How to add R-square to Observed vs Predicted Scatterplot for Random Forest regression analysis?
Article ID: KB0074198
Updated On:
| Products | Versions |
|---|---|
| Spotfire Statistica | 13.3,13.2,13.1,13.0,12.7,12.6,12.5 SP1,12 SP3,12 SP2,12.0A,12 SP1,11.0A,10.0 |
Description
In Random Forest regression analysis, to calculate and add R-square value to the Observed vs Predicted Scatterplot:
In the Random Forest Regression result dialog, under "Prediction" section, click on "Observed vs. Predicted", the scatterplot for Observed vs. Predicted values will be generated. The R-square for the relationship is not calculated and displayed in the scatterplot.
Issue/Introduction
Resolution
In Random Forest regression analysis, to calculate and add R-square value to the Observed vs Predicted Scatterplot:
1. In the Random Forest Regression result dialog, under "Prediction" section, click on "Predicted values".
2. A table with Observed and Predicted values will be displayed in analysis result workbook. Highlight the two columns, right click mouse and click on "Copy with Headers" from the drop-down menu.
3. Go to Statistica "Home | New | Spreadsheet" to create a new empty spreadsheet.
4. Paste the copied Observed and Predicted values columns to the new spreadsheet. Then go to Statistica "Graphs | Scatterplot". In the scatterplot dialog, click on "Variables" to select Observed and Predicted variables as X and Y axis.Â
5. In the scatterplot dialog, continue to go to "Advanced" tab, there are options user can select to display in the scatterplot: "R square", "Corr. and p" and "Regression equation".
Additional Information
Feedback
