Explain the trade-off between bias and variance, and how would you address overfitting in a random forest model for Avetta safety risk use cases?
