Fundamentals of soft models in textiles
Fundamentals of soft models in textiles
复制标题
DOI：
10.1533/9780857090812.1.45
复制
发表时间：
2011
期刊：
影响因子：
--
通讯作者：
J. Militký
中科院分区：
文献类型：
--
作者：
J. Militký
Abstract: Methods for building empirical models may be broadly divided into three categories: linear statistical methods, neural networks and nonlinear multivariate statistical methods. This chapter demonstrates the basic principles of empirical model building and surveys the criteria for parameter estimation. The development of regression-type models, including techniques for exploratory data analysis and reducing dimensionality, is described. Typical empirical models for linear and nonlinear situations are discussed, along with evaluation of model quality based on degree of fit, prediction ability and other criteria. The main techniques for building empirical models are compared. The second part of the chapter describes some variants of neural networks. Radial basis function (RBF) networks are described in detail. The application of RBF networks in modeling univariate and multivariate regression problems is discussed. Finally, the application of neural networks in color difference formulae and drape prediction is presented.
