As with the bestselling first edition, Computational Statistics Handbook with MATLABAr, Second Edition covers some of the most commonly used contemporary techniques in computational statistics. With a strong, practical focus on implementing the methods, the authors include algorithmic descriptions of the procedures as well as examples that illustrate the use of the algorithms in data analysis. Updated for MATLABAr R2007a and the Statistics Toolbox, Version 6.0, this edition incorporates many additional computational statistics topics. New to the Second Edition ac New functions for multivariate normal and multivariate t distributions ac Updated information on the new MATLAB functionality for univariate and bivariate histograms, glyphs, and parallel coordinate plots ac New content on independent component analysis, nonlinear dimensionality reduction, and multidimensional scaling ac New topics on linear classifiers, quadratic classifiers, and voting methods, such as bagging, boosting, and random forests ac More methods for unsupervised learning, including model-based clustering and techniques for assessing the results of clustering ac A new chapter on parametric models that covers spline regression models, logistic regression, and generalized linear models ac Expanded information on smoothers, such as bin smoothing, running mean and line smoothers, and smoothing splines With numerous problems and suggestions for further reading, this accessible text facilitates an understanding of computational statistics concepts and how they are employed in data analysis.We learned in a previous chapter that the EM (Expectation-Maximization) algorithm can be used to estimate a finite mixture, and we saw that several issues must be addressed. First, we need to specify the number of component densities inanbsp;...
|Title||:||Computational Statistics Handbook with MATLAB, Second Edition|
|Author||:||Wendy L. Martinez, Angel R. Martinez|
|Publisher||:||CRC Press - 2007-12-20|