This vignette demonstrates how to track network statistics of interest across simulations using EpiModelMpox.
We fit a layered network of size 100,000 (representing main, casual, and instantaneous partnerships among men and women). Here, we run network diagnostics using the EpiModel package to demonstrate the fit of networks to the target statistics (drawn from behavioral parameters estimated from the National Survey of Family Growth). For main and casual networks we use dynamic diagnostics, and for the one-time network we use a static diagnostic.
Main Network (Marriages & Cohabitations)
main_dynamic <- netdx(
nw_100000[[1]],
dynamic = TRUE, nsims = 10, nsteps = 500, ncores = 2
)
Network Diagnostics
-----------------------
- Simulating 10 networks
- Calculating formation statistics
main_dynamicEpiModel Network Diagnostics
=======================
Diagnostic Method: Dynamic
Simulations: 10
Time Steps per Sim: 500
Formation Diagnostics
-----------------------
Target Sim Mean Pct Diff Sim SE
edges 27473.090 27389.956 -0.303 13.933
nodematch.race.B 2169.121 2105.852 -2.917 3.246
nodematch.race.H 4089.055 4147.292 1.424 5.059
nodematch.race.W 14397.240 14318.004 -0.550 7.193
nodefactor.race.B 5516.648 5538.930 0.404 10.555
nodefactor.race.H 10354.139 10267.183 -0.840 8.984
nodefactor.race.W 33094.490 33095.073 0.002 23.850
nodecov.age 1967244.828 1959133.841 -0.412 770.509
nodecov.agesq 74129346.357 73732355.734 -0.536 23556.868
absdiff.sqrt(age_adj) 7076.843 6956.135 -1.706 4.517
nodefactor.deg_casual>0.TRUE 317.420 301.644 -4.970 2.075
offset(nodefactor.floor(age).15) NA 0.000 NA NaN
offset(nodefactor.floor(age).16) NA 0.000 NA NaN
Z Score SD(Sim Means) SD(Statistic)
edges -5.967 51.106 72.694
nodematch.race.B -19.491 10.264 16.574
nodematch.race.H 11.511 17.453 26.869
nodematch.race.W -11.016 36.828 44.730
nodefactor.race.B 2.111 14.739 49.065
nodefactor.race.H -9.679 43.602 50.958
nodefactor.race.W 0.024 71.419 113.036
nodecov.age -10.527 3071.695 4566.189
nodecov.agesq -16.852 110053.968 158105.474
absdiff.sqrt(age_adj) -26.720 26.209 28.131
nodefactor.deg_casual>0.TRUE -7.602 11.708 11.940
offset(nodefactor.floor(age).15) NA 0.000 0.000
offset(nodefactor.floor(age).16) NA 0.000 0.000
Duration Diagnostics
-----------------------
Target Sim Mean Pct Diff Sim SE Z Score SD(Sim Means) SD(Statistic)
edges 2835.426 2831.86 -0.126 2.285 -1.561 12.126 13.06
Dissolution Diagnostics
-----------------------
Target Sim Mean Pct Diff Sim SE Z Score SD(Sim Means) SD(Statistic)
edges 0 0 -0.196 0 -0.441 0 0Casual Network
casual_dynamic <- netdx(
nw_100000[[2]],
dynamic = TRUE, nsims = 10, nsteps = 500, ncores = 2
)
Network Diagnostics
-----------------------
- Simulating 10 networks
- Calculating formation statistics
casual_dynamicEpiModel Network Diagnostics
=======================
Diagnostic Method: Dynamic
Simulations: 10
Time Steps per Sim: 500
Formation Diagnostics
-----------------------
Target Sim Mean Pct Diff Sim SE Z Score SD(Sim Means)
edges 8102.937 7778.366 -4.006 22.476 -14.441 41.935
nodefactor.age_group.1 3677.577 3610.092 -1.835 8.406 -8.028 38.743
nodefactor.age_group.2 3366.224 3223.147 -4.250 10.485 -13.646 32.991
nodefactor.age_group.3 2708.317 2547.147 -5.951 13.001 -12.397 39.646
nodefactor.age_group.4 1788.003 1660.740 -7.118 5.424 -23.464 14.127
nodefactor.age_group.5 1751.400 1655.908 -5.452 8.472 -11.272 26.869
nodefactor.age_group.6 1533.364 1457.137 -4.971 5.954 -12.803 15.846
nodefactor.race.B 2431.117 2377.285 -2.214 9.070 -5.935 35.853
nodefactor.race.H 2836.619 2701.456 -4.765 10.055 -13.442 20.880
nodefactor.race.W 8680.129 8404.846 -3.171 18.575 -14.820 42.132
nodematch.race.B 841.279 849.468 0.973 3.020 2.711 13.630
nodematch.race.O 380.772 344.914 -9.417 2.371 -15.126 12.682
nodematch.race.W 3446.499 3402.777 -1.269 5.469 -7.995 18.687
concurrent 607.770 585.561 -3.654 2.892 -7.679 15.190
nodefactor.deg_main.1 317.420 324.550 2.246 2.209 3.228 12.063
SD(Statistic)
edges 125.114
nodefactor.age_group.1 59.293
nodefactor.age_group.2 65.333
nodefactor.age_group.3 71.954
nodefactor.age_group.4 34.029
nodefactor.age_group.5 51.390
nodefactor.age_group.6 37.567
nodefactor.race.B 58.583
nodefactor.race.H 59.716
nodefactor.race.W 121.807
nodematch.race.B 22.039
nodematch.race.O 17.788
nodematch.race.W 42.012
concurrent 22.250
nodefactor.deg_main.1 15.862
Duration Diagnostics
-----------------------
Target Sim Mean Pct Diff Sim SE Z Score SD(Sim Means) SD(Statistic)
edges 217.236 222.25 2.308 0.382 13.127 0.906 2.302
Dissolution Diagnostics
-----------------------
Target Sim Mean Pct Diff Sim SE Z Score SD(Sim Means) SD(Statistic)
edges 0.005 0.005 -0.115 0 -0.51 0 0.001One-Time Network
inst_dynamic <- netdx(nw_100000[[3]], dynamic = FALSE, nsims = 1000)
Network Diagnostics
-----------------------
- Simulating 1000 networks
- Calculating formation statistics
inst_dynamicEpiModel Network Diagnostics
=======================
Diagnostic Method: Static
Simulations: 1000
Formation Diagnostics
-----------------------
Target Sim Mean Pct Diff Sim SE Z Score SD(Sim Means)
edges 14.853 15.388 3.599 0.123 4.343 NA
nodefactor.race.B 5.799 6.079 4.834 0.094 2.968 NA
nodefactor.race.H 4.649 4.716 1.447 0.077 0.878 NA
nodefactor.race.W 17.501 18.052 3.151 0.171 3.229 NA
nodefactor.age_group.1 5.945 6.019 1.241 0.108 0.681 NA
nodefactor.age_group.2 6.766 6.845 1.170 0.108 0.732 NA
nodefactor.age_group.3 5.821 6.110 4.961 0.098 2.951 NA
absdiff.sqrt(age_adj) 2.489 2.619 5.197 0.030 4.369 NA
SD(Statistic)
edges 3.892
nodefactor.race.B 2.636
nodefactor.race.H 2.250
nodefactor.race.W 5.401
nodefactor.age_group.1 3.159
nodefactor.age_group.2 3.144
nodefactor.age_group.3 3.094
absdiff.sqrt(age_adj) 0.937