training network plot accuracy intead of rmse

hello everyone

I am using LSTM for data prediction and I use trainNetwork for it but When I run my cde the training plot only plots rmse and I want to plot accuracy ?

Here is my layers and Option what sholud I do

numResponses = 1 ;
featureDimension =1;
numHiddenUnits =200;

layers = [ ...
    sequenceInputLayer(featureDimension)
    lstmLayer(numHiddenUnits)
    fullyConnectedLayer(numResponses)
    regressionLayer
];
maxepochs = 500;
miniBatchSize = 45 ; 
options = trainingOptions('adam', ...  %%adam
    'MaxEpochs',maxepochs, ...
    'GradientThreshold',1, ...
    'Shuffle','every-epoch', ... 
    'ValidationData',{XVal_ZaMir,YVal_ZaMir}, ...
    'ValidationFrequency',25,...
    'InitialLearnRate',0.005, ...
    'MiniBatchSize',miniBatchSize, ...
    'LearnRateSchedule','piecewise', ...
    'LearnRateDropPeriod',50, ...
    'LearnRateDropFactor',0.1, ...
    'Verbose',1, ...
    'Plots','training-progress');

Answer 1

[Omega]

To plot accuracy instead of RMSE in the training progress graph when using trainNetwork with LSTM for a classification task, you need to ensure that your network and training options are set up for classification. This involves using a classification layer and specifying accuracy as a metric in the training options.

Here’s how you can do it:

1. Ensure the network is set up for classification: Use a softmax layer and a classification layer.
2. Specify accuracy as a metric: Use the trainingOptions function to specify accuracy as a metric.