# Quick and Easy Guide to Machine Learning Terminology: Learn Key Concepts in 5 Minutes

In this tutorial, we will explore key terms about machine learning. We’ll introduce some of the most important machine learning terminology and concepts to help you better understand this exciting field.

Before I start explaining it to you, I just want you to understand the differences between AI, Machine Learning, and Deep Learning with the following image:

Image Source

As you know, machine learning is an exciting and rapidly growing field that is changing the way we approach a wide variety of problems, from predicting customer churn to detecting fraud. But, as with any field, machine learning comes with its own set of terminology and jargon – as you can see in some tutorials like Linear Regression, Polynomial Regression or K-Nearest Neighbors.

## Framing the Problem

Before we dive into the terminology, it’s important to understand the context in which machine learning (ML) is typically used. In general, ML is used to help us solve problems where we don’t know the exact solution in advance. Instead, we have a set of inputs, and we want to train a machine learning model to predict an output based on those inputs.

To frame a machine learning problem, we typically start by defining our inputs (also known as features) and our output (also known as the target variable).

For example, let’s say we want to build a machine learning model to predict whether a customer is likely to churn from our subscription service. In this case, our inputs might include information about the customer’s demographics, usage patterns, and purchase history, while our output (the target variable) would be a binary indicator of whether the customer churned or not.

With this basic framing in mind, let’s dive into some of the key machine learning terminologies:

1. Model: A model is a mathematical representation of the relationships between our inputs and our output. In other words, it’s a way of encoding the patterns in our data so that we can make predictions based on new inputs. (Image Source)

2. Training: Training is the process of using our existing data to teach a machine learning model how to make predictions. During training, the model is presented with examples of inputs and the corresponding outputs, and it adjusts its internal parameters to better predict the outputs from the inputs. (Image Source)

3. Test: Once we’ve trained our model, we need to evaluate how well it performs on new, unseen data. This is typically done by setting aside a portion of our data as a test set and using it to see how well our model predicts the outputs based on the inputs. (Image Source)

4. Feature: A feature is a measurable aspect of the inputs that we think might be relevant to predicting the output. Features can be anything from demographic information to website clickstream data.
5. Label: The label (also known as the target variable) is the output that we’re trying to predict. In supervised learning (which we’ll discuss more below), the label is typically provided as part of our training data.
6. Supervised learning: Supervised learning is a type of machine learning in which we have both inputs and outputs (labels) for our training data. In other words, we’re trying to learn a mapping from inputs to outputs.
7. Unsupervised learning: Unsupervised learning is a type of machine learning in which we only have inputs and no corresponding outputs. The goal of unsupervised learning is typically to find patterns or structures in the data. An explanation of the last 3 points: label, supervised, and unsupervised (Image Source)

8. Overfitting: Overfitting occurs when a model is too complex and starts to memorize the training data rather than learning the underlying patterns. This can lead to poor performance on new, unseen data.
9. Underfitting: Underfitting occurs when a model is too simple and fails to capture the underlying patterns in the data. This can also lead to poor performance on new, unseen data. (Image Source)

## Conclusion

This is just a small sample of the many machine learning terms and concepts that you’re likely to encounter. However, it already allows you to understand some of the steps you did during the machine learning tutorials that you found here on Geekering.

What are you waiting for to get started? Shall I give you a first challenge? Click here.

### Extra Bonus – The difference between Machine Learning and Deep Learning:

Machine learning is when a computer learns how to do something by looking at examples. Just like how you learn how to draw a picture by looking at how other people draw, a computer can learn how to do things by looking at lots of examples.

Deep learning is a special way that computers can learn. It’s like when you build a tower out of blocks – you start with one block, and then add more and more until you have a really tall tower. Deep learning is like building a tower out of lots and lots of blocks – each block adds a little bit more to the tower, and eventually, you have a really big tower that can do something really cool, like recognize what’s in a picture or understand what someone is saying.

So, machine learning is like learning by looking at examples, and deep learning is a special way of learning that involves building really big towers out of lots of little blocks.

(Image Source)