modify documentation, modify responses

Author: kortirso

lib/learn_kit/knn.ex

@@ -90,10 +90,10 @@ defmodule LearnKit.Knn do
   ## Examples
 
       iex> classificator |> LearnKit.Knn.classify([feature: [-1, -2], k: 3, weight: "distance"])
-      {:ok, "Feature has label a1 with weight 1.5"}
+      {:ok, :a1}
 
   """
-  @spec classify(%LearnKit.Knn{data_set: data_set}, [tuple]) :: {atom, String.t}
+  @spec classify(%LearnKit.Knn{data_set: data_set}, [tuple]) :: {:ok, label}
 
   def classify(%Knn{data_set: data_set}, options \\ []) do
     try do
@@ -103,8 +103,8 @@ defmodule LearnKit.Knn do
       # modification of options
       options = Keyword.merge([k: 3, algorithm: "brute", weight: "uniform"], options)
       # prediction
-      {label, weight} = prediction(data_set, options)
-      {:ok, "Feature has label #{label} with weight #{weight}"}
+      {label, _} = prediction(data_set, options)
+      {:ok, label}
     rescue
       error -> {:error, error.message}
     end

lib/learn_kit/naive_bayes/gaussian.ex

@@ -116,14 +116,14 @@ defmodule LearnKit.NaiveBayes.Gaussian do
   ## Examples
 
       iex> classificator |> LearnKit.NaiveBayes.Gaussian.predict_proba([1, 2])
-      [a1: 0.0359, a2: 0.0039]
+      {:ok, [a1: 0.0359, a2: 0.0039]}
 
   """
-  @spec predict_proba(%LearnKit.NaiveBayes.Gaussian{fit_data: fit_data}, feature) :: predictions
+  @spec predict_proba(%LearnKit.NaiveBayes.Gaussian{fit_data: fit_data}, feature) :: {:ok, predictions}
 
   def predict_proba(%Gaussian{fit_data: fit_data}, feature) do
-    fit_data
-    |> classify_data(feature)
+    result = fit_data |> classify_data(feature)
+    {:ok, result}
   end
 
   @doc """
@@ -136,15 +136,14 @@ defmodule LearnKit.NaiveBayes.Gaussian do
   ## Examples
 
       iex> classificator |> LearnKit.NaiveBayes.Gaussian.predict([1, 2])
+      {:ok, {:a1, 0.334545454}}
 
   """
-  @spec predict(%LearnKit.NaiveBayes.Gaussian{fit_data: fit_data}, feature) :: prediction
+  @spec predict(%LearnKit.NaiveBayes.Gaussian{fit_data: fit_data}, feature) :: {:ok, prediction}
 
   def predict(%Gaussian{fit_data: fit_data}, feature) do
-    fit_data
-    |> classify_data(feature)
-    |> Enum.sort_by(&(elem(&1, 1)))
-    |> Enum.at(-1)
+    result = fit_data |> classify_data(feature) |> Enum.sort_by(&(elem(&1, 1))) |> Enum.at(-1)
+    {:ok, result}
   end
 
   @doc """
@@ -159,13 +158,13 @@ defmodule LearnKit.NaiveBayes.Gaussian do
       iex> classificator = LearnKit.NaiveBayes.Gaussian.new([{:label1, [[-1, -1], [-2, -1], [-3, -2]]}, {:label2, [[1, 1], [2, 1], [3, 2], [-2, -2]]}])
       iex> classificator = classificator |> LearnKit.NaiveBayes.Gaussian.fit
       iex> classificator |> LearnKit.NaiveBayes.Gaussian.score
-      0.857143
+      {:ok, 0.857143}
 
   """
-  @spec score(%LearnKit.NaiveBayes.Gaussian{data_set: data_set, fit_data: fit_data}) :: number
+  @spec score(%LearnKit.NaiveBayes.Gaussian{data_set: data_set, fit_data: fit_data}) :: {:ok, number}
 
   def score(%Gaussian{data_set: data_set, fit_data: fit_data}) do
-    fit_data
-    |> calc_score(data_set)
+    result = fit_data |> calc_score(data_set)
+    {:ok, result}
   end
 end

lib/learn_kit/naive_bayes/gaussian/classify.ex

@@ -19,6 +19,7 @@ defmodule LearnKit.NaiveBayes.Gaussian.Classify do
         class_fraction = 1.0 / labels_count
         feature_bayes = feature_mult(feature, fit_results, 1.0, 0)
         feature_bayes * class_fraction
+        |> Float.round(10)
       end
 
