Machine Learning
Generalization performance of support vector machines and other pattern classifiers
Advances in kernel methods
Making large-scale support vector machine learning practical
Advances in kernel methods
Support vector machines for dynamic reconstruction of a chaotic system
Advances in kernel methods
Learning with Kernels: Support Vector Machines, Regularization, Optimization, and Beyond
Learning with Kernels: Support Vector Machines, Regularization, Optimization, and Beyond
IEEE Transactions on Knowledge and Data Engineering
IEEE Transactions on Knowledge and Data Engineering
Training ν-Support Vector Classifiers: Theory and Algorithms
Neural Computation
Predictability, Complexity, and Learning
Neural Computation
Neural Computation
Fuzzy lattice neural network (FLNN): a hybrid model for learning
IEEE Transactions on Neural Networks
Input space versus feature space in kernel-based methods
IEEE Transactions on Neural Networks
Hi-index | 0.00 |
The paper presents a new effective approach for the construction of local Support Vector Machine (SVM) regression models for the prediction of non-stationary data. We illustrate that an analysis in the framework of dynamical systems theory can provide critically useful parameters for the effective training of the SVM predictors. A correlation dimension parameter is approximated and is used in order to obtain an appropriate dimensionality for the input space of the predictive SVM. The presented prediction framework can be utilized both for continuous signals and for the case where the observable variable is a discrete symbol, a circumstance very common in data mining problems. Using the information extracted from the correlation dimension computation, local Support Vector Machine models are trained and they are used only for local predictions. We apply this methodology to the difficult problem of evaluating the predictability of DNA sequences. The results support the importance of the estimation of the proper dimensionality of the embedding space by means of the correlation dimension. Additionally, they demonstrate the effectiveness of the presented SVM based prediction approach that is formulated under a dynamical systems reconstruction framework.