What is weighted voting in ensemble learning and when should it be used?
Updated May 16, 2026
Short answer
Weighted voting assigns different importance to models based on their performance.
Deep explanation
In weighted voting ensembles, each model contributes to the final prediction proportionally to its weight, typically based on validation accuracy or log loss. This improves over simple majority voting by giving stronger models more influence. It is especially useful when models have unequal performance or are trained on different data distributions.
Real-world example
Search ranking systems assign higher weight to deep learning models than linear models.
Common mistakes
- Assigning weights arbitrarily without validation-based justification.
Follow-up questions
- How are weights typically chosen?
- What happens if weights are poorly chosen?