David J. Hand, "Dark Data: Why What You Don't Know Matters" (Princeton UP, 2020)
New Books in Economics
00:00
Exploring the Dark Data Perspective in Machine Learning
This chapter discusses the utilization of dark data in designing study or analytic tools, highlighting the reframing of classical sampling as choosing what data to ignore. Various machine learning techniques like Bayesian priors, bootstrapping, and boosting are examined through a dark data lens, with boosting illustrated in diagnosing illnesses by creating data for misclassified instances. It also delves into the significance of dark data as a unifying framework for data analysis and the potential implications for education and addressing misunderstandings in statistical inference.
Transcript
Play full episode
