This article is a translation of Jason Brownlee's feature design article, which is linked here.
Feature design is an informal topic, but it is undoubtedly considered to be the key to successful application of machine learning.
When I created this guide, I did my best to learn and analyze all the materials in a wide and deep context.
You'll find out what the feature design is, what it solves, why it's important, how to design features, who does it well and where you can go to learn more and be good at it.
If you read an article about feature design, I hope it is this one.
"Feature design is another topic that doesn't seem worthy to be admired in any review articles or books, or even chapters in the book, but it's definitely the key to machine learning success. [...] The success of many machine learning is actually a success in designing a trait that learners can understand. "
--scott Locklin
Problems solved by Feature design
When your goal is to get the most likely results from a predictive model, you need to get the most out of all of your.
This includes getting the best results from the algorithm you are using. It also involves making the best use of the data for your algorithm.
How do you model your predictions to make the most of your data?
This is the process of feature design and the practice of solving problems.
"In fact, the success of all machine learning algorithms depends on how you present the data. "
--Mohammad
The importance of feature design
The characteristics of your data directly affect the predictive model you use and the results you will get.
[Feature selection] DIscover Feature Engineering, how to Engineer Features and how to Get good at It translation