MLJ NuSVM predict function not found

General Usage
1

I’m having an issue using MLJ NuSVR model. From the docs this should be working.

NuSVR = @load NuSVR pkg=LIBSVM                  
model = NuSVR(kernel=LIBSVM.Kernel.RadialBasis)  

# SVMRegressor = @load SVMRegressor pkg=MLJScikitLearnInterface
# model = SVMRegressor()

mach = machine(model, X, y) 
mach |> fit!

yhat = LIBSVM.predict(mach, Xnew)

I get this error:


MethodError: no method matching predict(::Machine{MLJLIBSVMInterface.NuSVR, true}, ::Vector{Float64})

Closest candidates are:
  predict(::LIBSVM.LinearSVC, ::AbstractArray)
   @ LIBSVM ~/.julia/packages/LIBSVM/IZl0L/src/ScikitLearnAPI.jl:11
  predict(::Union{LIBSVM.AbstractSVC, LIBSVM.AbstractSVR}, ::AbstractArray)
   @ LIBSVM ~/.julia/packages/LIBSVM/IZl0L/src/ScikitLearnAPI.jl:6
2

Checking Xnew’s type , should be same with X’s type

3 
import LIBSVM
import MLJ:predict
using MLJ


NuSVR = @load NuSVR pkg=LIBSVM                 ## model type
model = NuSVR(kernel=LIBSVM.Kernel.Polynomial) ## instance

X, y = make_regression(rng=123) ## table, vector
mach = machine(model, X, y) |> fit!

Xnew, _ = make_regression(3, rng=123)
yhat = predict(mach, Xnew)

import MLJ:predict . so works!

4

Thank you!