of the largest hypersphere that can be inscribed in the set of consistent hypotheses called the version space. However, this solution can be inaccurate if the version space is asymmetric or elongated. Several approaches have been proposed to utilize other possible centres of the version space that can improve the generalization performance. Morreti in 2003 proposed an algorithm for finding the centre of a general polytope, the so called p-Centre, using weighted projections. By applying this method, Bruckner in 2001 introduced a formulation for solving binary classification problems based on an approximation of the p-Centre of the version space, the so called p-Centre machine. In this paper, we extend the work by Bruckner and propose a kernel-based algorithm for regression analysis using the p-Centre method. The concept of the p-Centre of a polytope and version space is also explained. Furthermore, the applications of the proposed method are presented. The preliminary results indicate that the p-Centre-based kernel machine for regression has promising performance compared with the SVM for regression. [ABSTRACT FROM AUTHOR]

We reformulate the support vector machine approach to classification and regression problems using a different methodology than the classical 'largest margin' paradigm. From this, we are able to derive extremely simple quadratic programming problems that allow for general symbolic solutions to the classical problems of geometric classification and regression. We obtain a new class of learning machines that are also robust to the presence of small perturbations and/or corrupted or missing data in the training sets (provided that information about the amplitude of the perturbations is known approximately). A high performance framework for very large-scale classification and regression problems based on a Voronoi tessellation of the input space is also introduced in this work. Our approach has been tested on seven benchmark databases with noticeable gain in computational time in comparison with standard decomposition techniques such as SVMlight. [ABSTRACT FROM AUTHOR]

Technical indicators are used with two heuristic models, kernel principal component analysis and factor analysis in order to identify the most influential inputs for a forecasting model. Multilayer perceptron (MLP) networks and support vector regression (SVR) are used with different inputs. We assume that the future value of a stock price/return depends on the financial indicators although there is no parametric model to explain this relationship, which comes from the technical analysis. Comparison studies show that SVR and MLP networks require different inputs. Furthermore, proposed heuristic models produce better results than the studied data mining methods. In addition to this, we can say that there is no difference between MLP networks and SVR techniques when we compare their mean square error values. [ABSTRACT FROM AUTHOR]