Chapter 7. Advanced feature engineering


This chapter covers

  • Using advanced feature-engineering concepts to increase the accuracy of your machine-learning system
  • Extracting valuable features from text by using natural-language-processing techniques
  • Extracting meaning from images and using them as features in your machine-learning project

You explored the basic concepts behind feature engineering in chapter 5 and applied simple feature-engineering techniques to real-world data in chapter 6. In this chapter, you’ll look at more-sophisticated techniques that you can use when faced with types of data that have become common in today’s world. The two most important of these are text and images. This chapter presents advanced techniques for extracting features from text and image data, in order to use this data in your machine-learning pipelines.

7.1. Advanced text features

You already looked at simple feature engineering for text data in chapter 5. This section provides more details about the ideas behind these techniques, and presents more-advanced concepts that can improve the accuracy of your models even further.

7.2. Image features

7.3. Time-series features

7.4. Summary