Supervised Learning

Supervised learning is an approach for machine learning where your model goes over thousands of labeled entries in some data set.

For every input of the data set, your model returns an output. You use both the input and the output, combined with an Optimizer to adjust the model’s parameters.

You repeat this process many times. One iteration over the data set is called an epoch.

You’re supposed to split and randomize your dataset in at least 2 parts - a training one and a testing one.

• You use the training one for the actual learning process – adjusting your parameters.
• You then use the testing one to validate / evaluate the accuracy of the model, by giving it data that it’s never seen before. You don’t adjust the parameters when going over the testing dataset.