Go back to Flu Mortality
Result for the “SANN” method for the 1989-1900 epidemic:
Weston: The “SANN” method gave us some negative parameter values and a very poor result for the maxit = 300 (the sum doesn't look very good and of course the two models are entirely wrong).
For the guess of parameters beta1 = 0.3 alpha1 = 0.2 Ne = 1377 I1 = 1 beta2 = 0.3 alpha2 = 0.2 I2 = 1, we obtained
beta1 = -2.528945 alpha1 = 2.966377 Ne = 1379.536694 I1 = -1.366907 beta2 = 1.325298 alpha2 = 0.656774 I2 = 2.727558
So, this implies that f = 1379.536694/27653146 = 0.000049887
The covariance matrix was
[,1] [,2] [,3] [,4] [,5]
[1,] 883.59355 -2024.14285 21828.1736 845.71544 23.8237469 [2,] -2024.14285 4609.65943 -49987.2850 -1909.68368 -54.5543020 [3,] 21828.17363 -49987.28498 561433.5747 20885.42621 612.6809626 [4,] 845.71544 -1909.68368 20885.4262 796.68262 22.7940561 [5,] 23.82375 -54.55430 612.6810 22.79406 0.6697507 [6,] 26.81125 -61.39638 689.5311 25.65257 0.7540505 [7,] -11.57346 26.48650 -297.2874 -11.07079 -0.3198596
[,6] [,7]
[1,] 26.8112484 -11.5734587 [2,] -61.3963776 26.4865032 [3,] 689.5310942 -297.2874061 [4,] 25.6525660 -11.0707934 [5,] 0.7540505 -0.3198596 [6,] 0.8489429 -0.3626797 [7,] -0.3626797 0.1072153
The MLE was 3939.494
Here’s a plot of the model with the data (red is the data, blue is the sum of the two models, green is the lamb2 model [which can't be seen], and purple is the lamb1 model)