Random Forest Review

Random Forest Review
- What is random forest classification?

1. Sample the dataset to multiple testsets
2. Choose suitble number of features used in each testsets.
3. Create a Decision Tree model with each of the testsets
4. Majority vote / Aggregation from multiple models\

Tips:

• How many features to select in each tree? The number close to log or sqrt of the total feature is better.
• The random feature can help to reduce the correlation between trees.