Trait ndarray_stats::QuantileExt

source · [−]
pub trait QuantileExt<A, S, D> where
    S: Data<Elem = A>,
    D: Dimension, {
    fn argmin(&self) -> Result<D::Pattern, MinMaxError>
    where
        A: PartialOrd;
    fn argmin_skipnan(&self) -> Result<D::Pattern, EmptyInput>
    where
        A: MaybeNan,
        A::NotNan: Ord;
    fn min(&self) -> Result<&A, MinMaxError>
    where
        A: PartialOrd;
    fn min_skipnan(&self) -> &A
    where
        A: MaybeNan,
        A::NotNan: Ord;
    fn argmax(&self) -> Result<D::Pattern, MinMaxError>
    where
        A: PartialOrd;
    fn argmax_skipnan(&self) -> Result<D::Pattern, EmptyInput>
    where
        A: MaybeNan,
        A::NotNan: Ord;
    fn max(&self) -> Result<&A, MinMaxError>
    where
        A: PartialOrd;
    fn max_skipnan(&self) -> &A
    where
        A: MaybeNan,
        A::NotNan: Ord;
    fn quantile_axis_mut<I>(
        &mut self,
        axis: Axis,
        q: N64,
        interpolate: &I
    ) -> Result<Array<A, D::Smaller>, QuantileError>
    where
        D: RemoveAxis,
        A: Ord + Clone,
        S: DataMut,
        I: Interpolate<A>;
    fn quantiles_axis_mut<S2, I>(
        &mut self,
        axis: Axis,
        qs: &ArrayBase<S2, Ix1>,
        interpolate: &I
    ) -> Result<Array<A, D>, QuantileError>
    where
        D: RemoveAxis,
        A: Ord + Clone,
        S: DataMut,
        S2: Data<Elem = N64>,
        I: Interpolate<A>;
    fn quantile_axis_skipnan_mut<I>(
        &mut self,
        axis: Axis,
        q: N64,
        interpolate: &I
    ) -> Result<Array<A, D::Smaller>, QuantileError>
    where
        D: RemoveAxis,
        A: MaybeNan,
        A::NotNan: Clone + Ord,
        S: DataMut,
        I: Interpolate<A::NotNan>;
    fn __private__(&self, _: PrivateMarker);
}
Expand description

Quantile methods for ArrayBase.

Required Methods

Finds the index of the minimum value of the array.

Returns Err(MinMaxError::UndefinedOrder) if any of the pairwise orderings tested by the function are undefined. (For example, this occurs if there are any floating-point NaN values in the array.)

Returns Err(MinMaxError::EmptyInput) if the array is empty.

Even if there are multiple (equal) elements that are minima, only one index is returned. (Which one is returned is unspecified and may depend on the memory layout of the array.)

Example
use ndarray::array;
use ndarray_stats::QuantileExt;

let a = array![[1., 3., 5.],
               [2., 0., 6.]];
assert_eq!(a.argmin(), Ok((1, 1)));

Finds the index of the minimum value of the array skipping NaN values.

Returns Err(EmptyInput) if the array is empty or none of the values in the array are non-NaN values.

Even if there are multiple (equal) elements that are minima, only one index is returned. (Which one is returned is unspecified and may depend on the memory layout of the array.)

Example
use ndarray::array;
use ndarray_stats::QuantileExt;

let a = array![[::std::f64::NAN, 3., 5.],
               [2., 0., 6.]];
assert_eq!(a.argmin_skipnan(), Ok((1, 1)));

Finds the elementwise minimum of the array.

Returns Err(MinMaxError::UndefinedOrder) if any of the pairwise orderings tested by the function are undefined. (For example, this occurs if there are any floating-point NaN values in the array.)

Returns Err(MinMaxError::EmptyInput) if the array is empty.

Even if there are multiple (equal) elements that are minima, only one is returned. (Which one is returned is unspecified and may depend on the memory layout of the array.)

Finds the elementwise minimum of the array, skipping NaN values.

Even if there are multiple (equal) elements that are minima, only one is returned. (Which one is returned is unspecified and may depend on the memory layout of the array.)

Warning This method will return a NaN value if none of the values in the array are non-NaN values. Note that the NaN value might not be in the array.

Finds the index of the maximum value of the array.

Returns Err(MinMaxError::UndefinedOrder) if any of the pairwise orderings tested by the function are undefined. (For example, this occurs if there are any floating-point NaN values in the array.)

