Model > Evaluate regression

Evaluate regression model performance

To download the table as a csv-files click the top download button on the right of your screen. To download the plots at a png file click the lower download icon on the right of your screen.

Response variable

The numeric outcome, or response, variable of interest.

Predictor

Select one or more variables that can be used to predict the value of the response variable. This could be a variable or predicted values from a model (e.g., from a regression estimated using Model > Linear regression (OLS) or a Neural Network estimated using Model > Neural Network).

Show results for

If a filter is active (e.g., set in the Data > View tab) generate results for All data, Training data, Test data, or Both training and test data. If no filter is active calculations are applied to all data.

Example

Predictions were derived from a linear regression and an neural network with two nodes in the hidden layer on the diamonds data. The variables price and carat were log-transformed prior to estimation.The data is available through the Data > Manage tab (i.e., choose Examples from the Load data of type drop-down and press Load). The predictions shown below were generated in the Predict tab.

The test statistics show a small, but consistent, advantage for the NN.

Report > Rmd

Add code to Report > Rmd to (re)create the analysis by clicking the icon on the bottom left of your screen or by pressing ALT-enter on your keyboard.

Rsq, RSME, and MAE are plotted by default. It is possible to customize the plotted results through Report > Rmd. To change the plot use, for example:

plot(result, vars = "Rsq")

The plot can be further customized using ggplot2 commands (see example below)). See Data > Visualize for details.

plot(result, vars = "Rsq") +
  labs(caption = "Based on data from ...")

R-functions

For an overview of related R-functions used by Radiant to evaluate regression models see Model > Evaluate regression

Vincent R. Nijs, Rady School of Management (UCSD)