Data Science for Business: What You Need to Know about Data Mining and Data-Analytic Thinking

by Foster Provost

analysts often claim to prefer squared error because it strongly penalizes very large errors.

Squared error = strong penalisation of very large errors