Returns Err(MinMaxError::EmptyInput) if the array is empty.

Even if there are multiple (equal) elements that are maxima, only one index is returned. (Which one is returned is unspecified and may depend on the memory layout of the array.)

Example
use ndarray::array;
use ndarray_stats::QuantileExt;

let a = array![[1., 3., 7.],
               [2., 5., 6.]];
assert_eq!(a.argmax(), Ok((0, 2)));

Finds the index of the maximum value of the array skipping NaN values.

Returns Err(EmptyInput) if the array is empty or none of the values in the array are non-NaN values.

Even if there are multiple (equal) elements that are maxima, only one index is returned. (Which one is returned is unspecified and may depend on the memory layout of the array.)

Example
use ndarray::array;
use ndarray_stats::QuantileExt;

let a = array![[::std::f64::NAN, 3., 5.],
               [2., 0., 6.]];
assert_eq!(a.argmax_skipnan(), Ok((1, 2)));

Finds the elementwise maximum of the array.

Returns Err(MinMaxError::UndefinedOrder) if any of the pairwise orderings tested by the function are undefined. (For example, this occurs if there are any floating-point NaN values in the array.)

Returns Err(EmptyInput) if the array is empty.

Even if there are multiple (equal) elements that are maxima, only one is returned. (Which one is returned is unspecified and may depend on the memory layout of the array.)

Finds the elementwise maximum of the array, skipping NaN values.

Even if there are multiple (equal) elements that are maxima, only one is returned. (Which one is returned is unspecified and may depend on the memory layout of the array.)

Warning This method will return a NaN value if none of the values in the array are non-NaN values. Note that the NaN value might not be in the array.

Return the qth quantile of the data along the specified axis.

q needs to be a float between 0 and 1, bounds included. The qth quantile for a 1-dimensional lane of length N is defined as the element that would be indexed as (N-1)q if the lane were to be sorted in increasing order. If (N-1)q is not an integer the desired quantile lies between two data points: we return the lower, nearest, higher or interpolated value depending on the interpolate strategy.

Some examples:

  • q=0. returns the minimum along each 1-dimensional lane;
  • q=0.5 returns the median along each 1-dimensional lane;
  • q=1. returns the maximum along each 1-dimensional lane. (q=0 and q=1 are considered improper quantiles)

The array is shuffled in place along each 1-dimensional lane in order to produce the required quantile without allocating a copy of the original array. Each 1-dimensional lane is shuffled independently from the others. No assumptions should be made on the ordering of the array elements after this computation.

Complexity (quickselect):

  • average case: O(m);
  • worst case: O(m^2); where m is the number of elements in the array.

Returns Err(EmptyInput) when the specified axis has length 0.

Returns Err(InvalidQuantile(q)) if q is not between 0. and 1. (inclusive).

Panics if axis is out of bounds.

A bulk version of quantile_axis_mut, optimized to retrieve multiple quantiles at once.

Returns an Array, where subviews along axis of the array correspond to the elements of qs.

See quantile_axis_mut for additional details on quantiles and the algorithm used to retrieve them.

Returns Err(EmptyInput) when the specified axis has length 0.

Returns Err(InvalidQuantile(q)) if any q in qs is not between 0. and 1. (inclusive).

Panics if axis is out of bounds.

Example
use ndarray::{array, aview1, Axis};
use ndarray_stats::{QuantileExt, interpolate::Nearest};
use noisy_float::types::n64;

let mut data = array![[3, 4, 5], [6, 7, 8]];
let axis = Axis(1);
let qs = &[n64(0.3), n64(0.7)];
let quantiles = data.quantiles_axis_mut(axis, &aview1(qs), &Nearest).unwrap();
for (&q, quantile) in qs.iter().zip(quantiles.axis_iter(axis)) {
    assert_eq!(quantile, data.quantile_axis_mut(axis, q, &Nearest).unwrap());
}

Return the qth quantile of the data along the specified axis, skipping NaN values.

See quantile_axis_mut for details.

source

fn __private__(&self, _: PrivateMarker)

This method makes this trait impossible to implement outside of ndarray-stats so that we can freely add new methods, etc., to this trait without breaking changes.

We don’t anticipate any other crates needing to implement this trait, but if you do have such a use-case, please let us know.

Warning This method is not considered part of the public API, and client code should not rely on it being present. It may be removed in a non-breaking release.

Implementations on Foreign Types

source

fn __private__(&self, _: PrivateMarker)

Implementors