Struct linfa_bayes::GaussianNbParams

pub struct GaussianNbParams { /* fields omitted */ }

Gaussian Naive Bayes (GaussianNB)

The Gaussian Naive Bayes is a classification algorithm where the likelihood of the feature P(x_i | y) is assumed to be Gaussian, features are assumed to be independent, and the mean and variance are estimated using maximum likelihood.

Implementations

impl GaussianNbParams

pub fn params() -> Self

Create new GaussianNB model with default values for its parameters

pub fn var_smoothing(self, var_smoothing: f64) -> Self

Specifies the portion of the largest variance of all the features that is added to the variance for calculation stability

Trait Implementations

impl Debug for GaussianNbParams

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for GaussianNbParams

fn default() -> Self

Returns the “default value” for a type.

impl<F, D, L> Fit<'_, ArrayBase<D, Dim<[usize; 2]>>, L> for GaussianNbParams where
    F: Float,
    D: Data<Elem = F>,
    L: AsTargets<Elem = usize> + Labels<Elem = usize>, 

type Object = Result<GaussianNb<F>>

fn fit(&self, dataset: &DatasetBase<ArrayBase<D, Ix2>, L>) -> Self::Object

Fit the model

Example

let x = array![
    [-2., -1.],
    [-1., -1.],
    [-1., -2.],
    [1., 1.],
    [1., 2.],
    [2., 1.]
];
let y = array![1, 1, 1, 2, 2, 2];

let data = Dataset::new(x, y);
let model = GaussianNbParams::params().fit(&data)?;
let pred = model.predict(&data);

assert_abs_diff_eq!(pred, data.try_single_target()?);

impl<F, D, L> IncrementalFit<'_, ArrayBase<D, Dim<[usize; 2]>>, L> for GaussianNbParams where
    F: Float,
    D: Data<Elem = F>,
    L: AsTargets<Elem = usize> + Labels<Elem = usize>, 

type ObjectIn = Option<GaussianNb<F>>

type ObjectOut = Result<Option<GaussianNb<F>>>

fn fit_with(
    &self,
    model_in: Self::ObjectIn,
    dataset: &DatasetBase<ArrayBase<D, Ix2>, L>
) -> Self::ObjectOut

Incrementally fit on a batch of samples

Example

let x = array![
    [-2., -1.],
    [-1., -1.],
    [-1., -2.],
    [1., 1.],
    [1., 2.],
    [2., 1.]
];
let y = array![1, 1, 1, 2, 2, 2];

let mut clf = GaussianNbParams::params();
let mut model = None;

for (x, y) in x
    .axis_chunks_iter(Axis(0), 2)
    .zip(y.axis_chunks_iter(Axis(0), 2))
{
    model = clf.fit_with(model, &DatasetView::new(x, y))?;
}

let pred = model.as_ref().unwrap().predict(&x);

assert_abs_diff_eq!(pred, y);