|
|
Focus:
|
|
|
Background: Neural networks have been sucessfully used in the last years for data mining applications such as regression and classification. However, their training and particularly the time spent in the training cannot cope with the need of responsiveness demanded nowadays. The parallel approach is one way to speedup the process of training a neural net. |
|
|
Goal: Our study has focused on ways of parallelising a neural net improving not only the training speed, but also the execution speed. |