B. Mulgrew and C. F. N. Cowan. Adaptive Filters and Equalisers. Kluwer Academic Publ., Boston, 1988. Home/Search Document Not in Database Summary Related Articles Check |
This paper is cited in the following contexts:
Exploiting Population Information in Evolutionary Learning - Yao, Liu, Darwen (Correct)
....can be found by inverting r oo and multiplying it by r od according to Eq. 5) However, this method is time consuming because whenever a new training example becomes available it requires inversion and multiplication of matrices. An alternative is to use the recursive least square (RLS) algorithm [23](pp.31 33) to determine the weights. From Eq. 6) and Eq. 7) we can get r oo (i) r oo (i Gamma 1) o(i)o T (i) 8) and r od (i) r od (i Gamma 1) d(i)o(i) 9) Using Eqs. 5) 8) and (9) we can get w(i) w(i Gamma 1) k(i)e(i) 10) where e(i) d(i) Gamma w T (i Gamma ....
.... Gamma 1) o(i)o T (i) 8) and r od (i) r od (i Gamma 1) d(i)o(i) 9) Using Eqs. 5) 8) and (9) we can get w(i) w(i Gamma 1) k(i)e(i) 10) where e(i) d(i) Gamma w T (i Gamma 1)o(i) 11) and k(i) r Gamma1 oo (i)o(i) 12) A recursion for r Gamma1 oo (i) is given by [23] r Gamma1 oo (i) r Gamma1 oo (i Gamma 1) Gamma r Gamma1 oo (i Gamma 1)o(i)o T (i)r Gamma1 oo (i Gamma 1) 1 o T (i)r Gamma1 oo (i Gamma 1)o(i) 13) In our implementation of the above RLS algorithm, three runs were always performed with different initial r Gamma1 oo ....
B. Mulgrew and C. F. N. Cowan. Adaptive Filters and Equalisers. Kluwer Academic Publ., Boston, 1988.
Ensemble Structure of Evolutionary Artificial Neural Networks - Yao, Liu (1996) (2 citations) (Correct)
....linear combination methods which assign different weights to different individuals because some individuals are fitter than others [4; 5] The second method simply calculates the weights based on the error rate of each individual. The third method uses the recursive least square (RLS) algorithm [6](pp.31 33) to determine the weights. All individuals in a population will be considered in these two methods. The fourth method is a linear combination over a subset of all the individuals in a population. The motivation behind using a subset is that some of the individuals in a population may ....
....can be found by inverting r oo and multiplying it by r od according to Eq. 5) However, this method is time consuming because whenever a new training example becomes available it requires inversion and multiplication of matrices. An alternative is to use the recursive least square (RLS) algorithm [6](pp.31 33) to determine the weights. From Eq. 6) and Eq. 7) we can get r oo (i) r oo (i Gamma 1) o(i)o T (i) 8) and r od (i) r od (i Gamma 1) d(i)o(i) 9) Using Eqs. 5) 8) and (9) we can get w(i) w(i Gamma 1) k(i)e(i) 10) where e(i) d(i) Gamma w T (i Gamma ....
[Article contains additional citation context not shown here]
B. Mulgrew and C. F. N. Cowan. Adaptive Filters and Equalisers. Kluwer Academic Publ., Boston, 1988.
How to Make Best Use of Evolutionary Learning - Yao, Liu, Darwen (1996) (1 citation) (Correct)
....fi = 0:1; 0:25; 0:5 and 0:75. The ensemble output was O = N X j=1 w j o j (2) where o 1 ; o 2 ; o N were outputs from different individuals. The third method also used linear combination of all the individuals in the last generation. It used the recursive least square (RLS) algorithm [18](pp.31 33) to determine the weights. The fourth method was slightly more complicated. It used a genetic algorithm (GA) to search for a near optimal subset of the population so that the final integrated system would contain fewer EANNs. The weights were still determined by the RLS algorithm. 3.3 ....
B. Mulgrew and C. F. N. Cowan. Adaptive Filters and Equalisers. Kluwer Academic Publ., Boston, 1988.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC