pub enum Objective {
}
Expand description
Learning objective used when training a booster model.
Variants§
RegLinear
Linear regression.
RegLogistic
Logistic regression.
BinaryLogistic
Logistic regression for binary classification, outputs probability.
BinaryLogisticRaw
Logistic regression for binary classification, outputs scores before logistic transformation.
GpuRegLinear
GPU version of RegLinear
.
GpuRegLogistic
GPU version of RegLogistic
.
GpuBinaryLogistic
GPU version of RegBinaryLogistic
.
GpuBinaryLogisticRaw
GPU version of RegBinaryLogisticRaw
.
CountPoisson
Poisson regression for count data, outputs mean of poisson distribution.
SurvivalCox
Cox regression for right censored survival time data (negative values are considered right censored).
predictions are returned on the hazard ratio scale (i.e., as HR = exp(marginal_prediction)
in the proportional hazard function h(t) = h0(t) * HR
).
MultiSoftmax(u32)
Multiclass classification using the softmax objective, with given number of classes.
MultiSoftprob(u32)
Multiclass classification using the softmax objective, with given number of classes.
Outputs probabilities per class.
RankPairwise
Ranking task which minimises pairwise loss.
RegGamma
Gamma regression with log-link. Output is the mean of the gamma distribution.
RegTweedie(Option<f32>)
Tweedie regression with log-link. Takes an optional tweedie variance power parameter which controls the variance of the Tweedie distribution.
- Set closer to 2 to shift towards a gamma distribution
- Set closer to 1 to shift towards a Poisson distribution
range: (1, 2)
Set to None
to use XGBoost’s default (currently 1.5
).