I have physical model prediction data as well as actual data. From this I can calculate the error of each prediction data point through simple subtraction. I am hoping to train a neural network to be able to assign an error to the input of the physical model.
My current plan is to normalize the error of each data point and assign it as a label to each model input. So the NN would be trained (and validated)on a 1000 data points with the associated error as a label. Once the model is trained I would be able to input one data point and the output of the neural network would be a single class, that is the error. The purpose this would serve would be to tune the physical prediction model. Would this kind of architecture work? If so, would you recommend a feedforward or RNN? Thank you.
Dear Arthur Shune i recommend simple feed-forward for this simple task,if you have enough samples to teach the network the relation which is kind of y = x2-x1; then the network will do well on unseen data,but if your sample size is very short then your model will not do well and maybe just maybe you will have to use RNN,i have previously worked on similar task as neural network basic,you may want to see my simple work from here :https://www.kaggle.com/mobassir/hello-world-of-neural-network which demonstrated something very close to your problem,you can apply same strategy for your problem but make sure you have large sample to train the Network,thanks
Syed Mobassir Hossen That's exactly what I was looking for. Thank you for your help and advice. Cheers!
- Badges
- Science topic