PCA example throws error

[rkube]

Hi,
I'm trying to run the PCA example for the IRIS code but get an error when calling predict:

julia> M = fit(PCA, Xtr; maxoutdim=3)
PCA(indim = 4, outdim = 3, principalratio = 0.9957325846529407)

julia> Yte = predict(M, Xte)
ERROR: MethodError: no method matching predict(::PCA{Float64}, ::LinearAlgebra.Adjoint{Float64, Matrix{Float64}})
Closest candidates are:
  predict(::Discriminant, ::AbstractMatrix{T} where T) at /home/rkube/.julia/packages/MultivariateStats/HTpHt/src/lda.jl:26
Stacktrace:
 [1] top-level scope

Does anyone know what is going on here?

[wildart]

Which version are you using? What is your julia version? I cannot reproduce this error on 1.6.4 & 1.7.1.

[smartinsightsfromdata]

@rkube The error is relative to a conflict on the keyword transform (I believe the conflict being between MultivariateStats and RDatasets).

Change it as:

Yte = MultivariateStats.transform(M, Xte)

(but there are errors further down in the example that I haven't been able to solve - I'll create an issue on it)

[wildart]

The example has been updated in the new documentation. Please, take a look.