Hello,
I do not understand but it looks as if the trend line is not correct. anybody else experiencing this ?
1 0
5 17
10 30
15 34
20 40
25 44
30 48
35 52
40 55
45 59
50 61
55 62
60 65
65 66
70 68
75 72
80 77
85 79
90 87
95 95
100 100
trendline 3th degre polynomial
y = 0,0002x3 - 0,0406x2 + 2,5611x + 2,9455 R2= 0.9927 ( should be very good fit )
However if I use this function to generate the data, it does not fit at all ( y is maximum 53 for a x of 100 )
Last edited by ghislain; 02-21-2007 at 09:50 AM.
Are you using the trendline coefficients exactly as you've entered them (only to the 4th decimal place)? If so, you probably just need to get all the decimal places to show. I performed the regression using the LINEST function and, for the X^3 coefficient for example, I got something closer to .00025 -- an error of ~25% in that coefficient. Either use the LINEST function (=LINEST(knownys,knownxs^{1,2,3}) or increase the number of decimals shown in the chart trendline so you can get precise coefficients.
You are absolutely right, I have taken 6 decimals and that has done it.
Thank you very much
There are currently 1 users browsing this thread. (0 members and 1 guests)
Posting Permissions
LinkBack URL
About LinkBacks
Register To Reply
Bookmarks