Kontakt
Kako nakupovati
Dostava
Svetovalec pri nakupu
Regularized Least Squares Multi-Class Knowledge- Based Kernel Machines
Angleščina
142 b
30 dni za vračilo blaga
This book presents how a two-class discrimination §model with or without prior knowledge can be §extended to the case of multi-categorical §discrimination with or without prior knowledge. The §prior knowledge of interest is in the form of §multiple polyhedral sets belonging to one or more §categories, classes, or labels, and it is introduced §as additional constraints into a classification §model formulation. The solution of the knowledge-§based support vector machine (KBSVM) model for two-§class discrimination is characterized by a linear §programming (LP) problem, and this is due to the §specific norm (L1 or L ) that is used to compute the §distance between the two classes. The proposed §solutions to a classification problem is expressed §as a single unconstrained optimization problem with §(or without) prior knowledge via a regularized least §squares cost function in order to obtain a linear §system of equations in input space and/or dual space §induced by a kernel function that can be solved §using matrix methods or iterative methods.
O knjigi
Angleščina
Kategorije
Podarite to knjigo še danes
To je povsem preprosto
1 Dodajte knjigo v košarico in izberite dostavo kot darilo 2 V zameno vam bomo poslali kupon 3 Knjiga bo dostavljena na naslov obdarovancaPozdravljeni! Sem Libroamiko, vaš knjižni svetovalec.
Kako vam lahko pomagam?