Home/ Questions/Q 8733805
In Process

The Archive Base Latest Questions

Editorial Team
  • 0
Asked: 2026-06-13T09:43:26+00:00

I am using the neurolab for Machine learning classfication problem, Link :- http://code.google.com/p/neurolab/ My

  • 0

I am using the neurolab for Machine learning classfication problem,
Link :- http://code.google.com/p/neurolab/

My question is, can we train the neural network incrementally ?

To explain further, I have three parts of input data, I want to train the Neuro Net by

e = net.train(input_part1, output_part1, show=1, epochs=100, goal=0.0001)
e = net.train(input_part2, output_part2, show=1, epochs=100, goal=0.0001)
e = net.train(input_part3, output_part3, show=1, epochs=100, goal=0.0001)

will the train call with first two parts will be effective in predicting the
neural net parameters -OR- Will this use only last training data ?

Share

Leave an answer

You must login to add an answer.

Need An Account,

1 Answer

  1. Editorial Team

    Usually you will add all data together into one dataset and train the net on this training set. Training refers to setting the weights. Why not add all inputs (part1, part2, part3) into one dataset? Note, that there are techniques where learning and re-adjusting is part of the learning algorithm. If you want to do a plain algorithm you have one cycle of training and one cycle of performance.

    • 0