I have a question about Random Forest and when it comes to Features for the Decision Tree.
Let us assume that we have a Numeric Feature named:
Now let us assume that we have registered values for this Feature like this:
My question is. How should we choose what values to check for in the actual descision Tree.
For example we could use:
RSI > 11
RSI < 37
RSI >= 57
Is there a technique on how this generally should be done and also how Many of those different checks should be done. In the above example, I just randomly choose 3.
For example I do use GINI Impurity for the above 3 to know which one is the best split.
But that is not the same as actually in the first place know which values to choose and how many. So this sentence is my question then?