Struct OneHotEncoder 

pub struct OneHotEncoder { /* private fields */ }

Online one-hot encoder for string features.

Categories are discovered incrementally via update_strs. Before any category is seen, transform_strs returns an empty vector.

Implementations

impl OneHotEncoder

pub fn new() -> Self

Create a new empty encoder.

pub fn categories(&self) -> &[String]

The known categories, kept in sorted order.

pub fn category_index(&self, category: &str) -> Option<usize>

Find the index of category using binary search.

pub fn fit_one(&mut self, category: &str)

Add a category if not already present, keeping the list sorted.

pub fn transform_strs(&self, features: &[&str]) -> Result<Vec<f64>, RillError>

One-hot encode each string.

Output length = features.len() * categories.len(). Each group of categories.len() consecutive values has a 1.0 at the category index and 0.0 elsewhere.

Errors

• RillError::EmptyFeatures if features is empty.
• RillError::UnknownCategory if any string is not in the known categories.

Before any category has been seen, returns an empty vector.

pub fn update_strs(&mut self, features: &[&str]) -> Result<(), RillError>

Add all new categories from features and increment samples_seen.

Errors

Returns RillError::EmptyFeatures if features is empty.

pub fn samples_seen(&self) -> u64

How many samples have been seen.

pub fn reset(&mut self)

Reset the encoder to its initial empty state.

Trait Implementations

impl Clone for OneHotEncoder

fn clone(&self) -> OneHotEncoder

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for OneHotEncoder

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

Formats the value using the given formatter.

impl Default for OneHotEncoder

fn default() -> OneHotEncoder

Returns the “default value” for a type.