dimensionality reduction

KNIME Analytics Platform
1

hi

i'd like to reduce dimension of my data that has 104 attributes

and i want to dimensional reduction with random forest in knime

how can i do that???

if i have one workflow example it will be sooo good!

thanks so much

2

Hi,

The "examples server" can provide! Find the "Data Mining" subfolder starting with 002, in which the flow named "002004_Feature-Elimination_with_NaiveBayes" is probably right up your alley. Please note that I'm working with an old snapshot download of the examples server because to handle corporate firewalls, so in case the structure changed in the live version you may have to dig a little deeper!

HTH.

Cheers
E

3

Hi,

this blogpost might help you:

https://www.knime.org/blog/seven-techniques-for-data-dimensionality-reduction

The described workflow is available on the example server under 04_Analytics/01_Preprocessing/02_Techniques_for_Dimensionality_Reduction

The following forum thread also contains some information about how to select features using a random forest:

https://tech.knime.org/forum/knime-general/how-to-get-the-variable-importance-from-the-random-forest-model

Cheers,

nemad

4

Even better, thanks nemad!

E