gordan

Gauss-Jordan + Regression JavaScript library

Installation

npm i gordan --save

Usage

import Gordan from 'gordan'; // for ES6

const Gordan = require('gordan');  // for Node

Main API

Gordan.solveByGaussJordan(matrix)

• matrix: the augmented matrix, a bidimensional array

• Returns: the identity matrix with the solution coefficients

Gordan.getLinearRegressionRect(points)

• Returns: a list of points for the regression rect

Gordan.getQuadraticRegressionCurve(points)

• Returns: a list of points for the regression curve (from cuadratic equation)

Gordan.getRegressionPath(points, N)

• Returns: a list of points for an Nth grade equation (ax^N + bx^(N - 1) + cx^(N - 2) + ...)

• points: for all cases, a list of x, y points. The following formats are supported:

[
  [1, 2],
  [2, 2],
  [3, 3]
]

[
  { "x": 1, "y": 2 },
  { "x": 2, "y": 2 },
  { "x": 3, "y": 3 }
]

Secondary API

Gordan.addRows(row1, row2, [invert1[, invert2]])

• row1: first row to add, a number array

• row2: second row to add, a number array

• invert1: boolean, if present, values in row1 are multiplied by -1

• invert2: boolean, if present, values in row2 are multiplied by -1

• Returns: the addition of the 2 rows (number[])

Gordan.multiplyRow(row, value)

• row: the row to multiply, a number array

• value: each number in row is multiplied by this number

• Returns: a new row with the multipled values (number[])

Gordan.divideRow(row, value)

• row: the row to divide, a number array

• value: each number in row is divided by this number

• Returns: a new row with the divided values (number[])

Gordan.getSymbolValues(matrix)

• matrix: the augmented matrix, a bidimensional array

• Returns: the last column of the resulting identity matrix (number[])

Gordan.normalizePoints(points)

• points: an array of [x, y] or {x, y} points

• Returns: an array of points with {x, y} format

Gordan.getRegressionMatrixFromPoints(points, degreeOfEquation)```

• points: an array of [x, y] or {x, y} points

• degreeOfEquation: a number greater than zero

• Returns: the regression augmented matrix

Gordan.getRange(points, axis)

• points: an array of [x, y] or {x, y} points

• axis: a string 'x' or 'y'

• Returns: 'x'/'y' limits on the plane for the given points

Gauss-Jordan example

let gaussJordanMatrix = [
  [ 1, -2,  2, -3,   15 ],
  [ 3,  4, -1,  1,   -6 ],
  [ 2, -3,  2, -1,   17 ],
  [ 1,  1, -3, -2,   -7 ]
];
 
let solvedMatrix = Gordan.solveByGaussJordan(gaussJordanMatrix);

Results in:

1, 	0, 	0, 	0, 	2

0, 	1, 	0, 	0, 	-2

0, 	0, 	1, 	0, 	3

0, 	0, 	0, 	1, 	-1

Regression example

let points = [
  [1, 2],
  [1.5, 4],
  [2, 2.4],
  [3, 5],
  [4, 3],
  [5.5, 9.3],
  [5, 6],
  [6, 10],
  [6.5, 8.5],
  [6.5, 13],
  [7, 12.5]
];
 
let rect = Gordan.getLinearRegressionRect(points);
let curve = Gordan.getQuadraticRegressionCurve(points);
let gradeSixCurve = Gordan.getRegressionPath(points, 6);

Results in:

Here's the Chart.js code needed for that:

let chart = new Chart(document.getElementById('chart'), {
  type: 'scatter',
  data: {
    datasets: [{
      label: 'The points',
      data: Gordan.normalizePoints(points),
      backgroundColor: '#f40'
    }, {
      label: 'Linear regression rect',
      data: rect,
      backgroundColor: '#fcc'
    }, {
      label: 'Quadratic regression curve',
      data: curve,
      backgroundColor: '#cfc'
    }, {
      label: 'Grade 6 regression curve',
      data: gradeSixCurve,
      backgroundColor: '#ccf'
    }]
  },
  options: {
    responsive: true
  }
});