Struct ModelBuilderData 

pub struct ModelBuilderData<'a, M, Y, F>
where
    M: Glm,
    Y: Yval<M>,
    F: 'static + Float,
{ /* private fields */ }

Holds the data and all the specifications for the model and provides functions to adjust the settings.

Implementations

impl<'a, M, Y, F> ModelBuilderData<'a, M, Y, F>

where M: Glm, Y: Yval<M> + Copy, F: Float,

A builder to generate a Model object

pub fn linear_offset(self, linear_offset: Array1<F>) -> Self

Represents an offset added to the linear predictor for each data point. This can be used to control for fixed effects or in multi-level models.

pub fn freq_weights(self, freqs: Array1<usize>) -> Self

Frequency weights (a.k.a. counts) for each observation. Traditionally these are positive integers representing the number of times each observation appears identically.

pub fn var_weights(self, weights: Array1<F>) -> Self

Variance weights (a.k.a. analytic weights) of each observation. These could represent the inverse square of the uncertainties of each measurement.

pub fn no_constant(self) -> Self

Do not add a constant intercept term of 1s to the design matrix. This is rarely recommended, so you probably don’t want to use this option unless you have a very clear sense of why. Note that you can supply uniform or per-observation constant terms using ModelBuilderData::linear_offset.

pub fn no_standardize(self) -> Self

Don’t perform standarization (i.e. scale to 0-mean and 1-variance) of the design matrix. Note that the standardization is handled internally, so the reported result coefficients should be compatible with the input data directly, meaning the user shouldn’t have to interact with them.

pub fn build(self) -> RegressionResult<Model<M, F>, F>

where M: Glm, F: Float,