CODEWITHSUNDEEP
neural-networkartificial-intelligenceclassificationprediction
user1646196
user1646196

Reputation: 588

Multiple output data set for neural network

I've developed an Artificial Neural Network in C++ for an assignment and I was wondering if anyone knew of any multiple output data sets that I could use to test my network? I'm trying to prove that a modification I've done causes the network to perform better with a wide range of data sets.

I've tried a few sites but none of them state the number of outputs leading me to assume that they all have one output.

Upvotes: 2

Views: 671

Answers (2)

greeness
greeness

Reputation: 16124

UCI machine learning repository currently maintain 235 data sets as a service to the machine learning community.

A large percentage of them contains multiple types of labels, e.g., 

WALKING, WALKING_UPSTAIRS, WALKING_DOWNSTAIRS, SITTING, STANDING, LAYING

unacc, acc, good, vgood

Upvotes: 1

Marc Shivers
Marc Shivers

Reputation: 698

MNIST is a standard 10-class dataset of 20x20 pixel images of handwritten digits (0 - 9). You can find the dataset (along with performance results from various classifiers, including neural nets) at http://yann.lecun.com/exdb/mnist/

Upvotes: 0

Related Questions