segmentation using fastclus

[deega]

Hi,

I am trying to do segmentation using fastclus procedure but most of the data points fall into just one cluster. My input is a utility matrix that I created by customer purchase history and customer's demographic data. Could anybody please tell me whats wrong...

My input looks as follows

S.No	PromotionID	F	R	Total_P	return	A1	A2	A3	A4	A5	A6	A7	A8	A9	A10	A11	A12	A13	A14	A15	A16	A17	A18	A19	A20	Other_A	M1	M2	M3	M4	Other_M	O1	O2	O3	O4	O5	O6	O7	O8	O9	O10	Region	Age	Gender
104	-0.2383	1	8	1	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	1	0	0	0	0	0	-0.95472	0.383248	1
104	-0.2383	1	8	2	0	0	0	0	2	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	2	0	0	0	0	0	0	0	2	0	0	0	0	0	0	-1.11397	0.995225	1
104	-0.2383	1	7	1	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0.080419	-0.22873	2
104	-0.2383	1	8	2	0	0	0	0	0	0	0	2	0	0	0	0	0	0	0	0	0	0	0	0	0	0	2	0	0	0	0	0	0	0	0	2	0	0	0	0	0	-0.71584	-0.22873	1
104	2.136383	1	8	1	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	1	0	0	0	0	0	0	0	0	-0.63622	-0.22873	2
104	-0.2383	1	5	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0.637801	0.383248	2
104	-0.2383	1	8	1	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0.080419	-0.22873	1
104	-0.2383	1	8	3	0	0	0	0	0	0	0	0	3	0	0	0	0	0	0	0	0	0	0	0	0	0	0	3	0	0	0	0	0	0	0	3	0	0	0	0	0	0.239671	0.995225	2
104	-0.2383	1	8	2	0	0	0	0	0	0	0	0	0	2	0	0	0	0	0	0	0	0	0	0	0	0	0	2	0	0	0	0	0	0	0	2	0	0	0	0	0	0.558175	0.383248	1
104	-0.2383	2	8	3	0	0	0	2	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	2	1	0	0	0	0	0	0	0	2	0	0	0	0	1	1.274809	-0.22873	2
104	-0.2383	1	8	1	2	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	1	-0.63622	-0.84071	2
104	-0.2383	1	3	1	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	1	0	0	0	0	0	1.513687	-0.84071	1
104	-0.2383	1	8	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	1	0	0	0	0	0	-0.63622	-0.84071	1
104	-0.2383	1	3	1	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	1	0	0.319297	0.383248	2

[Ksharp]

I saw your data is category data not continuous data.

Firstly use PROC DISTANCE to get these category data's distance, then run PROC CLUSTER.

Search the following keyword at support.sas.com

categroy data cluster analysis

[deega]

@Ksharp
I might sound weird but is there a way to convert categorical data to continuous data ?

[Ksharp]

No way .

[Ksharp]

Or you could take a look this:

Overview: PRINQUAL Procedure
The PRINQUAL procedure performs principal component analysis (PCA) of qualitative, quantitative, or
mixed data. PROC PRINQUAL is based on the work of Kruskal and Shepard (1974); Young, Takane, and

polt two primary component at X-Y asix . and see which one belong to a cluster.
 

[deega]

Before Fastclus I tried distance and cluster but since I have large dataset (10000000 records) and I got error that it can not be used on such large data. Is there any other way ?

[Rick_SAS]

What code are you using for PROC FASTCLUS?

[deega]

Here is my code

proc fastclus data=std out=clus maxclusters=10;
var x--y;
run;

My variables are mix of categorical and continuous.

[Ksharp]

Cluster Analysis can't be apply to mixed data(categorical and continuous).

[Babloo]

Can we do cluster analysis only with continous varaibles?

[deega]

OK. If I convert my data to categorical as most of it is categorical then how can I cluster them, given that its large dataset.