Niklas Nebel wrote:
7th digit? You're not counting the digits behind the decimal point
only, are you?
I counted only the digits after the decimal point. But this is a test
suite that should give accurate results for a serious statistical
program. NIST lists the exact result and other packages do perform
accurate calculations.
R does indeed compute exact results for the mean. (not even the
slightest difference to the exact result)
SD is similarly computed to OOo (same error, which is acceptable,
5.587935e-10, given it is the SD).
I sorted again the data values ascending (see my previous posts), AND
indeed, Calc would become accurate, too. So, probably R just sorts this
data (it may however use more complex algorithms, too).
*Correlation Coefficient*
The correlation coefficient fails poorly:
- made column B = A+1
- C1: =CORREL(A1:A1001; B1:B1001)
- reported correl coeff. = -0.882
- should have been = -0.999
What's the "A+1" supposed to mean?
Anyway, the CORREL function implementation is straight forward and can
certainly be improved. If you want to try, go ahead.
I entered in column B(1): "=A1 + 1", that is I added the value "1" for
every element in column A. That said, data values in columns A and B
should be totally correlated. The NIST accepts a value of -0.999, but:
- Calc gives only a value of -0.882.
- R gives an absolute precise value of +1 (it reports positive values )
both with the cor(x,y) function and with the cor.test(x, y) function
Not even the slightest deviation from this value. This should be
definitely improved in Calc.
NOTE: sorting the values in Calc, breaks CORREL() [it gives an #VALUE
error, NO idea why]. Well, if I pair the non-ordered data in A with an
ordered (A+1), I see a very strange result: CORREL() gives -1.01, BUT
this coefficient CAN BE only between -1 and +1!!! SOME SERIOUS ERROR.
I will try to make more tests during the weekend.
Kind regards,
Leonard Mada
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
