w-statistic
Version:
A tool for statistical analysis.
226 lines (202 loc) • 5.72 kB
JavaScript
import get from 'lodash-es/get.js'
import each from 'lodash-es/each.js'
import size from 'lodash-es/size.js'
import isnum from 'wsemi/src/isnum.mjs'
import isbol from 'wsemi/src/isbol.mjs'
import iseobj from 'wsemi/src/iseobj.mjs'
import isearr from 'wsemi/src/isearr.mjs'
import cdbl from 'wsemi/src/cdbl.mjs'
import ss from './simpleStatistics.mjs'
/**
* 針對x,y數據進行線性回歸(y=mx+b)
*
* Unit Test: {@link https://github.com/yuda-lyu/w-statistic/blob/master/test/regLine.test.js Github}
* @memberOf w-statistic
* @param {Array} arr 輸入陣列，只提取有效數字(或為字串的數字)進行計算
* @param {Object} [opt={}] 輸入設定物件，預設{}
* @param {Number} [opt.interpX=null] 輸入經由回歸結果內插指定x值數字，預設null
* @param {Boolean} [opt.calcR2=false] 輸入是否計算r2值布林值，預設false
* @param {Boolean} [opt.useRegIntercept=true] 輸入是否回歸使用截距布林值，預設true
* @param {Boolean} [opt.useSync=false] 輸入是否使用同步函數布林值，預設false
* @returns {Object|Promise} 若useSync=true回傳回歸結果物件，若useSync=false則回傳Promise，此時若成功則resolve回歸結果物件，若失敗則reject錯誤訊息
* @example
*
* async function test() {
*
* let arr
* let r
*
* arr = [
* [1, 2.5],
* [2.5, 1.1],
* [4, 0.5],
* ]
* r = await regLine(arr)
* console.log(r)
* // => { m: -0.6666666666666664, b: 3.0333333333333323 }
*
* arr = [
* [1, 0.5],
* [2.5, 1.1],
* [4, 2.5],
* ]
* r = await regLine(arr)
* console.log(r)
* // => { m: 0.666666666666667, b: -0.3000000000000007 }
*
* arr = [
* [1, 0.5],
* [2.5, 1.1],
* [4, 2.5],
* ]
* r = await regLine(arr, { useRegIntercept: false }) //不使用截距, 也就是截距=0
* console.log(r)
* // => { m: 0.5698924731182796, b: 0 }
*
* arr = [
* [1, 2.5],
* [2.5, 1.1],
* [4, 0.5],
* ]
* r = await regLine(arr, { interpX: 2 })
* console.log(r)
* // => {
* // m: -0.6666666666666664,
* // b: 3.0333333333333323,
* // interpY: 1.6999999999999995
* // }
*
* arr = [
* [1, 2.5],
* [2.5, 1.1],
* [4, 0.5],
* ]
* r = await regLine(arr, { calcR2: true })
* console.log(r)
* // => {
* // m: -0.6666666666666664,
* // b: 3.0333333333333323,
* // r2: 0.949367088607595
* // }
*
* arr = [
* [1, 2.5],
* [2.5, 1.1],
* [4, 0.5],
* ]
* r = regLine(arr, { useSync: true }) //使用同步函數(sync)
* console.log(r)
* // => { m: -0.6666666666666664, b: 3.0333333333333323 }
*
* }
* test()
* .catch((err) => {
* console.log(err)
* })
*
*/
function regLine(arr, opt = {}) {
//interpX
let interpX = get(opt, 'interpX')
if (!isnum(interpX)) {
interpX = null
}
//calcR2
let calcR2 = get(opt, 'calcR2')
if (!isbol(calcR2)) {
calcR2 = false
}
//useRegIntercept, 是否回歸使用截距
let useRegIntercept = get(opt, 'useRegIntercept')
if (!isbol(useRegIntercept)) {
useRegIntercept = true
}
//useSync
let useSync = get(opt, 'useSync')
if (!isbol(useSync)) {
useSync = false
}
//_sync
let _sync = () => {
//check arr
if (!isearr(arr)) {
throw new Error(`arr is not an effective array`)
}
//rs
let rs = []
each(arr, (v) => {
let x = get(v, 0)
let y = get(v, 1)
if (isnum(x) && isnum(y)) {
rs.push([cdbl(x), cdbl(y)])
}
})
// console.log('rs', rs)
//check
if (size(rs) === 0) {
throw new Error(`no effective data`)
}
//r
let r = null
//useRegIntercept
if (!useRegIntercept) {
//無截距算斜率m
let sumx2 = 0
let sumxy = 0
each(rs, (v) => {
sumx2 += v[0] ** 2
sumxy += v[0] * v[1]
})
let m = null
if (sumx2 !== 0) {
m = sumxy / sumx2
}
r = {
m,
b: 0,
}
// console.log('r(no b)', r)
}
else {
//linearRegression
r = ss.linearRegression(rs)
// console.log('r', r)
}
//interpX
if (isnum(interpX)) {
interpX = cdbl(interpX)
let interpY = ss.linearRegressionLine(r)(interpX)
r.interpY = interpY
}
//calcR2
if (calcR2) {
let r2 = ss.rSquared(rs, (x) => ss.linearRegressionLine(r)(x))
r.r2 = r2
}
return r
}
//_async
let _async = async () => {
let r = null
try {
r = _sync()
if (iseobj(r)) {
return r
}
else {
return Promise.reject(`no effective data`)
}
}
catch (err) {
console.log(err)
return Promise.reject(err.message)
}
}
if (useSync) {
return _sync()
}
else {
return _async()
}
}
export default regLine