The machine learning perspective
What features do we really need to predict the income of a person? We will provide some ideas in this section.
You must learn how to explain and anticipate the output of an ML program before implementing it.
We will first display the training data using Facets Dive.
Displaying the training data with Facets Dive
We will continue to use the WIT_Model_Comparison_Ethical.ipynb
notebook in this section. We will begin by conducting a customized what-if ML investigation.
The U.S. Census Bureau is perfectly entitled to gather the information required for their surveys. The scope of ML analysis we are conducting is quite different. Our question is to determine whether or not we need the type of information provided by the U.S. Census Bureau to predict income.
We want to find a way to predict income that will break no laws and be morally acceptable. We want to remain ethical for applications other than U.S. census predictions...