Machine learning, Neural network with data in array format

Hi,

New to using NN in matlab. I'm looking to classify some data. Each event is contained in a 35x3 array. I have around 10k events so the matrix is 35x3x10000. The predictor is 10000x1 with values 0 or 1. I looked at the documentation and examples, but is not clear to me how I build and feed this data into the NN.

I'm doing this to classify simulation data for a neutron instrument. The events can be single or double scatter. Any help (and patience) will be greatly appreciated.

Rgds,

George

Answer 1

[KSSV]

You can give a try with this.

https://in.mathworks.com/help/deeplearning/ug/create-simple-deep-learning-network-for-classification.html

[George Suarez]

Thanks. I tried it before posting and didn't work. I'm sure is me no understanding how to configure the CNN.

%trainD is a 35x3x10000 array

trainDtarget = categorical(trainDtarget'); %10000x1 array

layers = [

imageInputLayer([size(trainD,1) size(trainD,2) 1])

convolution2dLayer(8,3,'Padding','same')

batchNormalizationLayer

reluLayer

maxPooling2dLayer(2,'Stride',2)

convolution2dLayer(3,16,'Padding','same')

batchNormalizationLayer

reluLayer

maxPooling2dLayer(2,'Stride',2)

convolution2dLayer(3,32,'Padding','same')

batchNormalizationLayer

reluLayer

fullyConnectedLayer(2)

softmaxLayer

classificationLayer];

options = trainingOptions('sgdm', ...

'InitialLearnRate',0.01, ...

'MaxEpochs',4, ...

'Shuffle','every-epoch', ...

'ValidationData',table(trainDtarget), ...

'ValidationFrequency',30, ...

'Verbose',false, ...

'Plots','training-progress');

net = trainNetwork(trainD,trainDtarget,layers,options);

I get an error

Error using trainNetwork (line 184)

Invalid network.

Caused by:

Layer 9: Input size mismatch. Size of input to this layer is different from the expected input size.

Inputs to this layer:

from layer 8 (size 17(S) × 1(S) × 16(C) × 1(B))

Answer 2

[yanqi liu]

yes，sir，may be upload your data mat file to analysis，or check

https://ww2.mathworks.cn/matlabcentral/answers/1649260-how-to-train-cnn-on-custom-dataset-in-matrix-form

make cnn model by the input data dimension