How do decision trees decide the value to be split upon for continous variables? [duplicate]

Question

I know that decision trees make the split based on some metric such as entropy, information gain, gini index etc. But for continous variables how does it figure the value at which to make a split. For example, my split is on age. How does it find out a specific age and decides values greater than this go to right, and vice versa.? Is it the mean value?

Answer 1

It tries all the possible splits and picks the one that reduces the loss function by the largest amount.