Correlation Matrix of time series across categories

RL997 — Fri, 26 Jul 2019 20:59:38 GMT

I have a set of data organized like this:

ID Date Return

A 1/1/2001 .01

A 2/1/2001 .005

.. … …

A 12/1/2001 -.001

B 1/1/2001 .009

B 2/1/2001 .015

… … …

B 12/1/2001 -.005

C 1/1/2001 .007

C 2/1/2001 .005

.. … …

C 12/1/2001 .002

I would like my output to show a matrix of ‘Return’ correlations along with p-values

A B C

A 1.00 .778 .562

(p val) (p val) (p val)

B .778 1.00 .342

(p val) (p val) (p val)

C .562 .342 1.00

(p val) (p val) (p val)

I can’t figure out how to do this with proc corr. The code I tried was this:

proc corr data = ReturnsByID;
var return;
by id;
ods output pearsonCorr = RetCorrMat;
run;

Any suggestions?

Thanks.

Re: Correlation Matrix of time series across categories

Ksharp — Sat, 27 Jul 2019 11:23:48 GMT

Then Change your data structure to fit the PROC CORR .

data have;
input ID  $        Date      : mmddyy10.         Return;
format date mmddyy10.;
cards;
A            1/1/2001            .01
A            2/1/2001            .005
A           12/1/2001          -.001
B            1/1/2001            .009
B            2/1/2001            .015
B            12/1/2001          -.005
C            1/1/2001            .007
C            2/1/2001            .005
C            12/1/2001          .002
;
proc sort data=have out=temp;
by  date;
run;
proc transpose data=temp out=want;
by  date;
var return;
id id;
run;
proc corr data=want;
var a b c;
run;

topic Correlation Matrix of time series across categories in Statistical Procedures

Correlation Matrix of time series across categories

Re: Correlation Matrix of time series across categories