I have started programming bat algorithm as a feature selection .i have misunderstanding of velocity concept, since the length of the feature subset is become too long after its pass to velocity equation:
Vnew= Vold+(Current-Best) * f;
S= 1/ ( 1+Math.exp(-Vnew));
X(t+1) = { 1 S>Rnd , 0 Rnd>=S)
In the intial population , the solution subset (selected feature "1s") length between 20-50 but the subsequent iteration the solution subset become too long ( the selected feature become more than 600)....
I think this should give you an insight.
Wei-Neng Chen; Jun Zhang; Chung, H.S.H.; Wen-Liang Zhong; Wei-gang Wu; Yu-hui Shi, "A Novel Set-Based Particle Swarm Optimization Method for Discrete Optimization Problems," Evolutionary Computation, IEEE Transactions on , vol.14, no.2, pp.278,300, April 2010
A good ariticle
Nakamura, R. Y., Pereira, L. A., Costa, K. A., Rodrigues, D., Papa, J. P., & Yang, X. S. (2012, August). BBA: A binary bat algorithm for feature selection. In Graphics, Patterns and Images (SIBGRAPI), 2012 25th SIBGRAPI Conference on (pp. 291-297)..
http://arxiv.org/pdf/1308.3900.pdf
- Badges
- Science topic