I am trying to get the bootstrap for a fitted pumas model ( the minimization was successful and the infer function ran successfully as well). The model is as follows:
pac_carb_ANC = @model begin
@param begin
tvPOP_CIRC0 ∈ RealDomain(lower=0.00)
tvPOP_MTT ∈ RealDomain(lower=0.00)
tvPOP_GAMMA ∈ RealDomain(lower=0.00)
tvPOP_SLOP ∈ RealDomain(lower=0.00)
tvPOP_SLOPC ∈ RealDomain(lower=0.000)
Ω_1 ∈ PDiagDomain(3)
Ω_2 ∈ PDiagDomain(1)
σ_prop ∈ RealDomain(lower=0.00)
end
@random begin
η_1 ~ MvNormal(Ω_1)
η_2 ~ MvNormal(Ω_2)
end
@covariates BSA SEX BILIRUBIN AGE Regimen # BSA (/m2) , SEX 0 Female 1 male Bilirubin (micromole), AGE (years)
@pre begin
Vc = 12.8
Vp = 252 * (BSA/1.8)^1.17
VMEL = 37.4 * 1.2^SEX * (BSA/1.8)^0.842 * (BILIRUBIN/7)^(-0.167) * (AGE/56)^(-0.352)
KMEL = 0.53
VMTR = 169 * 1.2^SEX * (BSA/1.8)^0.911
KMTR = 0.83 * 2.11^SEX
K21 = 1.15 * 0.893^SEX
Q = 20.1 * (BSA/1.8)^0.724
cl = 7.38
q = 90.5
vc = 11.9
vp = 8.23
CIRC0 = tvPOP_CIRC0 * exp(η_1[1])
MTT = tvPOP_MTT * exp(η_1[2])
SLOP = tvPOP_SLOP * exp(η_1[3])
SLOPC = tvPOP_SLOPC * exp(η_2[1])
GAMMA = tvPOP_GAMMA
NN = 3
KTR = ((NN+1)/MTT)
end
@init begin
Prol = CIRC0
Transit1 = CIRC0
Transit2 = CIRC0
Transit3 = CIRC0
Circ = CIRC0
end
@vars begin
CONC = (Central/Vc)
conc = (central_carb/vc)
EDRUG = (1-(SLOP*CONC))
EDRUGC = (1-(SLOPC*conc))
FEED = (abs(CIRC0/Circ)^GAMMA)
end
@dynamics begin
Central' = -((CONC*VMEL)/(CONC + KMEL)) + (K21*PERIPH1) - ((CONC*VMTR)/(CONC + KMTR)) + ((Q/Vp)*PERIPH2) - ((Q/Vc)*Central)
PERIPH1' = -(K21*PERIPH1) + ((CONC*VMTR)/(CONC + KMTR))
PERIPH2' = - ((Q/Vp)*PERIPH2) + ((Q/Vc)*Central)
#CARB' = -KC*CARB
central_carb' = -(cl/vc)*central_carb + (q/vp)*Periph_carb - (q/vc)*central_carb
Periph_carb' = (q/vc)*central_carb - (q/vp)*Periph_carb
# PD
Prol' = ((((KTR*Prol)*EDRUG*FEED*EDRUGC))-(KTR*Prol))
Transit1' = ((KTR*Prol)-(KTR*Transit1))
Transit2' = ((KTR*Transit1)-(KTR*Transit2))
Transit3' = ((KTR*Transit2)-(KTR*Transit3))
Circ' = ((KTR*Transit3)-(KTR*Circ))
end
@derived begin
CONC = @. (Central/Vc)
conc = (central_carb/vc)
E = @. Circ
ANC = @. Normal(E, E*σ_prop)
end
end
However the bootstrap runs for very long time and it keeps giving the following error:
bs= infer(fit_pacli_carb_ANC, Pumas.Bootstrap(rng=MersenneTwister(1234),
samples=200, ensemblealg = EnsembleThreads()))
┌ Warning: Interrupted. Larger maxiters is needed.
└ @ SciMLBase /builds/PumasAI/PumasSystemImages-jl/.julia/packages/SciMLBase/mndcy/src/integrator_interface.jl:331
^C
┌ Warning: Interrupted. Larger maxiters is needed.
└ @ SciMLBase /builds/PumasAI/PumasSystemImages-jl/.julia/packages/SciMLBase/mndcy/src/integrator_interface.jl:331
^C
Could you please let me know what does this warning message mean and if there is a way to make it run faster ?
