Types of machine learning

There are three major types of machine learning.

Supervised.

Unsupervised and reinforcement supervised learning refers to the case where we provide the algorithm

with inputs and their corresponding desired outputs based on this information.

It learns how to produce outputs as close to the ones we are looking for.

Both examples the one about the weather forecast and the one about preparing coffee Illustrated supervised

learning that was intentional as it is the focus of this course.

All right.

The second type of machine learning is unsupervised learning and unsupervised learning.

We feed inputs but there are no target outputs.

This means we don’t tell the algorithm exactly what our goal is.

Instead we ask it to find some sort of dependence or underlying logic in the data provided.

For instance imagine we administrate the Web site Katz and Dago’s dotcom users have the option to submit

photos of their cats or dogs to the Web site once a photo has been submitted.

We would like it to be automatically classified in the subpage cats or the subpage dogs.

In supervised learning we would train the algorithm on a data set of say 1000 cat photos and 1000 dog

photos each would be labeled cat or dog.

The model would then learn how to interpret an input picture as either a cat or a dog.

Minimizing mismatches over the training set.

Sometimes however we may not have the resources or the need to label the whole data set in the previous

dataset.

There was a person who manually labeled all 2000 pictures.

Now imagine a data set of two million pictures.

If it takes a person 5 seconds to label a picture a data set of two million observations would take

around 2800 hours or 345 working days to label all pictures.

With unsupervised learning though we can train the algorithm without labeling the photos or in an unsupervised

way.

We could simply ask it to split them into two groups based on visual similarities.

The result would be two groups that are unlabeled.

Once we have obtained that we can examine them and say oh yes the first set is dogs and the second set

is cats thinks algorithm unsupervised learning is especially useful when our goal is to split a dataset

into a certain number of categories which we do not know prior to implementing it.

That by the way is called clustering.

OK the final type of machine learning is reinforcement learning.

Without digging too deep into it with reinforcement learning we would treat a model to act in an environment

based on the rewards it receives.

It is much like training your pet and rewarding it with treats every time it achieves a goal.

Cits rolls over or gives you a pawn in the same way the machine learning algorithm can be taught how

to play Super Mario by rewarding it for progressing with an increase in score.

All right.

Supervised learning is the focus of this course.

That is because it is the simplest and most commonly used to exhaust the topic.

Supervised learning could be divided into additional subtypes classification and regression.

The difference is very straightforward classification supervised learning models provide outputs which

are categories such as cats or dogs in regression supervised learning models.

The outputs will be of numerical type.

For instance predicting the Eurodollar exchange rate will always give us a continuous number like 1

point to 1 or 1.1 9.

In this course we will create both classification and regression algorithms.

Stick around.

And thanks for watching.