Skip main navigation

Feature Extraction

In doing traditional machine learning methods such as decision trees, how do we know essential features to decide the target?
A cute machine is writing down

When we are doing traditional machine learning methods such as decision trees, we must know essential features to decide the target. So before we make any machine learning model, we choose potential essential features. This process calls feature extraction.

Feature Extraction is a representation of the given raw data, and we can now use these for classic machine learning algorithms to perform a task—for example, the classification of the data into several categories or classes.

Feature Extraction is usually quite complicated and requires detailed knowledge of the problem domain. Also, the process of extracting features requires many trials and errors. After learning from the trial and error, the feature extraction is adapted and finalized. In other words, feature extraction is a very complicated and time and effort consuming process.

However, deep learning does not always need the Feature Extraction step. The layers learn an implicit representation of the raw data without any extra effort. You will input the raw data. Then the automatic feature extraction is done over several layers of artificial neural networks. In other words, the feature extraction step is already part of the process in an artificial neural network. When we train the artificial neural networks, the feature extraction is optimized by the neural network to obtain the input data’s best possible abstract representation.

Let me give you one example to help you understand better. If you want to use a machine learning model to determine if a particular image is showing a cat or a dog, we humans first need to identify the unique features or features of a cat or a dog (shape of nose, hair, feet, ears, etc.) we extract the information about these and give them to the algorithm as input data. In this way, the algorithm would perform a classification of the images. In a deep learning model, you do not need to make input like this. The model would recognize these unique characteristics of a cat or a dog and make the correct identification of a cat or a dog.

This article is from the free online

Artificial Intelligence and Machine Learning for Business

Created by
FutureLearn - Learning For Life

Our purpose is to transform access to education.

We offer a diverse selection of courses from leading universities and cultural institutions from around the world. These are delivered one step at a time, and are accessible on mobile, tablet and desktop, so you can fit learning around your life.

We believe learning should be an enjoyable, social experience, so our courses offer the opportunity to discuss what you’re learning with others as you go, helping you make fresh discoveries and form new ideas.
You can unlock new opportunities with unlimited access to hundreds of online short courses for a year by subscribing to our Unlimited package. Build your knowledge with top universities and organisations.

Learn more about how FutureLearn is transforming access to education