topic Re: quantile Regression in Statistical Procedures

quantile Regression

thirupathireddy — Mon, 31 May 2021 04:46:49 GMT

I applied the Quantile regression technique to 10000 samples, and for the selected quantiles(5,10,50,75,90) estimated the coefficients, but I want to know the sample size for each quantile.

my question is

at 5% what is the sample size

at 10% what is the sample size

at 90% what is the sample size

Re: quantile Regression

PGStats — Mon, 31 May 2021 17:26:16 GMT

Do you want something like this ?

proc quantreg data=sashelp.heart;
where sex = "Male";
model weight = height / quantile=0.1 0.5 0.9;
output out=q residual=r / columnwise;
run;

proc sql;
select 
    quantile,
    sum(r>0) as n_greater,
    sum(r<0) as n_lower
from q
where sex = "Male"
group by quantile;
quit;

Re: quantile Regression

thirupathireddy — Mon, 31 May 2021 19:39:30 GMT

For the Sashelp. hearts after applying the quantile regression the total sample size is 2336

in that total sample size how many samples are belong to the First 0.1, 0.5, and 0.9 quantiles.

below output is showing

The FREQ Procedure

But I want output like below

Re: quantile Regression

PGStats — Tue, 01 Jun 2021 03:11:44 GMT

Maybe you simply need:

proc reg data=sashelp.heart;
where sex = "Male";
model weight = height;
output out=q residual=r;
run;

proc rank data=q out=qr groups=10;
var r;
ranks quantile;
run;

proc freq data=qr;
table quantile;
run;