In this workflow I apply the fourier transform to audio recordings of 4 different sources to generate cross sectional data for use in a classification task.
Using reduced data set for example purpose.
This is a companion discussion topic for the original entry at https://kni.me/w/V3gkrNaVmC5Z14xo
Hello KNIME Fellows,
I am learning a bit about ML with knime and I have encountered recently this algorithm, great stuff but I miss one generall little hint …
// If anyone could share a thought on what are “tree split criteria” of gradient boosted trees in generall or how this works within an algorithm itself … what are the strategies … for handling it automatically ?
Thou I could imagine or build many possible trees, I cannot see how KNIME node actually does it ? (I mean within " Gradient Boosted Trees Learner" itself)… For Example: If I had to predict the person height based on persons weight and and size of the foot, I could split the tree; say… all persons having footsize less then “some number” … but how is it determined witihn that particular node …?
… and is there any convenient way to show the whole tree of a model ? like in a scikit python … or like in a Decision Tree Predictor where I can go with right click … view: decision tree view …