Child pages
  • Impact on Target (8.0)

Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

This new feature allows comparing a segment with a selected benchmark (either the entire data set, or another segment). The comparison is based on how the observable variables of the segment can impact is in terms of Impact, a term combining the difference of the observable variables' mean values and their effect on the Target node value.

The Impact  for each observable variable   is computed as follow:

...

Example

Let's take the Perfume example for which we have defined five segments with the Breakout Variable Product, namely Prod3, Prod4, ProdG1, ProdG5 and Prod G6.

Below is the table with the Standardized Total Effects of all the observable variables on Purchase Intent, and the mean values of these variables, computed on the entire data set, and on the segment represented by Prod3.

Image Removed

These figures allows computing the Impacts of each variable.

Image Added

Prod3

Null Value Assessment

This option allows estimating if the segment and benchmark mean values of each observable variable are significantly different. Two tests are proposed for answering this question:

  • a Frequentist one, NHST t-test, the Null Hypothesis Significance Testing with the Welch's two-sample, two tailed t-test, and 
  • a Bayesian one, BEST, described in the paper by John K. Kruschke, "Bayesian Estimation Supersedes the t-test", Journal of Experimental Psychology: General, 2013.


Info

Below is the Bayesian network used in the BEST approach. We are assuming that the samples follow a Student's t-distribution. The segment and the benchmark have their own Image Added and Image Added, but they share the same Image Added.

Image Added



Info

The default Confidence Level has been set to 95%. This is the same for the both tests.

As for the Bayesian test, the Region of Practical Equivalence (ROPE) on the effect size around the null value has been set to [-0.1, 0.1]. The null value is declared to be rejected if the 95% HDI falls completely outside the ROPE.



Info

You can use Window | Preferences | Tools | Statistical Tools to modify the confidence level, as well as the Monte Carlo Marco Chain parameters that are used for inference in the Bayesian test.

Image Added