Tuesday, May 27, 2008

TIPS : an Intro to CHAID Analysis

CHAID Analysis : Cereal Case*

Below, step by step we give tips & tricks how to make CHAID Analysis with SPSS Software.

Firstly, from SPSS menu, click :
Analyse  Classify  Tree

Then, input ‘Preferred Breakfast’ into Dependent Variabel coloumn, and else variables to Independent Variables.

Click Ok.

Then Click Define Variables Properties at Classification Tree window
Click All Target Category and input to Exclude Coloumn. Click Continue.

At Classification Tree : Output, give checklist at ‘Tree sub menu’, ‘Topdown’ (in Orientation ), ‘Tables & Chart’ (in Node Contents), ‘Automatics’ (in Scale) and ‘Independent variables statistics’, ‘Node Definition’, then click Continue.

At Statistics submenu, give checklist at all item at Model & Node Performancesubmenu, then click continue.

We have result as below:

Model Summary

Specifications Growing Method CHAID
Dependent Variable Preferred breakfast
Independent Variables Age category, Gender, Marital status, Lifestyle
Validation NONE
Maximum Tree Depth 3
Minimum Cases in Parent Node 100
Minimum Cases in Child Node 50
Results Independent Variables Included Age category, Marital status, Lifestyle
Number of Nodes 13
Number of Terminal Nodes 8
Depth 2


Misclassification Costs

Observed Predicted
Breakfast Bar Oatmeal Cereal
Breakfast Bar .000 1.000 1.000
Oatmeal 1.000 .000 1.000
Cereal 1.000 1.000 .000
Dependent Variable: Preferred breakfast


For Cereal. Look at big Chart i.e node 5,8,9 with N=3, 12, 31, with Sum=139. look at table Classification below.

Classification

Observed Predicted
Breakfast Bar Oatmeal Cereal Percent Correct
Breakfast Bar 112 34 85 48.5%
Oatmeal 13 251 46 81.0%
Cereal 84 116 139 41.0%
Overall Percentage 23.8% 45.6% 30.7% 57.0%
Growing Method: CHAID
Dependent Variable: Preferred breakfast

(*) Data Cereal taken from SPSS

seja o primeiro a comentar!

Post a Comment

Edited By JuraganTAHU Design by Usuário ^