Bioequivalence and Bioavailability Forum • Interlude II (simulations)

Interlude II (simulations) [RSABE / ABEL]

posted by PharmCat – Russia, 2020-08-20 00:26 (1338 d 22:12 ago) – Posting: # 21890
Views: 5,564

Hi Helmut!

using DataFrames, CSV, ReplicateBE, LinearAlgebra



path      = dirname(@__FILE__)

dfa       = CSV.File(path*"/sim500-2x3x3.csv") |> DataFrame

dfa.logpk = log.(dfa.PK)



res = Vector{Any}(undef, 500)

pd  = Vector{Any}(undef, 500)

os  = Vector{Any}(undef, 500)

for i in 1:500

    df = filter(r -> i == r.set , dfa)

    res[i] = ReplicateBE.rbe!(df, dvar = :logpk, subject = :subject, formulation = :treatment, period = :period, sequence = :sequence, g_tol = 1e-10, singlim = 1e-12)

    pd[i]  = isposdef(Symmetric(res[i].result.H))

    os[i]  = ReplicateBE.optstat(res[i])

    println(i)

end

println("Positive definite: ", sum(pd), "(",sum(pd)/500*100,"%)")

println("Converged: ", sum(os), "(",sum(os)/500*100,"%)")

Positive definite: 332(66.4%)
Converged: 500(100.0%)

I can make table with results if necessary.

Complete thread:

Partial replicate design: reference(s)? Helmut 2020-08-13 14:22
- Partial replicate design: reference(s)? ElMaestro 2020-08-13 15:23
  - Partial replicate design: reference(s)? Helmut 2020-08-13 17:03
- Hyslop's Alternative Cross-over Designs for Individual Bioequivalence mittyri 2020-08-13 16:21
  - Terry’s homebrew Helmut 2020-08-13 17:08
    - impressive homebrew mittyri 2020-08-13 18:09
      - impressive indeed Helmut 2020-08-14 13:21
- Partial replicate design: reference(s)? zizou 2020-08-13 23:24
  - Donald’s model Helmut 2020-08-14 11:43
    - Donald’s model - model for (logistic) groups d_labes 2020-08-14 14:03
      - Donald’s model - model for (logistic) groups Helmut 2020-08-14 14:06
- Interlude I (sample sizes, problems & remedies) Helmut 2020-08-14 19:37
- Interlude II (simulations) Helmut 2020-08-19 21:31
  - Interlude II (simulations)PharmCat 2020-08-19 22:26