Fitting with Poisson random variables1 Recall that the normal fitting procedure is to choose parameters of a fitting function F(xi;c) that will minimize the chi-square: N 2 n i F xi ; c 2 (46) F xi ; c The probability of seeing an entire Poisson data set is i 1 N P ni ; F xi ; c e Define F xi ;c ni ! i 1 2 Poisson 2ln P ni ; F xi ; c F xi ; c ni (50) (51) So that maximizing Eq. (50) is equivalent to minimizing N 2 Poisson 2 F xi ; c ni ln F xi ; c ln ni ! (52) i 1 The term with ni! is independent of the fitting constants and can thus be dropped. Thus minimizing 2Poisson is equivalent to minimizing N P2 ,m 2 F xi ; c ni ln F xi ; c (53) i 1 Bob Coldwell speculation N F x ; c P2 ,m i ni 2 (54) 1 0 cm cm F xi ; c i 1 This is slightly different from minimizing the usual 2 in Eq. (46) 2 2 N F x ; c n F x ; c ni F xi ; c i i i 0 2 (55) cm cm F xi ; c F 2 xi ; c i 1 Simplify, but only a bit 2 2 N F x ; c ni F xi ; c i ni 0 2 1 (56) cm cm F 2 xi ; c F xi ; c i 1 The second term coming from the derivative of the error in Eq (46) is wrong. – Careful – it lowers the chi-square of (46) by maximizing i in N 2 i 1 n i F xi ; c i2 2 (57) This is heavily based on Bob Deserio’s same named section in Deserio’s UFlab Poisson Statistics The use of F(x,c) goes beyond the original. 1 Notice, that Eq (57) with i independent of c – but somehow equal to F does minimize for the correct value of F. This says that the current method for a Poisson situation requires. 1. Determine a best estimate for (xi). 2. With (xi) fixed, determine F(xi,c) by minimizing Eq. (57) 3. Re-determine (xi) iterate. This procedure will minimize Eq (57) with de-facto i2 F xi , c , but at the end Eq (56) will be 2 N F x ; c i ni F xi ; c 2 cm cm F 2 xi ; c i 1 2 (58) In the case that F = , this term is 2 2 N ni 2 (59) 2 2 n2 2 2 i 1 This was shown in Derivation of the Poisson distribution.doc#SigmaP to be equal to implying that 2 2 (60) So that 2 2 2 (61) 2 Using 2 2 N ni 2 2 2 2 i 1 2 0 2 m m n 2 2 2 2 2 2 2 2 2 2 2 (62) (63) 1 2 or m 1 This is not quite in agreement with Bob Deserio’s results of Poisson = 2.9 Gauss = 2 Compare predicted 2.6 from the above.