Understand Tree Decision

[AnaBerta]

Hello

I am starting with Knime and I have created a workflow to evaluate different calcification algorithms. For my data, the best is Tree because a I have a good accuracy.
My question is related about how can I understand the algorithm, and what are the fields of my data that have more influence in the prediction.
I am trying to determine if a customer is going to leave the company and I have 50 fields in my database, but I don´t know what are more important fields to decide that a customers is going to leave.

Is there any way to decide this?

Ana

[mlauber71]

I recently wrote this entry. Maybe the links there might help you explore more about machine learning

How to compare different models

OK, there is the Binary Classification Inspector that would give you an idea about the model. Binary Classification Inspector Example I created two Meta nodes for binary and regression models that would run a certain set of parameters and create some graphics (you would need some R packages). Input would just be a solution column with the truth either as 0/1 string or the number and a “submission” which is the score from 0-1. You could collect the results from the statistic tables and see whi…

[AnaBerta]

Thank you for you answer.
I need to analyze better but as I understand this workflow compare two models but when you have selected a models, it´s possible to know what fields have more weight in the decision?

Ana

[ipazin]

Hello @AnaBerta,

and welcome to KNIME and it’s Community!

@AnaBerta

it´s possible to know what fields have more weight in the decision?

Sure there is. In case of Decision Tree algorithm (which has tree structure as output) the top-level features are the more important that the others.

Check this topic with further topics linked for more about it:

Feature Importance using decision tree

Hi Everybody, I have a question is it possible to extract feature importance in Knime using decision tree learner and predictor? Using sklearn I usual use the following code for it: from sklearn.tree import DecisionTreeClassifier dt = DecisionTreeClassifier() dt.fit(X_train, y_train) dt.score(X_test, y_test) dt.feature_importances_ For sure I could use the python node, but I try to transfer my normal tasks to KNIME without using python. Does somebody have an idea?

Br,
Ivan

[AnaBerta]

Thank you for your answer. I am going to test.

Ana