Feature selection-exercise- I am not understanding how to do

Hi, I have to finish a workflow for an only course. But I don’t find the node they ask me to use. The argument is backward feature elimination. And in the practice solution given by the professor from 2016 knime it is used a backward feature elimination start 1:1 connected to a naivebayes 3.7 and a weka predictor. Then there is a backward f elimination end. Which is connected to a backward feature elimination filter, that is connected also to a partition b test. In my knime there is a meta node backward feature elimination which is composed by a feature selection loop start, a partitioning, anaive bayes learner, a n b predictor, a scorer, a feature selection loop end and filter. I don’t understand how to do. Can you help me please?

Dear @lorenza1992,

You can use the great https://nodepit.com to search for nodes. A search for “backward feature” reveals that the nodes have been deprecated. Looking at the search results, there is a Feature Selection Loop Start (1:1) that has a “Feature selection strategy” option with the following description:

Here you can choose between two selection strategies: Forward Feature Selection and Backward Feature Elimination.

Hence, you should use this (and similar) nodes to replace the Backward Feature Elimination Start (1:1) node.

Best,
Stefan

2 Likes

Thank you! Yes yesterday I solved my problem by myself just like you are advising me! Thank you again for your comment

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.