In the previous chapter, we established the need to automate the process of building machine learning models. In this chapter, we explain what Automated Machine Learning is, the different techniques involved in this process, and how they all come together. We will also give a quick overview of automated ML on Microsoft Azure Machine Learning.
In Chapter 1, we discussed how coming up with a good machine learning model can be time-consuming and tedious, given all the possible combinations to explore. Automated Machine Learning is a recent development in machine learning focused on making that entire process easy, with the goal of bringing efficiency to data scientists as well as enabling non–data scientists to build models.
Let’s go through the stages of the machine learning process and see how Automated Machine Learning can help at each stage.
As briefly discussed in the previous chapter, real-world data is not clean and requires a lot of effort to get to a usable state. Understanding input data is a crucial step toward formulating the machine learning problem.
Automated Machine Learning can help here by analyzing the data and automatically detecting the data type of each column. Column types could be Boolean, numeric (discrete or continuous), or text. Automatically detecting these column types helps with subsequent stages like feature engineering.
In many cases, Automated Machine ...