Download Cluster Analysis Using SPSS Start with an existing data file. and more Lecture notes Computational and Statistical Data Analysis in PDF only on Docsity!
Dr. Gaurav Bansal Bus Adm 216: Cluster Analysis Using SPSS Start with an existing data file.
marketvalues.sav
Dataset screen shot Next step: Click on analyze >> Classify >> Two Step Cluster
This will then open the following window. You can select both categorical and continuous variables for this analysis.
We selected House street as the categorical variable. We selected house price, square feet and days in market as continuous variables. After selecting the variables, click on the OUTPUT button.
Next step: In the following window click on the CREATE CLUSTER MEMBERSHIP variable
You can go to the data view there you will find a new variable added to your list of variables towards the end. This variable identifies the cluster membership of all the observations in your dataset.
Next: We can identify from the SPSS output that the cluster quality is good.
Next: Then click on Graphs and then select Chart Builder. Select Scatter / Dot plots. Then select GROUPED SCATTER plot.
Then select Sq Ft as x variable and Purchase price as y variable. Use Two Step Cluster Membership variable to SET COLOR (shown in the upper right corner of the graph window). You need to drag and drop the variables at the right place. Then hit ok.
In the above TWO STEP analysis we could choose both categorical and continuous variables and the algorithm automatically identifies the suitable number of clusters possible with the variables and the data set provided. However there is another approach as well. That is to first use the Hierarchical cluster modeling and examine the dendogram. The visual examination of the dendogram would then be helpful in identifying the correct number of clusters. You can then plot them using the scatter plot (as shown above). You need to be creative in selecting the right mix of x and y variables to demonstrate the clusters on your scatter plot. The hierarchical cluster modeling only allows use of continuous variables.
