| REFERENCES (2009) | |||||||||||||||
|
|||||||||||||||
Abstract | |||||||||||||||
| HOD process. Therefore, there is a tradeoff between memory and computing time. HOD provides a direct solution for the learning algorithm. In comparison, tuning the MC algorithm to provide lower error rates by increasing the number of samples will result in less iterations to reach convergence, but that will come at the expense of a higher computation time. In the end, the time needed to attain smaller error values makes the decimation method a better choice. | |||||||||||||||
Publication details | |||||||||||||||
| |||||||||||||||