how to design custom ANN using Deep Network designer app MATLAB

Hello Everyone, I Hope you are doing well.

I want to create a simple ANN using Deep network Designer app or using code.

ANN contain input layer, 10 neurons in hidden layer with sigmoid activation and output layer with classifciaton and softmax layer

I have the dataset of shape 250x1000. I have attached the dataset below. which contain label as well.

I also want to be label in catogorical form . Like 1 name as 'Class1'

How can i do it in MATLAB

[Med Future]

@Abolfazl Chaman Motlagh i want input as 1x1000,

five classes have 50 samples total of 250x1000

Answer 1

[yanqi liu]

clc; clear all; close all;
load Datasetn
layers = [
    imageInputLayer([1000 1 1])
    fullyConnectedLayer(10)
    reluLayer
    fullyConnectedLayer(5)
    softmaxLayer
    classificationLayer];

opts = trainingOptions('adam', ...
    'MaxEpochs',200, ...
    'Shuffle','every-epoch', ...
    'Plots','training-progress', ...
    'Verbose',false, ...
    'ExecutionEnvironment', 'cpu', ...
    'ValidationPatience',Inf);
xc = reshape(dataset', [1000,1,1,250]);
% 'Class1' 
yc = [];
for i = 1 : length(label)
    yc{i,1} = ['Class'  num2str(double(label(i)))];
end
yc = categorical(yc);
disp(yc)

net = trainNetwork(xc,yc,layers,opts);

[Med Future]

@yanqi liu Can i divide the dataset between 80% train and 20% test?

[yanqi liu]

yes，sir

clc; clear all; close all;
load Datasetn
idx = randperm(length(label)) ;
dataset = dataset(idx,:);
label = label(idx,:);
m = round(length(label)*0.8) ;  
dataset1 = dataset(1:m,:); label1 = label(1:m,:);
dataset2 = dataset(1+m:end,:); label2 = label(1+m:end,:);

layers = [
    imageInputLayer([1000 1 1])
    fullyConnectedLayer(10)
    reluLayer
    fullyConnectedLayer(5)
    softmaxLayer
    classificationLayer];

opts = trainingOptions('adam', ...
    'MaxEpochs',200, ...
    'Shuffle','every-epoch', ...
    'Plots','training-progress', ...
    'Verbose',false, ...
    'ExecutionEnvironment', 'cpu', ...
    'ValidationPatience',Inf);
xc = reshape(dataset1', [1000,1,1,m]);
% 'Class1' 
yc = [];
for i = 1 : length(label1)
    yc{i,1} = ['Class'  num2str(double(label1(i)))];
end
yc = categorical(yc);
disp(yc)

net = trainNetwork(xc,yc,layers,opts);