-
Book Overview & Buying
-
Table Of Contents
Learning Apache Mahout
By :
Let's start by understanding what is meant by feature engineering. Feature engineering is performed after data cleansing and preparation, before or even during model training. It aims to provide better representation of the data to the machine learning algorithm. Feature engineering as a process has multiple outcomes and can impact the overall modeling exercise in many ways. Feature engineering can be focused to increase model accuracy and generalization, decrease the computation requirements for large and wide datasets, and make the model simpler. Generally, a practitioner aims to do all of these. Feature engineering can be divided into four major tasks: feature construction, feature extraction, feature selection, and dimensionality reduction. We will discuss the four tasks shortly.
Before we discuss feature engineering, let's revisit the definition of features first:
Change the font size
Change margin width
Change background colour