Deep Neural Network Training (Non-Image Classification)

Dear MATLAB Community,

I'm tying to train a Neural Network for Non-Image Classification task:

• I have a table 30x14 called Dataset containing 13 Columns which are the Features and the last column (14th) has the labels.
• Each Row in Dataset is one example.
• I also have the features saved as a 30x13 Matrix called "features" containg the features and a 30x1 Matrix called "labels" containing the labels.
• I also build an architecture using Deep Neural Network Designer Application in MATLAB and saved my architecture in a variable called "layers".
• I need to build and train a neural network for classifying future examples and assess the accuray of my model.
• I have options set as

options = trainingOptions('sgdm', ...
    'MaxEpochs',20,...
    'InitialLearnRate',1e-4, ...
    'Verbose',false, ...
    'Plots','training-progress');

Attempts:

(1) net = trainNetwork(Dataset,layers,options)
(2) net = trainNetwork(features,labels,layers,options)

Both attemps failed and gave errors

(1)

Error using trainNetwork (line 165)

Invalid training data table for classification. Predictors must be in the first column of

the table, as a cell array of image paths or images. Responses must be after the first

column, as categorical labels.

(2)

Error using trainNetwork (line 165)

Too many input arguments.

Caused by:

Error using trainNetwork>iParseInputArguments (line 290)

Too many input arguments.

Anyone has any idea what to do ?

Thank You

Answer 1

[JESUS DAVID ARIZA ROYETH]

For your problem I recommend that you use patternnet

example:

[x,t] = iris_dataset; %x is 4x150 (4 features,150 samples) , y is 3x150 (3 total labels,150 samples)
net = patternnet(10);%create net
net = train(net,x,t);%train net
view(net)
y = net(x);%evaluate net
perf = perform(net,t,y);%test performance
classes = vec2ind(y);%all net outputs 
futureclassification=vec2ind(net([6.7 3.1 5.6 2.4]'))%your future data here

[Hasan Abed Al Kader Hammoud]

Thanks for the reply but I want to use my own architecture.