machine learning - Testing The Trained Neural Network - Matlab -


i want use neural network classify handwritten digits of mnist dataset

i have created 2 layer neural network 100 hidden unit , trained using 60,000 * 784 trainimages matrix , 60,000 * 1 trainlabels

net = newff(trainimages,trainlabels,100) 

how test , calculate error rate of trained network 10,000 * 784 testimages

training:

[net,tr]=train(net,trainimages',trainlabels'); 

testing:

predictedlabels = sim(net,testimages'); error_rate = 1- mean(predictedlabels == testlabels'); 

Comments

Popular posts from this blog

php - regexp cyrillic filename not matches -

c# - OpenXML hanging while writing elements -

sql - Select Query has unexpected multiple records (MS Access) -