API

Training

# NeuroTreeModels.fit — Function.

function fit(
    config::NeuroTreeRegressor,
    dtrain;
    feature_names,
    target_name,
    weight_name=nothing,
    offset_name=nothing,
    deval=nothing,
    metric=nothing,
    print_every_n=9999,
    early_stopping_rounds=9999,
    verbosity=1,
    return_logger=false
)

Training function of NeuroTreeModels' internal API.

Arguments

• config::NeuroTreeRegressor

• dtrain: Must be a AbstractDataFrame

Keyword arguments

• feature_names: Required kwarg, a Vector{Symbol} or Vector{String} of the feature names.

• target_name Required kwarg, a Symbol or String indicating the name of the target variable.

• weight_name=nothing

• offset_name=nothing

• deval=nothing Data for tracking evaluation metric and perform early stopping.

• metric=nothing: evaluation metric tracked on deval. Can be one of:

  • :mse

  • :mae

  • :logloss

  • :mlogloss

  • :gaussian_mle

• print_every_n=9999

• early_stopping_rounds=9999

• verbosity=1

• return_logger=false

source

Inference

# NeuroTreeModels.infer — Function.

infer(m::NeuroTreeModel, data)

Return the inference of a NeuroTreeModel over data, where data is AbstractDataFrame.

source