||Topic: Combine two methods: batch & incremental
||Posted: 26-May-11 03:57
I have a dataset which contains some snapshots of my results. each snapshot contains some data, and different snapshots belong to different time stamps.
Now, I want to train a neural network. I want to train it in this way:
-to give each snapshot's data in batch form to my NN.
-to give different snapshots in incremental form to my NN.
But, I don't know hove to design a NN to combine both batch and incremental learning in it.
I appreciate any help.