normalize train data for Gaussian NB

Author: kortirso

CHANGELOG.md

@@ -7,6 +7,7 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 ## Unreleased
 ### Added
 - preprocessing normalization for data set
+- normalize train data for Gaussian NB
 
 ### Modified
 - normalization can be selected for KNN

lib/learn_kit/knn/classify.ex

@@ -31,7 +31,7 @@ defmodule LearnKit.Knn.Classify do
       defp check_normalization(data_set, options) do
         type = Keyword.get(options, :normalization)
         case type do
-          t when t in ["minimax", "z_normalization"] -> normalize(data_set, options, type)
+          t when t in ["minimax", "z_normalization"] -> normalize(data_set, type)
           _ -> data_set
         end
       end
@@ -72,7 +72,7 @@ defmodule LearnKit.Knn.Classify do
       end
 
       # normalize each feature
-      defp normalize(data_set, options, type) do
+      defp normalize(data_set, type) do
         coefficients = find_coefficients_for_normalization(data_set, type)
         Enum.map(data_set, fn {key, features} ->
           {

lib/learn_kit/naive_bayes/gaussian.ex

@@ -7,6 +7,7 @@ defmodule LearnKit.NaiveBayes.Gaussian do
 
   alias LearnKit.NaiveBayes.Gaussian
 
+  use Gaussian.Normalize
   use Gaussian.Fit
   use Gaussian.Classify
   use Gaussian.Score
@@ -78,6 +79,29 @@ defmodule LearnKit.NaiveBayes.Gaussian do
     %Gaussian{data_set: data_set}
   end
 
+  @doc """
+  Normalize train data
+
+  ## Parameters
+
+    - classifier: %LearnKit.NaiveBayes.Gaussian{}
+    - type: none/minimax/z_normalization, default is none, optional
+
+  ## Examples
+
+      iex> classifier = classifier |> LearnKit.NaiveBayes.Gaussian.normalize_train_data("minimax")
+      %LearnKit.NaiveBayes.Gaussian{
+        data_set: [a1: [[0.6666666666666666, 0.8], [1.0, 1.0]], b1: [[0.0, 0.0]]],
+        fit_data: []
+      }
+
+  """
+  @spec normalize_train_data(%Gaussian{data_set: data_set}, String.t()) :: %Gaussian{data_set: data_set, fit_data: fit_data}
+
+  def normalize_train_data(%Gaussian{data_set: data_set}, type \\ "none") when is_binary(type) do
+    %Gaussian{data_set: normalize_data(data_set, type), fit_data: []}
+  end
+
   @doc """
   Fit train data
 

lib/learn_kit/naive_bayes/gaussian/normalize.ex

@@ -0,0 +1,37 @@
+defmodule LearnKit.NaiveBayes.Gaussian.Normalize do
+  @moduledoc """
+  Module for fit functions
+  """
+
+  alias LearnKit.Preprocessing
+
+  defmacro __using__(_opts) do
+    quote do
+      defp normalize_data(data_set, type) do
+        case type do
+          t when t in ["minimax", "z_normalization"] -> normalize(data_set, type)
+          _ -> data_set
+        end
+      end
+
+      # normalize each feature
+      defp normalize(data_set, type) do
+        coefficients = find_coefficients_for_normalization(data_set, type)
+        Enum.map(data_set, fn {key, features} ->
+          {
+            key,
+            Enum.map(features, fn feature -> Preprocessing.normalize_feature(feature, coefficients, type) end)
+          }
+        end)
+      end
+
+      # find coefficients for normalization
+      defp find_coefficients_for_normalization(data_set, type) do
+        Enum.reduce(data_set, [], fn {_, features}, acc ->
+          Enum.reduce(features, acc, fn feature, acc -> [feature | acc] end)
+        end)
+        |> Preprocessing.coefficients(type)
+      end
+    end
+  end
+end

test/learn_kit/naive_bayes/gaussian_test.exs

@@ -14,27 +14,39 @@ defmodule LearnKit.NaiveBayes.GaussianTest do
   end
 
   test "add train data to classificator" do
-    %Gaussian{data_set: data_set} = Gaussian.new
-                    |> Gaussian.add_train_data({:a1, [1, 2]})
-                    |> Gaussian.add_train_data({:a1, [1, 3]})
-                    |> Gaussian.add_train_data({:b1, [2, 3]})
+    %Gaussian{data_set: data_set} =
+      Gaussian.new
+      |> Gaussian.add_train_data({:a1, [1, 2]})
+      |> Gaussian.add_train_data({:a1, [1, 3]})
+      |> Gaussian.add_train_data({:b1, [2, 3]})
 
     assert data_set == [b1: [[2, 3]], a1: [[1, 3], [1, 2]]]
   end
 
+  test "normalize data set", state do
+    %Gaussian{data_set: data_set} = state[:classificator] |> Gaussian.normalize_train_data("minimax")
+
+    assert data_set ==
+      [
+        label1: [[0.3333333333333333, 0.25], [0.16666666666666666, 0.25], [0.0, 0.0]],
+        label2: [ [0.6666666666666666, 0.75], [0.8333333333333334, 0.75], [1.0, 1.0], [0.16666666666666666, 0.0]]
+      ]
+  end
+
   test "fit data set", state do
     %Gaussian{fit_data: fit_data} = state[:classificator] |> Gaussian.fit
 
-    assert fit_data == [
-                        label1: [
-                                  %{mean: -2.0, standard_deviation: 0.816496580927726, variance: 0.6666666666666666},
-                                  %{mean: -1.3333333333333333, standard_deviation: 0.4714045207910317, variance: 0.2222222222222222}
-                                ],
-                        label2: [
-                                  %{mean: 1.0, standard_deviation: 1.8708286933869707, variance: 3.5},
-                                  %{mean: 0.5, standard_deviation: 1.5, variance: 2.25}
-                                ]
-                        ]
+    assert fit_data ==
+      [
+        label1: [
+                  %{mean: -2.0, standard_deviation: 0.816496580927726, variance: 0.6666666666666666},
+                  %{mean: -1.3333333333333333, standard_deviation: 0.4714045207910317, variance: 0.2222222222222222}
+                ],
+        label2: [
+                  %{mean: 1.0, standard_deviation: 1.8708286933869707, variance: 3.5},
+                  %{mean: 0.5, standard_deviation: 1.5, variance: 2.25}
+                ]
+      ]
   end
 
   test "return probability estimates for the feature", state do