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);