Chapter 3. Data Preprocessing
Anyone who is interested in machine learning (ML) would have certainly heard that 80% of a data scientist or machine learning engineer's time is spent on preparing the data, and the remaining 20% is spent on building and evaluating the model. The considerable time spent preparing the data is considered as an investment to construct a good model. A simple model this is made using an excellent dataset outpaces a complex model developed using a lousy dataset. In real life, finding a reliable dataset is very difficult. We have to create and nurture good data. You must be wondering, how do you create good data? This is something that we will discover in this chapter. We will study everything that is needed to create an excellent and viable dataset. In theory, good is relative to what task we have at hand and how we perceive and consume the data. In this chapter, we will walk you through the following topics:
- Data transformation
- Feature selection
- Dimensionality reduction...
