I am working with RandomForestClassifier and I would like to be able to analyse the decision path which each decision tree takes for a single data point. What I understand is that the final prediction the random forest meta classifier makes is an aggregation of all the trees from the class with the most votes for each data point. Each data point that the classifier is predicting lands on only one leaf node so from all the trees it will take the class which is predicted the most from each of the 200 leaf nodes.

What I am interested in analysing is the decision path of a single data point through the decision trees of the random forest. I would love to see the final decision path that is the aggregation made by the meta classifier.

I have used Graphviz to visualise the decisions the classifier is taking, this will show me the path for a data point for one decision tree, however, I would like to find a way of getting this for all the trees involved in the aggregation made by the meta classifier.

Is this possible?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.