clojure-math

Incomplete bindings to Commons Math, EJML, and other math-related bits of code, to make up for the lack of NumPy and SciPy in Clojure.

API documentation

Usage

Vector algebra

Supported vector operations are: plus, minus, scale, norm, normalize, dist, dot, and cross.

Currently supported vector representations: any clojure.lang.Seqable, double arrays, Vector2D from Vector3D from Commons Math 3, DenseMatrix64F from EJML.

To create Commons Math's vectors from Clojure sequences, use to-vector:

(use 'clojure.math.geometry)
(use '[clojure.math.internal.cmath3 :only [to-vector]])

(let [x (to-vector [1 2 3])
      y (to-vector [4 5 6])]
  (cross x y))

Using Commons Math's vector constructors directly results in faster operations, but you don't get a Seqable instance.

Convolution and filtering

For discrete convolution of one-dimensonal sequences:

user=> (use 'clojure.math.signal)
nil
user=> (conv [1 2 3 4 5] [0.25 0.5 0.25] :valid)
[2.0 3.0 4.0]

The third parameter to conv is optional; acceptable values are: :full (default), :same, and :valid. The semantics is the same as for mode parameter of numpy.convolve.

Currently supported data types are: double arrays and vectors.

For median filter use medfilt:

user=> (use 'clojure.math.signal)
nil
user=> (medfilt [0 0 1 0 0 0 2 0 3 0 0] 5)
(0 0 0 0 0 0 0 0 0 0 0)

Geometric fitting

Fitting lines using total least squares

user=> (use 'clojure.math.geometry.fit)
nil
user=> (fit-line [[0 1] [1 2] [2 3.1] [4 5]])
{:point [1.75 2.775], :direction [-0.7059474712429025 -0.7082641935363889]}

License