Numeric JavaScript

You should install node and npm.

For the jupyter kernel, I used this: https://www.npmjs.com/package/ijavascript

There's also this: https://github.com/notablemind/jupyter-nodejs

I bet there are others.

Numeric JS

Following along from http://www.numericjs.com/workshop.php

I am using the minified JS file in this repo.



In [1]:

    
var n = require('./numeric-1.2.6.min.js');









    Out[1]:





undefined



In [2]:

    
// Linear algebra example. We start with a matrix.
var A = [[1,2,3],
         [4,5,6],
         [7,3,9]];
A









    Out[2]:





[ [ 1, 2, 3 ], [ 4, 5, 6 ], [ 7, 3, 9 ] ]



In [3]:

    
// Let's also make a vector.
var x = [3,1,2];
x









    Out[3]:





[ 3, 1, 2 ]



In [4]:

    
// Matrix-vector product.
var b = n.dot(A, x);
b









    Out[4]:





[ 11, 29, 42 ]



In [5]:

    
// Matrix inverse.
var Ainv = n.inv(A);
Ainv









    Out[5]:





[ [ -0.9, 0.30000000000000004, 0.09999999999999998 ],
  [ -0.1999999999999999,
    0.39999999999999997,
    -0.19999999999999998 ],
  [ 0.7666666666666666, -0.3666666666666667, 0.09999999999999998 ] ]



In [6]:

    
// Let's check it:
n.dot(Ainv,b);









    Out[6]:





[ 3, 1.0000000000000018, 1.9999999999999982 ]



In [7]:

    
// Determinant
numeric.det(A);









    Out[7]:





-29.999999999999996



In [8]:

    
// Eigenvalues.
ev = numeric.eig(A)









    Out[8]:





{ lambda: 
   { x: [ 13.98887031606829, -1.0436794353124923, 2.054809119244198 ],
     y: undefined },
  E: { x: [ [Object], [Object], [Object] ], y: undefined } }

Hidden Tao

Need to try this too: https://github.com/hiddentao/linear-algebra



In [ ]: