What is ensemble diversity measurement and how is it quantified?

Ensemble diversity measures how differently models behave on the same inputs.

Diversity is a key factor in ensemble performance. It can be quantified using disagreement measures, correlation coefficients, Q-statistics, or entropy-based metrics. High diversity means models make different errors, allowing ensemble aggregation to cancel mistakes. However, too much diversity with weak models can hurt performance. Effective ensembles balance accuracy and diversity.