bagging
1
7.5
You train several models on different parts of the data and combine their answers to make predictions more reliable
- noun
- /ˈbæɡɪŋ/
- Specialized
:
agregación
- One common method in machine learning is bagging, which combines predictions from multiple models to improve accuracy.
Examples
-
The bagging approach reduces variance by training several models on different resampled subsets of the original data.
-
I did find their bagging technique useless as well.
Blog text (17) -
Using bagging techniques in statistics can lead to more stable and reliable results compared to single model predictions.
Synonyms
3
6.7
A set of computer models whose results are combined to make better predictions
is a specific method using resampled subsets and averaging to reduce variance
6.5
Combining results from different computer models to make one better result
focuses on resampling and averaging to reduce variance rather than weighting
How Complex
Surface Forms
Etymology
Bagging in machine learning is short for the phrase 'bootstrap aggregating', but a helpful image is of many small 'bags' of data: you train a model on each bag and then combine their answers. So, bagging means using many models together to make predictions more stable and accurate.