lazymatrix::RowStats

LazyMatrix for Rust · 0.9.0

Source

pub trait RowStats<F: Scalar>: MatrixShape + MatrixErrorType

Row statistics used to fit independently normalized samples.

Implementations sum duplicate cells before nonlinear statistics and include implicit zeros. Present vectors have length nrows; inactive components are None. Raw zero scales remain zero here. Fitting replaces them with one.

Methods

dense_row_normalization_stats

fn dense_row_normalization_stats(
    &self,
    spec: RowNormalization
) -> RowNormalizationStats<F> where Self: MatrixRead<F> + Sized

Source

Compute raw row statistics from dense indexed storage.

Dense adapters can delegate their fitting hook to this implementation. It uses O(nrows + ncols) storage, including one working row for quantiles.

row_normalization_stats

fn row_normalization_stats(
    &self,
    spec: RowNormalization
) -> Result<RowNormalizationStats<F>, <Self as MatrixErrorType>::Error>

Source

Compute the centers and raw scales requested by spec.

Backends can share scans through this hook. Returns the first storage error encountered while computing the statistics.

Implementation of RowStats

Source

impl<M: lazymatrix::RowStats<F>, F: Scalar> RowStats<F> for SprsCsc<M>

Implementation of RowStats

Source

impl<M: lazymatrix::RowStats<F>, F: Scalar> RowStats<F> for SprsCsr<M>