Newer
Older
using StatsBase, Distributions
using AbstractMCMC: MCMCSerial #, MCMCThreads, MCMCDistributed
using CairoMakie
using Revise
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
function analyse(chains)
t = TruncatedNormal(0, 1, -1, 1)
println(" Target: mean = ", mean(t), ", std = ", std(t))
accept = [ getindex.(c, 1) for c in chains ]
states = [ getindex.(c, 2) for c in chains ]
rejection_rate = 1 .- mean.(accept)
println(" Rejection rate: ", mean(rejection_rate) * 100, " ± ", std(rejection_rate) * 100, " % ")
m = mean.( states )
s = std.( states )
println(" Mean ", mean(m), " ± ", std(m) )
println(" Std ", mean(s), " ± ", std(s) )
println("\n")
fig = Figure(); ax = Axis(fig[1,1])
for s in states[1:1]
hist!(ax, s, normalization=:pdf)
end
x = [-1:.01:1 ... ]; y = pdf.(t, x)
lines!(ax, x, y, label="N(0,1)")
display(fig);
end
begin # Vanilla MetropolisHastings
w = CyclicWalk(-1, 1, .5)
s = MetropolisHastings(w)
f = LogDensity(x -> -x^2/2)
print("Sampling π_∞")
@time chains = sample(f, s, MCMCSerial(), 10000, 4)
analyse(chains)
end
begin # Sample Based Energy Function
w = CyclicWalk(-1, 1, .5)
s = MetropolisHastings(w)
n = 100
z = [ rand(2) .* 2 .- 1 for i=1:n ]
f = SampledLogDensity(z, (x,z) -> - ( sum(z.^2) < x^2 )/2 )
print("Sampling π_$n")
@time chains = sample(f, s, MCMCSerial(), 10000, 4)
analyse(chains)
end