filename = "creditcard.csv";
tbl = readtable(filename,'TextType','String');

fraudRows = tbl(tbl.Class == 1, :);

% Determine the amount of oversampling needed
numFraud = height(fraudRows);
numNonFraud = height(tbl) - numFraud;
numReplications = numNonFraud - numFraud;  % Adjust this as needed

% Randomly replicate fraudRows
replicatedRows = fraudRows(randi(numFraud, numReplications, 1), :);

% Combine the original data with the replicated rows
tblAugmented = [tbl; replicatedRows];

% Shuffle the combined dataset
tblAugmented = tblAugmented(randperm(height(tblAugmented)), :);
tbl = tblAugmented;

labelName = "Class";
tbl = convertvars(tbl,labelName,'categorical');

classNames = categories(tbl{:,labelName});
numObservations = size(tbl,1);
numObservationsTrain = floor(0.7*numObservations);
numObservationsValidation = floor(0.15*numObservations);
numObservationsTest = numObservations - numObservationsTrain - numObservationsValidation;
idx = randperm(numObservations);
idxTrain = idx(1:numObservationsTrain);
idxValidation = idx(numObservationsTrain+1:numObservationsTrain+numObservationsValidation);
idxTest = idx(numObservationsTrain+numObservationsValidation+1:end);
tblTrain = tbl(idxTrain,:);
tblValidation = tbl(idxValidation,:);
tblTest = tbl(idxTest,:);
numFeatures = size(tbl,2) - 1;
numClasses = numel(classNames);
 
% More complex architecture
layers = [
    featureInputLayer(30)% 148 input features
    fullyConnectedLayer(500)
    batchNormalizationLayer
    reluLayer
    dropoutLayer(0.5)
    fullyConnectedLayer(100)
    batchNormalizationLayer
    reluLayer
    dropoutLayer(0.5)
    fullyConnectedLayer(numClasses)
    softmaxLayer
    classificationLayer];


% Rest of your code for training

% Update training options
options = trainingOptions('adam', ...
    'InitialLearnRate',0.00001, ...
    'LearnRateSchedule', 'piecewise', ...
    'LearnRateDropFactor', 0.2, ...
    'LearnRateDropPeriod', 50, ...
    'MaxEpochs',300, ...
    'MiniBatchSize', 128, ...
    'ValidationData',tblValidation, ...
    'ValidationFrequency',30, ...
    'Verbose',false, ...
    'Plots','training-progress');

miniBatchSize = 15;
net = trainNetwork(tblTrain,labelName,layers,options);
YPred = classify(net,tblTest(:,1:end-1),'MiniBatchSize',miniBatchSize);
YTest = tblTest{:,labelName};
accuracy = sum(YPred == YTest)/numel(YTest);
figure
confusionchart(YTest,YPred);
nnnn = grp2idx(YTest);
mmmm = grp2idx(YPred);
[c_matrixp,Result2]= confusion.getMatrix(nnnn,mmmm);
figure
plotConfMat(c_matrixp,{'Non-Fraud','Fraud'});