Statistical learning theory

About: Statistical learning theory is a research topic. Over the lifetime, 1618 publications have been published within this topic receiving 158033 citations.

Statistical Learning Theory for Spatial Data

Abstract: A current problem across many different fields is how to handle, to understand and to model data if there are too of the well-established approaches for working with spatially distributed data. It is a model-dependent approach based on the exploratory analysis and modelling of spatial correlation structure. On the other hand, recent explosive growth in the development of adaptive methods for learning from data have resulted in data-driven and model-free contemporary approaches, particularly based on Statistical Learning Theory —SLT (Vapnik-Chervonenkis theory) [1].

A Method of Asphalt Pavement Performance Evaluation Based on POS-LSSVM

Abstract: The support vector machine based on statistical learning theory is applied to establish a model for asphalt pavement evaluation.The particle swarm optimization,which can avoid the man-made blindness and enhance the efficiency and capability of forecasting,was used to optimize the parameters of the of least squares support vector machine(LS-SVM).PCI,RQI,BPN and SSI were selected as evaluation indexes to establish the model for pavement performance evaluation based on PSO-LSSVM.The results show that the method is feasible and effective for evaluation of asphalt pavement performance.

Application of LS-SVM on slurry pipeline critical deposition velocity prediction

Abstract: As an important parameter in slurry pipeline transportation, the critical deposition velocity has long been the subject of many experts and scholars at home and abroad. However, during slurry pipeline hydraulics design, there are some problems in slurry pipeline critical deposition formulas, such as the various forms of it, the big error of the calculated value and the narrow scope of application in it. Least squares support vector machine (LS-SVM) is a machine learning method based on statistical learning theory, which can avoid many shortcomings of the traditional neural network. Therefore, this paper introduced the least square support vector machine to predict the critical deposition velocity based on the analysis of the main factors that influence it. After the sample test, the simulation results show that this model can achieve good results and the accuracy of predicted results is higher when compared with the empirical formula of representation.

ARMA Criterion of Feature Extraction in Support Vector Regression:Empirical Research of Our National Social Total Retail Sales of Consumer Goods Prediction

Abstract: According to the method of local grey support vector regression in statistical learning theory,the ARMA criterion of feature extraction for single variable financial time series prediction is put forward.The experimental result of our national social total retail sales of consumer goods demonstrates that the ARMA criterion can objectively and accurately perform feature extraction and gain the higher accuracy of prediction.

An application of the learning theory to wavelet based signal denoising

Abstract: In this paper the statistical learning theory is applied to signal denoising using wavelets. The methodology is based on the estimation of the functional relationship between the Vapnik-Chervonenkis (VC) dimension and approximation complexity. Experimental results confirm the basic assumptions.