Introduction
Hey there, today we are going to demystify feature engineering. It seems like a tough topic to cover but I hope at the end of this article you will at least understand the basics of it.
From Wikipedia, Feature engineering is a machine learning method that uses data to create new variables that are not included in the training set.
It can generate new features for both supervised and unsupervised learning.
Makes data transformations easier and faster while improving model's accuracy.
Feature engineering Techniques
- Data Cleaning: this is tidying up your data. You address missing information, correct errors, and remove any inconsistencies.
- Data Transformation: this is data reshaping or adjusting. Example: scaling large numbers down or normalizing data so that it fits within a certain range. The important factor is to make these changes without altering the data meaning.
- Feature Extraction: This is where we explore existing data and create new features that can offer new insights. This makes the model simpler and faster without losing useful details.
- Feature Selection: involves picking out the pieces of data that are most closely related to your target prediction. This gets rid of unnecessary information, making the model more focused.
- Feature Iteration: This is all about trial and error. The process of adding or removing certain features, test how they impact the model, and keep the ones that improve the model’s performance.
Types of Features in Machine Learning
- Numerical Features: These are numbers that can be measured, are straightforward and continuous in nature. Example : age.
- Categorical Features: These are categorical. Example, eye color.
- Time-series Features: data that is recorded over time. Example; Stocks.
- Text Features: these are features made from words or text. Example : Customer reviews
Conclusion
I hope I explained the terms well, I do believe this are the few things you need to know as a beginner theoretically. Next time we discuss feature engineering it will be on more practical terms.
Till next time!
Top comments (0)