How to classify feature vectors of an image using fully connected layer

I have extracted image features using 3 different pretrained network and contenated them. Now I want to classify them using fully connected layer.

%%clear workspace;

path = char('E:\Mine\cluster'); %pass to this variable your complet data set path

net=alexnet();

net1=googlenet();

net2=resnet101();

imds = imageDatastore(fullfile(path),'IncludeSubfolders',true, 'LabelSource', 'foldernames');

augImds=augmentedImageDatastore(net.Layers(1, 1).InputSize(1:2),imds);

augImds1=augmentedImageDatastore(net1.Layers(1, 1).InputSize(1:2),imds);

augImds2=augmentedImageDatastore(net2.Layers(1, 1).InputSize(1:2),imds);

Labels=imds.Labels;

% count the number of images

numClass=numel(countcats(Labels));

% feature extraction with the pre-trained network

featurea=squeeze(activations(net,augImds,'fc8'));

featureg=squeeze(activations(net1,augImds,'inception_5b-1x1'));

featurer=squeeze(activations(net2,augImds,'res5b'));

combined_Vector = featurea( : ), featureg(:), featurer( : );

layer = fullyConnectedLayer(2,'Name','fc1');