I am getting the following error when I try to use the kmean function:
Julia version 1.5.3
All the packages are updated
using Clustering
using Distances
using Statistics
X = rand(5, 1000)
R = kmeans(X, 20; maxiter=200, display=:iter)
I can’t reproduce the error:
julia> using Clustering
julia> using Distances
julia> using Statistics
julia> X = rand(5, 1000);
julia> R = kmeans(X, 20; maxiter=200, display=:iter)
Iters objv objv-change | affected
-------------------------------------------------------------
0 2.041035e+02
1 1.453690e+02 -5.873454e+01 | 20
2 1.369302e+02 -8.438822e+00 | 20
3 1.339511e+02 -2.979075e+00 | 20
4 1.320703e+02 -1.880780e+00 | 20
5 1.311573e+02 -9.130492e-01 | 20
6 1.304643e+02 -6.929885e-01 | 19
7 1.298840e+02 -5.802423e-01 | 15
8 1.293135e+02 -5.705580e-01 | 15
9 1.287156e+02 -5.978245e-01 | 15
10 1.277974e+02 -9.181926e-01 | 15
11 1.272536e+02 -5.438166e-01 | 15
12 1.269656e+02 -2.880654e-01 | 15
13 1.266944e+02 -2.712121e-01 | 15
14 1.263861e+02 -3.082485e-01 | 18
15 1.262145e+02 -1.716252e-01 | 10
16 1.261344e+02 -8.009206e-02 | 4
17 1.261112e+02 -2.315701e-02 | 2
18 1.260852e+02 -2.602578e-02 | 3
19 1.260378e+02 -4.740786e-02 | 2
20 1.260236e+02 -1.416028e-02 | 5
21 1.259967e+02 -2.694733e-02 | 4
22 1.259770e+02 -1.970601e-02 | 5
23 1.259224e+02 -5.454997e-02 | 5
24 1.258796e+02 -4.281383e-02 | 3
25 1.258489e+02 -3.076725e-02 | 6
26 1.257414e+02 -1.074401e-01 | 7
27 1.255379e+02 -2.034701e-01 | 11
28 1.252256e+02 -3.123244e-01 | 11
29 1.250767e+02 -1.488813e-01 | 15
30 1.248356e+02 -2.411395e-01 | 13
31 1.245145e+02 -3.211148e-01 | 14
32 1.240144e+02 -5.000809e-01 | 13
33 1.236681e+02 -3.462645e-01 | 11
34 1.234899e+02 -1.782160e-01 | 6
35 1.234166e+02 -7.331246e-02 | 2
36 1.234007e+02 -1.591534e-02 | 4
37 1.232947e+02 -1.059802e-01 | 4
38 1.231907e+02 -1.040324e-01 | 4
39 1.231226e+02 -6.811605e-02 | 4
40 1.230901e+02 -3.242738e-02 | 0
41 1.230901e+02 0.000000e+00 | 0
K-means converged with 41 iterations (objv = 123.09013843381697)
KmeansResult{Array{Float64,2},Float64,Int64}([0.8097632681462386 0.5942172853403593 … 0.41041465915327047 0.2871475920064936; 0.4423317451301203 0.7727740451896872 … 0.20892369888129614 0.33620661064553803; … ; 0.2638807518254864 0.7631791027492311 … 0.2706224597515288 0.7700235299058837; 0.7422600696416772 0.787743528767687 … 0.3020438637946737 0.20631202895324988], [1, 15, 2, 19, 12, 6, 9, 20, 5, 8 … 10, 17, 9, 9, 15, 11, 14, 3, 2, 12], [0.23008209271154545, 0.0925096532284333, 0.18198522158592567, 0.0936893822488023, 0.03152772952743099, 0.1628032272375204, 0.14118752288088166, 0.1298405914640286, 0.1311384618873661, 0.07454307862249632 … 0.008549293563550275, 0.14804896139196555, 0.16342915243080292, 0.14462644500488597, 0.1275835264533618, 0.18969570228617894, 0.07135802657533619, 0.06272456250491709, 0.18269737453993606, 0.19772766113312157], [36, 55, 57, 53, 40, 51, 44, 57, 52, 48, 49, 46, 51, 56, 53, 49, 37, 63, 49, 54], [36, 55, 57, 53, 40, 51, 44, 57, 52, 48, 49, 46, 51, 56, 53, 49, 37, 63, 49, 54], 123.09013843381697, 41, true
Maybe check your versions:
(@v1.5) pkg> st Clustering Distances
Status `~/.julia/environments/v1.5/Project.toml`
[aaaa29a8] Clustering v0.14.2
[b4f34e82] Distances v0.10.3
This is what I get (The same as yours):
(@v1.5) pkg> st Clustering Distances
Status `~/.julia/environments/v1.5/Project.toml`
[aaaa29a8] Clustering v0.14.2
[b4f34e82] Distances v0.10.3
This is very annoying. I don’t know what is the problem. It used to work but it just stopped working and giving me errors. I restarted Julia but it didn’t help.
At first I faced the same issue. after updating to the latest versions I started getting
. Illegal inttoptr
%37 = ptrtoint double addrspace(13)* %36 to i64
Illegal inttoptr
%56 = inttoptr i64 %55 to i8 addrspace(13)*
Absolutely meaningless description for me
Only kmeans craches. The other methods (kmedoids, dbscan and so on) work correctly
If one wants, KMeans (Kmedoids, GMM,…) is also available in the BetaML package:
julia> using BetaML
julia> X = rand(5,1000)
5×1000 Matrix{Float64}:
0.0715517 0.100949 0.3514 0.503785 … 0.157863 0.0471136 0.16259 0.575626
0.856428 0.221503 0.0362716 0.543169 0.608833 0.504425 0.197979 0.255666
0.63662 0.202124 0.404668 0.425744 0.942191 0.649821 0.263136 0.0224259
0.622316 0.28606 0.260308 0.217047 0.404194 0.534265 0.475211 0.22486
0.257915 0.547401 0.655035 0.2384 0.484519 0.728864 0.908609 0.0515379
julia> R = kmeans(X,20)
([10, 10, 11, 10, 11], [0.09908971856173475 0.11528474365765297 … 0.18562366412084563 0.040032244612623044; 0.13755257307199592 0.13716311464831238 … 0.22218234341572918 0.06714182899564042; … ; 0.7914210997464357 0.5090954214895226 … 0.8436798914287494 0.5280047635069359; 0.8298839542566969 0.5309737924801818 … 0.8802385707236331 0.5551143478899533])
julia> S = kmeans(X,20,initStrategy="random")
([7, 14, 11, 6, 14], [0.6645401579066887 0.24639879500963735 … 0.35925334730152403 0.39712960954993387; 0.7034044964027021 0.28444705440179846 … 0.18323978520955944 0.2791541389879703; … ; 0.7069184750950341 0.4924644518658278 … 0.6396535310804825 0.3863272262424791; 0.5599258108009191 0.4799037791696229 … 0.616686683602457 0.5139395438797025])
Note in this case the importance of the centroid init strategy…
Doc: Clustering · BetaML.jl Documentation
(or use the MLJ interface)
The same issue with BetaML. The problem is systematic
пн, 6 сент. 2021 г., 16:29 Antonello Lobianco via JuliaLang <julialang@discoursemail.com>:
Then reinstall Julia and clean the Julia directory…