Thanks
vijay
November 17, 2022, 1:02am
2
It looks like some of the resampled datasets have issues. Do you get these warnings during the fitting of the original model
Actually the fitting of the original model runs very smooth with no warnings. here is the output
Iter Function value Gradient norm
0 5.080082e+03 2.657559e+01
* time: 0.02542591094970703
1 5.079981e+03 2.004248e+01
* time: 149.61270904541016
2 5.079961e+03 1.497125e+01
* time: 293.2913498878479
3 5.079924e+03 1.386949e+01
* time: 404.4286630153656
4 5.079919e+03 1.275754e+01
* time: 519.6313679218292
5 5.079840e+03 4.859565e+00
* time: 631.5725409984589
6 5.079835e+03 3.599118e+00
* time: 770.0869359970093
7 5.079796e+03 1.848765e+00
* time: 882.0380258560181
8 5.079784e+03 1.772230e+00
* time: 986.7223229408264
9 5.079773e+03 7.275160e-01
* time: 1067.3959460258484
10 5.079765e+03 8.293510e-01
* time: 1146.2089738845825
11 5.079762e+03 1.422620e-01
* time: 1223.3514399528503
12 5.079762e+03 2.062762e-02
* time: 1297.7948999404907
13 5.079762e+03 2.165612e-03
* time: 1366.6582679748535
14 5.079762e+03 1.802472e-03
* time: 1433.9624688625336
15 5.079762e+03 3.306187e-03
* time: 1499.6750829219818
16 5.079762e+03 3.313480e-03
* time: 1578.7967178821564
17 5.079762e+03 3.667299e-03
* time: 1661.0921380519867
18 5.079762e+03 3.121457e-03
* time: 1731.8161239624023
19 5.079762e+03 3.241176e-03
* time: 1805.8876328468323
20 5.079762e+03 3.255183e-03
* time: 1882.551826953888
21 5.079762e+03 3.278371e-03
* time: 2026.176521062851
22 5.079762e+03 3.269786e-03
* time: 2173.9755730628967
23 5.079762e+03 3.274075e-03
* time: 2320.8510229587555
24 5.079762e+03 3.277968e-03
* time: 2469.7986998558044
25 5.079762e+03 3.278351e-03
* time: 2620.940274953842
26 5.079762e+03 3.282005e-03
* time: 2767.507360935211
27 5.079762e+03 3.285948e-03
* time: 2915.770865917206
28 5.079762e+03 3.290329e-03
* time: 3069.461163043976
29 5.079762e+03 3.286230e-03
* time: 3169.5120148658752
30 5.079762e+03 3.286927e-03
* time: 3299.151221036911
31 5.079762e+03 3.287002e-03
* time: 3405.686693906784
32 5.079762e+03 3.287002e-03
* time: 3466.962636947632
FittedPumasModel
Successful minimization: true
Likelihood approximation: Pumas.FOCE
Log-likelihood value: -5079.7618
Number of subjects: 405
Number of parameters: Fixed Optimized
0 10
Observation records: Active Missing
ANC: 2274 0
Total: 2274 0
---------------------------
Estimate
---------------------------
tvPOP_CIRC0 6.2787
tvPOP_MTT 143.71
tvPOP_GAMMA 0.19585
tvPOP_SLOP 5.8194
tvPOP_SLOPC 0.0058817
Ω_1₁,₁ 0.23317
Ω_1₂,₂ 0.031576
Ω_1₃,₃ 0.11131
Ω_2₁,₁ 3.3201
σ_prop 0.36922
---------------------------
And this is the output of infer function:
symptotic inference results using sandwich estimator
Successful minimization: true
Likelihood approximation: Pumas.FOCE
Log-likelihood value: -5079.7618
Number of subjects: 405
Number of parameters: Fixed Optimized
0 10
Observation records: Active Missing
ANC: 2274 0
Total: 2274 0
---------------------------------------------------------------------------
Estimate SE 95.0% C.I.
---------------------------------------------------------------------------
tvPOP_CIRC0 6.2787 0.18622 [ 5.9137 ; 6.6437 ]
tvPOP_MTT 143.71 4.2255 [135.43 ; 151.99 ]
tvPOP_GAMMA 0.19585 0.011701 [ 0.17291 ; 0.21878 ]
tvPOP_SLOP 5.8194 0.2963 [ 5.2386 ; 6.4001 ]
tvPOP_SLOPC 0.0058817 0.0020094 [ 0.0019434; 0.00982 ]
Ω_1₁,₁ 0.23317 0.024729 [ 0.1847 ; 0.28164 ]
Ω_1₂,₂ 0.031576 0.0083073 [ 0.015294 ; 0.047858]
Ω_1₃,₃ 0.11131 0.022134 [ 0.067931 ; 0.15469 ]
Ω_2₁,₁ 3.3201 0.9243 [ 1.5085 ; 5.1317 ]
σ_prop 0.36922 0.01312 [ 0.34351 ; 0.39494 ]
---------------------------------------------------------------------------