Dedicated to Roxanne Halitsky, Wife and Mother.
Theta Criteria are based on operator's theory and matrices decompositions. The main idea is to evaluate differences between sets of operators' eigenvalues & eigenvectors or eigenfunctions.
Theta Criteria will lead to more efficient analysis, optimization and prognosis of multivariate systems and applications. Theta Criteria can be used for data set mining and image processing. The book is designed as a complementary tool for applied mathematics methods.
We assume the Reader has standard undergraduate knowledge of advanced calculus, matrix theory, operator theory, approximation methods and measure theory.
The book consists of Introduction and four Chapters.
In Chapter I: Existing Statistical Methods for Multivariate Data Processing we briefly review existing methods and software packages for multivariate systems analysis. These methods are based upon Multivariate General Linear Hypothesis (MGLH).
In line with MGLH, all dataset variables are linear, additive and relationships models are linear series of weighted terms.
We also mention the following methods: Multiple Regression, Discriminant Function Analysis, Canonical Analysis, Principle Components Analysis and formal linear algebra methods. Lastly, we discuss our methods, Theta Criteria, which are constructed on norms of weighted differences of matrices ordered eigenvectors.
In Chapter II: Theta Criteria Formal Study, we placed our main formal results. At this moment we considered only positively-defined matrices.
In Chapter III: Theta Criteria Numerical Study we have numerically studied Theta Criteria on sets of varied matrices. We compare Theta Criteria accuracy with existing matrices' norms and invariants.
In Chapter IV: Theta Criteria Applications, we briefly discuss Theta Criteria application areas.
At this moment we are only able to "scratch the surface" of Theta Criteria universe. We are convinced that our methods will find their places in various complex applications.
In our Download Section you can grab Amazon Kindle eBook (Amazon Prime® may be required) or Free PDF. You can also purchase Softcover Book on Amazon.
Finally, you may contact us anytime.
Mathematics Subject Classification:
46G12 Measures and integration on abstract linear spaces
58C40 Spectral theory; eigenvalue problems
Mathematics / Probability & Statistics / Multivariate Analysis
© Steve Halitsky / Edward Halitsky. All Rights Reserved.
No part of this book or website may be translated or reproduced in any form without specific written permission from the Authors.
Cover Photo: Øresundsbroen bridge from Sweden to Denmark. Copyright © Roger Petersson.
MATLAB® is a registered trademark of MathWorks, Inc.
SAS® is a registered trademark of SAS Institute, Inc.
SPSS® is a registered trademark of SPSS, an IBM Company.