Understand, Manage, and Prevent Algorithmic Bias: A Guide for Business Users and Data Scientists

© Tobias Baer 2019

Tobias BaerUnderstand, Manage, and Prevent Algorithmic Biashttps://doi.org/10.1007/978-1-4842-4885-0_10

10. Biases Introduced by the Algorithm Itself

Tobias Baer¹ 

(1)

Kaufbeuren, Germany

 

In the discussion so far, you have experienced algorithms as neutral and fact-driven and actively pursuing the goal of debiasing decision-making. The types of algorithmic biases reviewed all originated outside of the algorithm, such as in real-world biases or inadequate data. In this chapter, we will dive deeper in how an algorithm works and discover situations in which an algorithm “randomly” introduces new biases in the sense of prejudice against specific profiles of instances. Much of this can be considered noise, but every once in a while, ...