topic How to compute Mahalanobis distance in SAS in SAS Programming

How to compute Mahalanobis distance in SAS

kbhagat — Tue, 14 Nov 2023 18:46:29 GMT

Hello Everyone,

Can anyone help me to find out how to calculate Mahalanobis distance with two samples?

Re: How to compute Mahalanobis distance in SAS

ballardw — Tue, 14 Nov 2023 19:53:35 GMT

What do you mean by "two samples"? Two different data sets? One data set and two different variables? One data set, one variable with two groups indicated by a different variable?

Perhaps you should include a small sample of the data that you currently have as working data step code so we don't have to guess.

Re: How to compute Mahalanobis distance in SAS

kbhagat — Tue, 14 Nov 2023 20:50:06 GMT

Thank you for your response. I have two datasets: CIREN and CISS. We are trying to find out how similar CIREN data is to CISS. CISS is a random sample of the whole country's database. CIREN is a particular kind of crash. The variables we are using are Age, weight, gender, height, ISS score, and AIS score. I know how to calculate the Mahalanobis distance in one dataset. I am trying to find out what is going to be the equation if we have two datasets.

Re: How to compute Mahalanobis distance in SAS

sbxkoenk — Tue, 14 Nov 2023 21:04:25 GMT

You need to append both datasets CIREN and CISS.

Do not forget to make an extra column (named "SourceDS" $15) with two distinct values : "row from CIREN" and "row from CISS".

Then do a canonical discriminant analysis to profile CIREN versus CISS (as if you would analyze heterogeneity between clusters and homogeneity within clusters).

proc candisc data=MyData out=outcan distance anova;
   class SourceDS;
   var Age weight gender height ISS_score and AIS_score;
run;

BR, Koen

Re: How to compute Mahalanobis distance in SAS

sbxkoenk — Tue, 14 Nov 2023 21:07:50 GMT

For Mahalanobis distance, see here :

What is Mahalanobis distance?
By Rick Wicklin on The DO Loop February 15, 2012
https://blogs.sas.com/content/iml/2012/02/15/what-is-mahalanobis-distance.html
How to use Mahalanobis distance to find outliers in multivariate data?
Detecting outliers in SAS: Part 3: Multivariate location and scatter
By Rick Wicklin on The DO Loop February 2, 2012
https://blogs.sas.com/content/iml/2012/02/02/detecting-outliers-in-sas-part-3-multivariate-location-and-scatter.html
How to compute Mahalanobis distance in SAS
By Rick Wicklin on The DO Loop February 22, 2012
https://blogs.sas.com/content/iml/2012/02/22/how-to-compute-mahalanobis-distance-in-sas.html

Koen

Re: How to compute Mahalanobis distance in SAS

kbhagat — Tue, 14 Nov 2023 21:34:28 GMT

Thank you so much. I have to try this method. I will get back to you if I have follow-up question

Re: How to compute Mahalanobis distance in SAS

kbhagat — Tue, 14 Nov 2023 21:35:06 GMT

Thank you so much