Public API
fit_evotree
EvoTrees.fit_evotree — Functionfit_evotree(
+ params::EvoTypes{L},
+ dtrain;
+ target_name,
+ fnames=nothing,
+ w_name=nothing,
+ offset_name=nothing,
+ deval=nothing,
+ metric=nothing,
+ early_stopping_rounds=9999,
+ print_every_n=9999,
+ verbosity=1,
+ return_logger=false,
+ device="cpu")Main training function. Performs model fitting given configuration params, dtrain, target_name and other optional kwargs.
Arguments
params::EvoTypes: configuration info providing hyper-paramters.EvoTypescan be one of:dtrain: A Tables compatible training data (named tuples, DataFrame...) containing features and target variables.
Keyword arguments
target_name: name of target variable.fnames = nothing: the names of thex_trainfeatures. If provided, should be a vector of string withlength(fnames) = size(x_train, 2).w_name = nothing: name of the variable containing weights. Ifnothing, common weights on one will be used.offset_name = nothing: name of the offset variable.deval: A Tables compatible evaluation data containing features and target variables.metric: The evaluation metric that wil be tracked ondeval. Supported metrics are::mse: mean-squared error. Adapted for general regression models.:rmse: root-mean-squared error (CPU only). Adapted for general regression models.:mae: mean absolute error. Adapted for general regression models.:logloss: Adapted for:logisticregression models.:mlogloss: Multi-class cross entropy. Adapted toEvoTreeClassifierclassification models.:poisson: Poisson deviance. Adapted toEvoTreeCountcount models.:gamma: Gamma deviance. Adapted to regression problem on Gamma like, positively distributed targets.:tweedie: Tweedie deviance. Adapted to regression problem on Tweedie like, positively distributed targets with probability mass aty == 0.:gaussian_mle: Gaussian maximum log-likelihood. Adapted toEvoTreeMLEmodels withloss = :gaussian_mle.:logistic_mle: Logistic maximum log-likelihood. Adapted toEvoTreeMLEmodels withloss = :logistic_mle.
early_stopping_rounds::Integer: number of consecutive rounds without metric improvement after which fitting in stopped.print_every_n: sets at which frequency logging info should be printed.verbosity: set to 1 to print logging info during training.return_logger::Bool = false: if set to true (default),fit_evotreereturn a tuple(m, logger)where logger is a dict containing various tracking information.device="cpu": Hardware device to use for computations. Can be either"cpu"or"gpu". Following losses are not GPU supported at the moment:l1,:quantile,:logistic_mle.
fit_evotree(
+ params::EvoTypes{L};
+ x_train::AbstractMatrix,
+ y_train::AbstractVector,
+ w_train=nothing,
+ offset_train=nothing,
+ x_eval=nothing,
+ y_eval=nothing,
+ w_eval=nothing,
+ offset_eval=nothing,
+ early_stopping_rounds=9999,
+ print_every_n=9999,
+ verbosity=1)Main training function. Performs model fitting given configuration params, x_train, y_train and other optional kwargs.
Arguments
params::EvoTypes: configuration info providing hyper-paramters.EvoTypescan be one of:
Keyword arguments
x_train::Matrix: training data of size[#observations, #features].y_train::Vector: vector of train targets of length#observations.w_train::Vector: vector of train weights of length#observations. Ifnothing, a vector of ones is assumed.offset_train::VecOrMat: offset for the training data. Should match the size of the predictions.x_eval::Matrix: evaluation data of size[#observations, #features].y_eval::Vector: vector of evaluation targets of length#observations.w_eval::Vector: vector of evaluation weights of length#observations. Defaults tonothing(assumes a vector of 1s).offset_eval::VecOrMat: evaluation data offset. Should match the size of the predictions.metric: The evaluation metric that wil be tracked onx_eval,y_evaland optionallyw_eval/offset_evaldata. Supported metrics are::mse: mean-squared error. Adapted for general regression models.:rmse: root-mean-squared error (CPU only). Adapted for general regression models.:mae: mean absolute error. Adapted for general regression models.:logloss: Adapted for:logisticregression models.:mlogloss: Multi-class cross entropy. Adapted toEvoTreeClassifierclassification models.:poisson: Poisson deviance. Adapted toEvoTreeCountcount models.:gamma: Gamma deviance. Adapted to regression problem on Gamma like, positively distributed targets.:tweedie: Tweedie deviance. Adapted to regression problem on Tweedie like, positively distributed targets with probability mass aty == 0.:gaussian_mle: Gaussian maximum log-likelihood. Adapted toEvoTreeMLEmodels withloss = :gaussian_mle.:logistic_mle: Logistic maximum log-likelihood. Adapted toEvoTreeMLEmodels withloss = :logistic_mle.
early_stopping_rounds::Integer: number of consecutive rounds without metric improvement after which fitting in stopped.print_every_n: sets at which frequency logging info should be printed.verbosity: set to 1 to print logging info during training.fnames: the names of thex_trainfeatures. If provided, should be a vector of string withlength(fnames) = size(x_train, 2).return_logger::Bool = false: if set to true (default),fit_evotreereturn a tuple(m, logger)where logger is a dict containing various tracking information.device="cpu": Hardware device to use for computations. Can be either"cpu"or"gpu". Following losses are not GPU supported at the moment:l1,:quantile,:logistic_mle.
predict
MLJModelInterface.predict — Functionpredict(model::EvoTree, X::AbstractMatrix; ntree_limit = length(model.trees))Predictions from an EvoTree model - sums the predictions from all trees composing the model. Use ntree_limit=N to only predict with the first N trees.
importance
EvoTrees.importance — Functionimportance(model::EvoTree; fnames=model.info[:fnames])Sorted normalized feature importance based on loss function gain. Feature names associated to the model are stored in model.info[:fnames] as a string Vector and can be updated at any time. Eg: model.info[:fnames] = new_fnames_vec.