Which of the following approaches is used to mitigate the "curse of dimensionality" in machine learning?
- A) Increasing the number of features to improve model performance
- B) Using feature selection or extraction techniques to reduce the number of dimensions
- C) Increasing the dataset size to match the number of dimensions
- D) Decreasing the learning rate to reduce model complexity