I want to build a model that generates a risk score based on a decision tree of six levels. All these levels and options have the same weight (1).
The first level has three options. The second level has the same four options for each of the options in the first level. The third level features two options regardless of the response in level two. The fourth level has two options regardless of the response in level three. If the answer in level four is two, we skip level five and go straight to level six. The fifth and the sixth levels have only one option regardless of the answer in the previous levels.
The tree looks like this:
1 2 3 [1]
1234 1234 1234 [2]
12 12 12 [3]
12 12 12 [4]
1 1 1 [5]
1 1 1 [6]
If information in level five is provided and this information matches with the data inputted by another user, then the risk score for this user is added by another point, and so on. As the risk score can grow indefinitely based on network connections, I would like to limit it to $0-10$.
I am not quite sure whether this a probability problem as the weight of the levels and options are all the same across this decision tree.
Hope this question makes sense and the explanation of the problem is clear enough.
Thanks,