Recursive nonlinear estimation. A geometric approach (Q1921013): Difference between revisions

The author shows that statistical estimation procedures may be considered as distance minimizing. This is called a geometric view of estimation. A unifying approach is given involving maximum likelihood estimation, Kullback-Leibler distance, Shannon information measure and Bayesian estimation. A comprehensive discussion of the subject, many interconnections of related literature and a lot of evaluated numerical examples are presented. Although the mathematician may miss a more rigorous treatment of some subjects, the book may be of interest for both theoretically and practically oriented people.