What solvers can be used for LSTM timeseries forecast.

Hi, I have a problem running the example openExample('nnet/TimeSeriesForecastingUsingDeepLearningExample').

First of all, this example wasn't found. After I copied the code and pasted in the editor and run it, it says that I used a invalid solver. Can anyone help me with this? What solver options can I use?

[James Lu]

This is the linke to the documentation page: https://www.mathworks.com/help/nnet/examples/time-series-forecasting-using-deep-learning.html

Answer 1

[mizuki]

In R2017b, you can use only 'sgdm' (Stochastic gradient descent with momentum optimizer). In R2018a, there are two additional optimizer options - 'rmsprop' and 'adam.'

See the "solverName" section under trainingoptions documentation: https://.mathworks.com/help/nnet/ref/trainingoptions.html