Hi,
I am currently trying to step through the tutorial posted at https://tutorials.pumas.ai/html/introduction/introduction.html
(side note, a lot of the documentation links are not being found because they have an exta /dev in the URL)
I am using my own data, so I unfortunately can’t post too much detail on that.
The first thing I noticed it that my gof plots do not have local regression lines (otherwise they work)
Then, when running the VPC, I get an error LinearAlgebra.LAPACKException(2)
I appreciate any help in making things work.
Running Pumas v1.0.2, PumasPlots v0.1.3
Thanks,
Sven
Hi Sven
The first thing I noticed it that my gof plots do not have local regression lines (otherwise they work)
Could you please show us the full output of ] st
? It might be an issue with one of the plotting packages so
I’d like to know exactly which version you are using.
Then, when running the VPC, I get an error LinearAlgebra.LAPACKException(2)
Please see the tip box at https://docs.pumas.ai/analysis/vpc/#Visual-Predictive-Check-(VPC). We are working on intercepting the exception to provide a better error message.
(side note, a lot of the documentation links are not being found because they have an exta /dev in the URL)
Thanks for letting us know. We made a lot of changes to the websites last week and indeed it looks like it broke some links. We are working on fixing it.
Thank you for the quick reply. Increasing bandwidth as suggested fixed the VPC.
This is my output for st
[c52e3926] Atom v0.12.14 ⚲
[6e4b80f9] BenchmarkTools v0.5.0
[4f18b42c] BusinessDays v0.9.11
[336ed68f] CSV v0.7.7
[49dc2e85] Calculus v0.5.1
[aaaa29a8] Clustering v0.14.1
[8f4d0f93] Conda v1.4.1
[3a865a2d] CuArrays v2.2.2
[717857b8] DSP v0.6.7
[a93c6f00] DataFrames v0.21.5
[864edb3b] DataStructures v0.17.20
[31a5f54b] Debugger v0.6.5
[7806a523] DecisionTree v0.10.7
[0c46a032] DifferentialEquations v6.15.0
[31c24e10] Distributions v0.23.2
[7876af07] Example v0.5.3
[587475ba] Flux v0.10.4
[59287772] Formatting v0.4.1
[38e38edf] GLM v1.3.9
[28b8d3ca] GR v0.46.90
[09f84164] HypothesisTests v0.10.0
[7073ff75] IJulia v1.21.2
[6218d12a] ImageMagick v1.1.5
[916415d5] Images v0.22.4
[c601a237] Interact v0.10.3
[6042db11] JDBC v0.5.0
[033835bb] JLD2 v0.1.14
[682c06a0] JSON v0.21.0
[494afd89] JavaCall v0.7.4
[4076af6c] JuMP v0.21.3
[ea4eb333] JuliaRunClient v0.2.2
[480116ec] JuliaWebAPI v0.6.2
[e5e0dc1b] Juno v0.8.2 ⚲
[5ab0869b] KernelDensity v0.5.1
[1902f260] Knet v1.3.7
[b964fa9f] LaTeXStrings v1.1.0
[2ec943e9] Libz v1.0.1
[093fc24a] LightGraphs v1.3.3
[9c8b4983] LightXML v0.9.0
[f0e99cf1] MLBase v0.8.0
[ee78f7c6] Makie v0.10.0
[dbeba491] Metalhead v0.5.0
[3ae66bd7] Miletus v1.0.0
[b8a86587] NearestNeighbors v0.4.4
[49dea1ee] Nettle v0.5.0
[a15396b6] OnlineStats v1.5.3
[429524aa] Optim v0.20.1
[3b7a836e] PGFPlots v3.2.1
[d96e819e] Parameters v0.12.1
[fae87a5f] ParserCombinator v2.0.0
[b98c9c47] Pipe v1.3.0
[4722fa14] PkgAuthentication v0.1.1
[91a5bcdd] Plots v0.29.9
[3349acd9] ProtoBuf v0.8.0
[4f2c3c20] Pumas v1.0.2
[0c61e9cd] PumasPlots v0.1.3
[b7b41870] PumasTutorials v0.0.1
[438e738f] PyCall v1.91.4
[d330b81b] PyPlot v2.9.0
[fcd29c91] QuantEcon v0.16.2
[6f49c342] RCall v0.13.7
[ee283ea6] Rebugger v0.3.3
[f2b01f46] Roots v1.0.4
[3646fa90] ScikitLearn v0.6.2
[60ddc479] StatPlots v0.9.2
[90137ffa] StaticArrays v0.12.4
[2913bbd2] StatsBase v0.32.2
[65254759] StatsMakie v0.2.3
[f3b207a7] StatsPlots v0.14.7
[24249f21] SymPy v1.0.7
[40c74d1a] TableView v0.6.1
[1d978283] TensorFlow v0.11.0
[b8865327] UnicodePlots v1.2.0
[44d3d7a6] Weave v0.10.2
[e88e6eb3] Zygote v0.4.22
I forgot to ask, what is the output of versioninfo()
?
I think that some of the installed packages are causing other packages not to be up-to-date. It’s fixed in a the upcoming binary release but I assume it will be easier for you to simply adjust the installed packages.
In that case, could you please try to ] rm
the two packages StatPlots
(notice that it’s not StatsPlots
) and Metalhead
and then ] up
, then restart your Julia session, and report back to us if the local regression line appears in the plot.
Thank you Andreas, your suggested approach fixed the problem.