8.5 Classifying ML algorithms

“We can classify ML tools by how they learn (extract information) from data. Different varieties of ML use different algorithms that invoke different assumptions about the principles underlying intelligence (Domingos 2015). We can also categorize ML tools by the kind of experience theyare allowed to have during the learning process (Goodfellow et al. 2016), and we use this latter categorization here.” (Molina and Garip 2019, 28)
“In supervised machine learning (SML), the algorithm observes an output (Y) for each input(X). That output gives the algorithm a target to predict and acts as a teacher. In unsupervisedmachine learning (UML), the algorithm only observes the input. It needs to make sense of thedata without a teacher providing the correct answers—in fact, there are often no correct answers.” (Molina and Garip 2019, 28)

Molina, Mario, and Filiz Garip. 2019. “Machine Learning for Sociology.” Annu. Rev. Sociol., July.