Miscellaneous refactor

README.md

@@ -2,15 +2,15 @@
 
 # Numerix
 
-A collection of (potentially) useful mathematical functions. At the moment it has a number of distance and correlation functions. The plan is to implement other special functions, statistics, probability distributions, maybe even machine learning algorithms.
+A collection of useful mathematical in Elixir, with a slant towards statistics, linear algebra and machine learning.
 
 ## Installation
 
 Add `numerix` to your list of dependencies in `mix.exs`:
 
 ```elixir
   def deps do
-    [{:numerix, "~> 0.0.6"}]
+    [{:numerix, "~> 0.1.0"}]
   end
 ```
 
@@ -32,6 +32,26 @@ Check out the [API reference](https://hexdocs.pm/numerix/api-reference.html) for
 
 ## Features
 
+### Statistics
+
+* Mean
+* Weighted mean
+* Median
+* Mode
+* Range
+* Variance
+* Population variance
+* Standard deviation
+* Population standard deviation
+* Moment
+* Kurtosis
+* Skewness
+* Covariance
+* Weighted covariance
+* Population covariance
+* Quantile
+* Percentile
+
 ### Correlation functions
 
 * Pearson
@@ -54,22 +74,18 @@ Check out the [API reference](https://hexdocs.pm/numerix/api-reference.html) for
 * Logit
 * Logistic
 
-### Statistics
+### Window functions
 
-* Mean
-* Weighted mean
-* Median
-* Mode
-* Range
-* Variance
-* Population variance
-* Standard deviation
-* Population standard deviation
-* Moment
-* Kurtosis
-* Skewness
-* Covariance
-* Weighted covariance
-* Population covariance
-* Quantile
-* Percentile
+* Gaussian
+
+### Linear algebra
+
+* Dot product
+* L1-norm
+* L2-norm
+* p-norm
+* Vector subtraction and multiplication
+
+### Kernel functions
+
+* RBF

lib/correlation.ex

@@ -1,10 +1,11 @@
 defmodule Numerix.Correlation do
-  alias Numerix.{Common, Statistics}
-
   @moduledoc """
   Statistical correlation functions between two vectors.
   """
 
+  import Numerix.LinearAlgebra
+  alias Numerix.{Common, Statistics}
+
   @doc """
   Calculates the Pearson correlation coefficient between two vectors.
   """
@@ -14,21 +15,15 @@ defmodule Numerix.Correlation do
   def pearson(vector1, vector2) do
     sum1 = vector1 |> Enum.sum
     sum2 = vector2 |> Enum.sum
-
     sum_of_squares1 = vector1 |> square |> Enum.sum
     sum_of_squares2 = vector2 |> square |> Enum.sum
-
-    sum_of_products =
-      vector1
-      |> Stream.zip(vector2)
-      |> Stream.map(fn {x, y} -> x * y end)
-      |> Enum.sum
+    sum_of_products = vector1 |> dot_product(vector2)
 
     size = length(vector1)
     num = sum_of_products - (sum1 * sum2 / size)
-    density = :math.sqrt(
-      (sum_of_squares1 - :math.pow(sum1, 2) / size)
-      * (sum_of_squares2 - :math.pow(sum2, 2) / size))
+    density = (sum_of_squares1 - :math.pow(sum1, 2) / size)
+      |> Kernel.*(sum_of_squares2 - :math.pow(sum2, 2) / size)
+      |> :math.sqrt
 
     case density do
       0.0 -> 0.0

lib/statistics.ex

@@ -232,7 +232,7 @@ defmodule Numerix.Statistics do
   defp sum_powered_deviations(xs, n) do
     x_mean = mean(xs)
     xs
-    |> Enum.map(fn x -> :math.pow(x - x_mean, n) end)
+    |> Stream.map(fn x -> :math.pow(x - x_mean, n) end)
     |> Enum.sum
   end