topic Clustering in SAS Studio

Clustering

NVN — Thu, 08 Jun 2017 13:41:02 GMT

I am trying to implement the K-means Clustering algorithm. I imported the data set(CREDIT) to sas , replaced the missing values (NA) with 0 of the variable that i need ( MonthlyIncome ) but variable have the character data type. But when i try to convert this variable to numeric to do the Scaling process with the following code the log is showing the following errors. please help me..

error 1: Found " Monthly_Income" when expecting " ;"

error2: Statement "Monthly_Income" is not valid;

Code:

Proc standerd data=CREDIT mean=0 std=1;

out=CREDIT1;

var age MonthlyIncome debtratio etc...............;

Var Monthly_Inc=(MonthlyIncome, best12.);

run;

Re: Clustering

stat_sas — Thu, 08 Jun 2017 14:07:00 GMT

Hi,

First convert variable in data step and then standardize.

Re: Clustering

NVN — Thu, 08 Jun 2017 14:43:13 GMT

Data step..? you mean..? at the time of importing the data set to sas.? or change the type directly.?

Re: Clustering

ballardw — Thu, 08 Jun 2017 15:14:03 GMT

@NVN wrote:

Data step..? you mean..? at the time of importing the data set to sas.? or change the type directly.?

Best would be to insure that the data was read into SAS as a numeric.

If that isn't practical for some reason you can use a data step to change:

data want;

set have (rename=(income=textincome));

income = input(textincome,best10.);

/* assign format or labels as desired*/

drop textincome;

run;

the rename is so you could use the same varaible name as you cannot change sas data types directly.

Re: Clustering

NVN — Thu, 08 Jun 2017 19:56:43 GMT

It is working.. Thank you soo much guys.