how to get good accuracy in mlp & k-fold crossvalidation code

hello, can anyone help me...

I found a problem in k fold cross validation, my research uses 4 classes of right hand, left hand and right, left foot movements.

I need the mlp & k fold cross validation code with good accuracy

please help me, email: [e-mail]

for df = 1:10

n = 1;

for c = 1:6520

if indices (1,c) == df

K_fold_test(n) = c;

n = n +1;

end

end

n = 1;

for c = 1:6520

if indices (1,c) ~= df

K_fold_train(n) = c;

n = n +1;

end

end

%test = (indices == df);

%train = ~test;

%train = train.';

U_testing = T_data_time(1:19,K_fold_test); %data testing MLP K_fold

%test = test.';

U_training = T_data_time(1:19,K_fold_train); % data training MLP K_fold

Target_test_MLP = Target_new(1:4,K_fold_test);

Target_train_MLP = Target_new(1:4,K_fold_train);

% h =[16];

h = [8 2]; %Jumlah node pada hidden layer

% h = [16 2];

%X = (T_var_time_nn_stdz).'; %pilih salah satu feature

% Data_train_MLP = (U).';

% Data_test_MLP = (UT).';

% Data_target_test = (Target_test_MLP).';

% Data_target_train = (Target_train_MLP).';

lambda = 1e-2;

[model, mse] = mlp(U_training, Target_train_MLP, h, lambda);

% plot(mse);

% disp(['T = [' num2str(Data_target_test) ']']);

Y_norm = mlpPred(model, U_testing);

% disp(['Y = [' num2str(Y) ']']);

Y_Max = max(Y_norm);

Y_output = zeros(size(Y_norm));

match = (Y_norm == Y_Max );

Y_output(match) = 1;

Y_output(~match)= 0;

Y_Round = round(Y_norm);

Y_R_all = [ Y_R_all Y_Round ] ;

plotconfusion(Target_test_MLP, Y_output)

figure

%Y_Round = round(Y);

%Y_R_all = [ Y_R_all Y_Round ] ;

end