
上QQ阅读APP看书,第一时间看更新
When should we use a random forest?
The following are some examples of when to use random forests:
- When model interpretation is not the most important criterion. Interpretation will not be as easy as a single tree.
- When model accuracy is most important.
- When you want robust classification, regression, and feature selection analysis.
- To prevent overfitting.
- Image classification.
- Recommendation engines.