@logistics-ts/classification

Inventory and demand classification for logistics-ts: ABC (value), XYZ (variability), FSN (movement), the ABC-XYZ policy matrix, and the Syntetos–Boylan–Croston demand-pattern classifier that drives forecasting method selection. Every classifier returns an Explained<T> result.

Install

npm i @logistics-ts/classification

What's in it

• abc — Pareto value classification (the "vital few").
• xyz — demand-variability classification from a bucketed series.
• fsn — fast / slow / non-moving split.
• abcXyzMatrix — combines ABC + XYZ into a policy-hinted cell per item.
• classifyDemandPattern — smooth / erratic / intermittent / lumpy (Syntetos–Boylan–Croston quadrant); this is what @logistics-ts/forecasting's autoForecast routes on.

Quick start

import { bucketize, generateExampleData } from '@logistics-ts/core'
import { abc, abcXyzMatrix, xyz } from '@logistics-ts/classification'

const { demand } = generateExampleData({ items: 6, seed: 2 })
const series = bucketize(demand, 'month')

const items = series.map((s) => ({
  itemId: s.itemId,
  volume: s.buckets.reduce((sum, b) => sum + b.quantity, 0),
  unitValue: demand.find((d) => d.itemId === s.itemId)?.unitPrice ?? 0,
}))

const abcResult = abc(items, { by: 'value' })
const xyzResult = xyz(series)
const matrix = abcXyzMatrix(abcResult.value, xyzResult.value)

console.log(matrix.value) // [{ itemId, class: 'AX'..'CZ', policyHint }, ...]

In the umbrella package

@logistics-ts/classification is re-exported as the classification namespace from logistics-ts. It depends only on @logistics-ts/core.

Links

• Full docs, API map, and examples
• Other logistics-ts packages

License

MIT