www.gusucode.com > nnet 案例源码 matlab代码程序 > nnet/ClassifyUsingSoftmaxLayerExample.m

    %% Classify Using Softmax Layer
% Load the sample data.

% Copyright 2015 The MathWorks, Inc.

[X,T] = iris_dataset;
%%
% |X| is a 4x150 matrix of four attributes of iris flowers: Sepal length,
% sepal width, petal length, petal width.
%
% |T| is a 3x150 matrix of associated class vectors defining which of the three
% classes each input is assigned to. Each row corresponds to a dummy
% variable representing one of the iris species (classes). In each column, a 1 in one
% of the three rows represents the class that particular sample (observation or example) belongs
% to. There is a zero in the rows for the other classes that the
% observation does not belong to.
%%
% Train a softmax layer using the sample data.
net = trainSoftmaxLayer(X,T);
%%
% Classify the observations into one of the three classes using the trained
% softmax layer.
Y = net(X);
%%
% Plot the confusion matrix using the targets and the classifications
% obtained from the softmax layer.
plotconfusion(T,Y);