I want to train a neural network with input of some array and output such as a time series ( 2D matrix):
input = [ a1, a2, ..., an] ====> n = 1,2,...,10
output = [ t1 , t2 , ... , tm ;
f(t1), f(t2), ... , f(tm)] m => 1,2, ..., j j is variable between 500 to 1000]
where n is small (less than 10), and m is varied for each input, It is also large such as time series, f in above output matrix is a function with high nonlinearity.
In other word the output is the solution of ODE with variable step size and input is the parameters of ODE equation.
I want to create an accurate neural network to predict a good output graph by giving a new input parameters. Which types of neural network you suggest.?
Please see the attached figure.
Keyword: Artificial Neural Networks
Deep Learning
Multilayer Perceptron
Convolutional Neural Network
LSTM
NN
CNN
RNN and CNN both are good for time serise.
But RNNs are better suited to analyzing temporal, sequential data, such as text or videos. A CNN has a different architecture from an RNN. CNNs are "feed-forward neural networks" that use filters and pooling layers, whereas RNNs feed results back into the network.
So you can apply RNN...
Dear Mohsen Rezaei ,
Types of Neural NetworksPerceptron Feed Forward Neural Network Multilayer Perceptron Convolutional Neural Network Radial Basis Function Neural Network Recurrent Neural Network LSTM –Long Short-Term Memory Sequence to Sequence models Modular Neural Network FAQs
https://www.mygreatlearning.com/blog/types-of-neural-networks/
One of best practices in case of outputs of various values (e.g. values are way greater or less than one) is to normalize the labels in the training phase.
You can use simple ANN, autoencoder CNN , or encoder followed with ANN in your case.
- Badges
- Science topic
- Similar topics
- Methodology
- Research Methodology
- Research Papers