stats/data_binning_utils library

Data binning helpers: width, quantile, boundary, counts (roadmap #585).

Binning groups continuous values into a small set of buckets for histograms, bucketed metrics, or feature discretization. Width binning uses uniform intervals; quantile binning uses cut points so each bucket holds a similar count.

Functions

binByBoundaries(List<num> values, List<num> boundaries) → List<int>

Assign each value in values to a bin using ascending boundaries.

binByWidth(List<num> values, {required num min, required num max, required int bins}) → List<int>

Assign each value in values to an equal-width bin over [min, max].

binCounts(List<int> binIndices, int bins) → List<int>

Frequency of each bin index in binIndices, over bins buckets.

quantileBoundaries(List<num> values, int bins) → List<num>

The bins-1 internal cut points splitting sorted values into bins roughly equal-count groups (quantile boundaries).