Classification of FashionMNIST
FashionMNIST is a dataset of small, labeled fashion images meant as a more difficult replacement to the overused MNIST handwritten digits dataset. This example shows how to train LeNet on this dataset to classify images into the 10 available categories.
Get the necessary data from the Wolfram Data Repository.
This a a random sample of the dataset.
Train a small model to classify the images, using 10% of the data as a validation set.
Use the test dataset to compute the accuracy of the final classifier.
Visualize the confusion matrix of the classifier.
Visualize the average receiver operating characteristic (ROC) curve for every class. This plot shows the relation between the true positive rate (or recall) and the false positive rate (or fall-out) in a binary classifier.
Compute the average ROC across all the classes.
Use FeatureSpacePlot3D with the features extracted by the net to build a 3D visualization for the dataset content.
