How to Do a Mixed-Mode Tree Analysis

Introduction

This article describes how to perform a mixed-mode tree analysis in Q. A Mixed-Mode Tree can be created which predicts the Questions to analyze using predictor questions.

The image below shows a tree which predicts main phone company by age, education and gender.

Requirements

• A categorical variable to be used as Target data.
• At least one variable to select as a Predictor variable. 

Method

1. The first step depends on which version of Q you are using:
   • In Displayr, select Anything > Advanced Analysis > Machine Learning > Mixed-Mode Tree.
   • In Q Create > Classifier > Mixed-Mode Trees
   • Older versions of Q:
     • Create and Segments
     • Select splitting by questions (tree)
2. Select the questions to be used to form the segments in the Questions to analyze dialog box.
3. Select the predictor variables:
   • Select the variables into the lower box (that is, the box that is below the box for Questions to analyze.
4. If necessary, modify the default options. Note that:

• By default Q will automatically select the number of segments using the Bayesian information criterion. You can alternatively specify a specific number of segments by selecting the Manual option. Alternatively, you can select a different information criteria by clicking Advanced.
• The Question Type of the questions that are analyzed determines how the latent class model is conducted. For example, when analyzing a Pick One - Multi any scale points are ignored and Q treats the data as categorical; if it is converted to a Ranking Q focuses on understanding relativities; if it is converted to Number - Multi Q treats the data as being numeric.

Next

How to Perform a Mixed-Mode Cluster Analysis

How to Interpret Trees