       # multiply together the feature probabilities for all of the features in a label for given values

lib/learn_kit/naive_bayes/gaussian/score.ex

@@ -28,7 +28,7 @@ defmodule LearnKit.NaiveBayes.Gaussian.Score do
       end
 
       defp check_feature(feature, fit_data, label) do
-        {predicted_label, _} = Gaussian.predict(%Gaussian{fit_data: fit_data}, feature)
+        {:ok, {predicted_label, _}} = Gaussian.predict(%Gaussian{fit_data: fit_data}, feature)
         if predicted_label == label, do: 1, else: 0
       end
     end

test/learn_kit/knn_test.exs

@@ -28,6 +28,6 @@ defmodule LearnKit.KnnTest do
                     |> LearnKit.Knn.add_train_data({:a2, [3, 2]})
                     |> LearnKit.Knn.add_train_data({:a2, [-2, -2]})
 
-    assert {:ok, "Feature has label a1 with weight 1.5"} = LearnKit.Knn.classify(classificator, [feature: [-1, -2], k: 3, weight: "distance"])
+    assert {:ok, :a1} = LearnKit.Knn.classify(classificator, [feature: [-1, -2], k: 3, weight: "distance"])
   end
 end

test/learn_kit/naive_bayes/gaussian_test.exs

@@ -1,20 +1,60 @@
 defmodule LearnKit.NaiveBayes.GaussianTest do
   use ExUnit.Case
 
-  alias LearnKit.{NaiveBayes}
+  alias LearnKit.NaiveBayes.Gaussian
 
   test "create new knn classificator with empty data set" do
-    assert %NaiveBayes.Gaussian{data_set: data_set} = NaiveBayes.Gaussian.new
+    assert %Gaussian{data_set: data_set} = Gaussian.new
 
     assert data_set == []
   end
 
   test "add train data to classificator" do
-    %NaiveBayes.Gaussian{data_set: data_set} = NaiveBayes.Gaussian.new
-                    |> NaiveBayes.Gaussian.add_train_data({:a1, [1, 2]})
-                    |> NaiveBayes.Gaussian.add_train_data({:a1, [1, 3]})
-                    |> NaiveBayes.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 "fit data set" do
+    classificator = LearnKit.NaiveBayes.Gaussian.new([{:label1, [[-1, -1], [-2, -1], [-3, -2]]}, {:label2, [[1, 1], [2, 1], [3, 2], [-2, -2]]}])
+    %LearnKit.NaiveBayes.Gaussian{fit_data: fit_data} = classificator |> LearnKit.NaiveBayes.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}
+                                ]
+                        ]
+  end
+
+  test "return probability estimates for the feature" do
+    classificator = LearnKit.NaiveBayes.Gaussian.new([{:label1, [[-1, -1], [-2, -1], [-3, -2]]}, {:label2, [[1, 1], [2, 1], [3, 2], [-2, -2]]}])
+    classificator = classificator |> LearnKit.NaiveBayes.Gaussian.fit
+
+    assert {:ok, result} = classificator |> LearnKit.NaiveBayes.Gaussian.predict_proba([1, 2])
+    assert result == [label1: 0.0, label2: 0.017199571]
+  end
+
+  test "return exact prediction for the feature" do
+    classificator = LearnKit.NaiveBayes.Gaussian.new([{:label1, [[-1, -1], [-2, -1], [-3, -2]]}, {:label2, [[1, 1], [2, 1], [3, 2], [-2, -2]]}])
+    classificator = classificator |> LearnKit.NaiveBayes.Gaussian.fit
+
+    assert {:ok, result} = classificator |> LearnKit.NaiveBayes.Gaussian.predict([1, 2])
+    assert result == {:label2, 0.017199571}
+  end
+
+  test "returns the mean accuracy on the given test data and labels" do
+    classificator = LearnKit.NaiveBayes.Gaussian.new([{:label1, [[-1, -1], [-2, -1], [-3, -2]]}, {:label2, [[1, 1], [2, 1], [3, 2], [-2, -2]]}])
+    classificator = classificator |> LearnKit.NaiveBayes.Gaussian.fit
+
+    assert {:ok, result} = classificator |> LearnKit.NaiveBayes.Gaussian.score
+    assert result == 0.857143
+  end
 end