REGR_R2
The higher the correlation, the more accurate the set of data used for prediction. When the highest correlation (close to 1) is obtained, the predicted data is most accurate.
For example:
create table test(id int, num int);
insert into test values(1, 298),(2, 368),(3, 458),(4, 328),(5, 108),(6, 667),(7, 123),(8, 555);
Calculate the correlation between two sets of data.
select count(*),regr_r2(t.num,test.num) from
(select id-1 as id, num from test) as t,test where t.id=test.id ;
COUNT(*) | REGR_R2(T.NUM,TEST.NUM)
----------+--------------------------
28 | 0.6088521468006012
(1 row)
REGR_AVGX
regr_avgx(y, x) actually calculates the average of x (mathematical expectation), y has no role here.
Top comments (0)