sample-quantile

find the sample value(s) for one or many probabilities

• Introduction • API • Utilities • License •

var quartiles = sampleQuantile([0,1,2,3,4,5,6], [0.25, 0.5, 0.75]) // [1,3,5]
var minimum = sampleQuantile([2, 4, 6, 8], 0) // 2
var maximum = sampleQuantile([2, 4, 6, 8], 1) // 8
var median = sampleQuantile([1, 2], 0.5) // 1.5
var unsortedMax = = sampleQuantile([3, 1, 2], 1, true) // 3

Introduction

There are a number of npm packages to help derive the quantile or percentile values from a sample set with different assumptions and limitations. This implementation is based on the following:

• only works with probabilities between and including 0 and 1
• In-between probabilities are interpolated (R-6 presented here and discussed here)
• works with single probability to return a single value
• works with arrays of probabilities to return an array of values
• samples are not sorted by default to prevent overhead, but can be sorted if specified

API

sampleQuantile(arrayOfSamples, singleOrManyProbabilities[, sortFirst]) // --> singleOrManyValues

• arrayOfSamples is an array of sorted samples. If unsorted, set the sortFirst flag to true
• singleOrManyProbabilities single probability number (0 <= x <= 1) or an array of numbers
• sortFirst optional flag to duplicate and sort the data first. The source of samples remains unchanged

Utilities

When performance is important, the subfunctions for single numbers and for arrays can be called directly to avoid the internal checks:

require('sample-quantile/quantile')(samples, probability) : number require('sample-quantile/quantiles')(samples, probabilities) : Array<number>

License

Released under the MIT License