On Wed, 3 Dec 2003 17:51:00 +1100, "miller" <[EMAIL PROTECTED]> wrote:
>The mean of the larger of the two is m + 0.564s. >The standard deviation is 0.826s So, those answer looks same to the results posted previously as >> [EQ-4] m(w) = m + s/Sqrt(Pi) = m + 0.564s, >> [EQ-5] s(w)^2 = (-1+Pi)*s^2/Pi = (0.826s). On the other hand, this new pdf is not exactly Gaussian distribution as you noticed, and, moreover, my test shows that the difference between them is very slight. What's your opinion in this feature? Best Regards, James K. ([EMAIL PROTECTED]) - Any remarks, proposal and/or indicator would be respected. - Private opinions: These are not the opinions from my affiliation. [Homepage] http://myhome.naver.com/txdiversity . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
