cancel
Turn on suggestions
Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.
Showing results for 
Show  only  | Search instead for 
Did you mean: 
Try the Materials Informatics Toolkit, which is designed to easily handle SMILES data. This and other helpful add-ins are available in the JMP® Marketplace
Choose Language Hide Translation Bar
SophieC32
SophieC32
Level I

How are calculated "Contribution" in Predictor Screening ?

Created: Apr 8, 2022 03:58 AM  |  Last Modified: Jun 8, 2023 5:47 PM (1482 views)

Hello,

 

I would like to know precisely with the formula, how are calculated the contributions in Predictor Screening ?

Best regards,

 

Sophie

0 Kudos
1 ACCEPTED SOLUTION

Accepted Solutions
Ross_Metusalem
Staff Ross_Metusalem
Staff
Solution

Re: How are calculated "Contribution" in Predictor Screening ?

Apr 8, 2022 02:02 PM (1464 views)  |  Posted in reply to message from SophieC32 04-08-2022

The Contribution column in Predictor Screening corresponds to the Column Contributions you'll find in JMP's and JMP Pro's other decision tree-based platforms (Partition, Bootstrap Forest, Boosted Tree). The values in the column are the total sum of squares (for continuous responses) or G^2 (for categorical responses) attributed to each predictor, specifically taken from a Bootstrap Forest model with default settings (except the number of trees, which can be specified). See this help page for the calculations, and if you haven't worked with decision trees before, I recommend reading the Partition help more generally to better understand these calculations.

Ross Metusalem
JMP Academic Ambassador

View solution in original post

1 REPLY 1
Ross_Metusalem
Staff Ross_Metusalem
Staff
Solution

Re: How are calculated "Contribution" in Predictor Screening ?

Apr 8, 2022 02:02 PM (1465 views)  |  Posted in reply to message from SophieC32 04-08-2022

The Contribution column in Predictor Screening corresponds to the Column Contributions you'll find in JMP's and JMP Pro's other decision tree-based platforms (Partition, Bootstrap Forest, Boosted Tree). The values in the column are the total sum of squares (for continuous responses) or G^2 (for categorical responses) attributed to each predictor, specifically taken from a Bootstrap Forest model with default settings (except the number of trees, which can be specified). See this help page for the calculations, and if you haven't worked with decision trees before, I recommend reading the Partition help more generally to better understand these calculations.

Ross Metusalem
JMP Academic Ambassador