# HP 40g - Statistics Question 1

When I tried to find logistic fit for the data below, HP 39gs shows that there is a fit error with the reason stating that there is invalid statistics data. However, when I retry by changing the data in column C2 by multiplying them with 100, I was able to get the fit. I suspect that the logistic fit is not able to find the fit if the y-value is less than 1.

C1: 0 1 2 3 4 5 6 7 8 9 10

C2: .02 .04 .09 .18 .33 .54 .72 .86 .91 .96 .98

I’m guessing here since I don’t know what the internal methods used in calculating the logistic curve are. However, what I suspect is that the routine that does the calculation of the logistic curve calls, as part of its internal routine, the routine that is used for the exponential curve. Now that one also has problems with certain data - in this case with negatives. Try puting in a set of data that follows a roughly exponential curve, with one of the y values that are close to the x axis being negative instead of a very small positive. The result is an error message saying "Fit Erorr: Invalid Statistics Data LN(Neg)". The reason for this is because the internal exponential routine takes logs of the y values in order to linearise it before doing a best fit. Now if the Logistic fit is taking logs of the data as part of its working (which I would think is quite likely) then the result will be negative values for any data values 0

Logistic Model:

Applicable to HP 39g, HP 39g+, HP 39gs, HP 40gs

Updated On: 13.09.15