Updating sample tutorial applications.

accord-net · Dec 8, 2014 · 7076f10 · 7076f10
1 parent fa1d92b
commit 7076f10
Show file tree

Hide file tree

Showing 32 changed files with 1,254 additions and 145 deletions.
diff --git a/...assification/Classification Sample.csproj → ...on/0. Introduction/0. Introduction.csproj b/...assification/Classification Sample.csproj → ...on/0. Introduction/0. Introduction.csproj
@@ -34,36 +34,37 @@
     <WarningLevel>4</WarningLevel>
   </PropertyGroup>
   <ItemGroup>
-    <Reference Include="Accord, Version=2.13.1.0, Culture=neutral, PublicKeyToken=fa1a88e29555ccf7, processorArchitecture=MSIL">
+    <Reference Include="Accord, Version=2.14.0.0, Culture=neutral, PublicKeyToken=fa1a88e29555ccf7, processorArchitecture=MSIL">
       <SpecificVersion>False</SpecificVersion>
-      <HintPath>..\packages\Accord.2.13.1\lib\net40\Accord.dll</HintPath>
+      <HintPath>..\packages\Accord.2.14.0\lib\net40\Accord.dll</HintPath>
     </Reference>
-    <Reference Include="Accord.Controls, Version=2.13.1.0, Culture=neutral, PublicKeyToken=fa1a88e29555ccf7, processorArchitecture=MSIL">
+    <Reference Include="Accord.Controls, Version=2.14.0.0, Culture=neutral, PublicKeyToken=fa1a88e29555ccf7, processorArchitecture=MSIL">
       <SpecificVersion>False</SpecificVersion>
-      <HintPath>..\packages\Accord.Controls.2.13.1\lib\net40\Accord.Controls.dll</HintPath>
+      <HintPath>..\packages\Accord.Controls.2.14.0\lib\net40\Accord.Controls.dll</HintPath>
     </Reference>
-    <Reference Include="Accord.Imaging, Version=2.13.1.0, Culture=neutral, PublicKeyToken=fa1a88e29555ccf7, processorArchitecture=MSIL">
+    <Reference Include="Accord.Imaging, Version=2.14.0.0, Culture=neutral, PublicKeyToken=fa1a88e29555ccf7, processorArchitecture=MSIL">
       <SpecificVersion>False</SpecificVersion>
-      <HintPath>..\packages\Accord.Imaging.2.13.1\lib\net40\Accord.Imaging.dll</HintPath>
+      <HintPath>..\packages\Accord.Imaging.2.14.0\lib\net40\Accord.Imaging.dll</HintPath>
     </Reference>
-    <Reference Include="Accord.IO">
-      <HintPath>..\packages\Accord.IO.2.13.1\lib\net40\Accord.IO.dll</HintPath>
+    <Reference Include="Accord.IO, Version=2.14.0.0, Culture=neutral, PublicKeyToken=fa1a88e29555ccf7, processorArchitecture=MSIL">
+      <SpecificVersion>False</SpecificVersion>
+      <HintPath>..\packages\Accord.IO.2.14.0\lib\net40\Accord.IO.dll</HintPath>
     </Reference>
-    <Reference Include="Accord.MachineLearning, Version=2.13.1.0, Culture=neutral, PublicKeyToken=fa1a88e29555ccf7, processorArchitecture=MSIL">
+    <Reference Include="Accord.MachineLearning, Version=2.14.0.0, Culture=neutral, PublicKeyToken=fa1a88e29555ccf7, processorArchitecture=MSIL">
       <SpecificVersion>False</SpecificVersion>
-      <HintPath>..\packages\Accord.MachineLearning.2.13.1\lib\net40\Accord.MachineLearning.dll</HintPath>
+      <HintPath>..\packages\Accord.MachineLearning.2.14.0\lib\net40\Accord.MachineLearning.dll</HintPath>
     </Reference>
-    <Reference Include="Accord.Math, Version=2.13.1.0, Culture=neutral, PublicKeyToken=fa1a88e29555ccf7, processorArchitecture=MSIL">
+    <Reference Include="Accord.Math, Version=2.14.0.0, Culture=neutral, PublicKeyToken=fa1a88e29555ccf7, processorArchitecture=MSIL">
       <SpecificVersion>False</SpecificVersion>
-      <HintPath>..\packages\Accord.Math.2.13.1\lib\net40\Accord.Math.dll</HintPath>
+      <HintPath>..\packages\Accord.Math.2.14.0\lib\net40\Accord.Math.dll</HintPath>
     </Reference>
-    <Reference Include="Accord.Neuro, Version=2.13.1.0, Culture=neutral, PublicKeyToken=fa1a88e29555ccf7, processorArchitecture=MSIL">
+    <Reference Include="Accord.Neuro, Version=2.14.0.0, Culture=neutral, PublicKeyToken=fa1a88e29555ccf7, processorArchitecture=MSIL">
       <SpecificVersion>False</SpecificVersion>
-      <HintPath>..\packages\Accord.Neuro.2.13.1\lib\net40\Accord.Neuro.dll</HintPath>
+      <HintPath>..\packages\Accord.Neuro.2.14.0\lib\net40\Accord.Neuro.dll</HintPath>
     </Reference>
-    <Reference Include="Accord.Statistics, Version=2.13.1.0, Culture=neutral, PublicKeyToken=fa1a88e29555ccf7, processorArchitecture=MSIL">
+    <Reference Include="Accord.Statistics, Version=2.14.0.0, Culture=neutral, PublicKeyToken=fa1a88e29555ccf7, processorArchitecture=MSIL">
       <SpecificVersion>False</SpecificVersion>
-      <HintPath>..\packages\Accord.Statistics.2.13.1\lib\net40\Accord.Statistics.dll</HintPath>
+      <HintPath>..\packages\Accord.Statistics.2.14.0\lib\net40\Accord.Statistics.dll</HintPath>
     </Reference>
     <Reference Include="AForge, Version=2.2.5.0, Culture=neutral, PublicKeyToken=c1db6ff4eaa06aeb, processorArchitecture=MSIL">
       <HintPath>..\packages\AForge.2.2.5\lib\AForge.dll</HintPath>

diff --git a/.../Classification/Classification/Program.cs → ...Classification/0. Introduction/Program.cs b/.../Classification/Classification/Program.cs → ...Classification/0. Introduction/Program.cs
@@ -1,4 +1,5 @@
-using System.Data;
+using System;
+using System.Data;
 using Accord.Controls;
 using Accord.IO;
 using Accord.MachineLearning.Bayes;
@@ -7,13 +8,12 @@
 using Accord.MachineLearning.VectorMachines;
 using Accord.MachineLearning.VectorMachines.Learning;
 using Accord.Math;
+using Accord.Neuro.Learning;
 using Accord.Statistics.Distributions.Univariate;
 using Accord.Statistics.Kernels;
 using Accord.Statistics.Models.Regression;
 using Accord.Statistics.Models.Regression.Fitting;
-using System;
 using AForge.Neuro;
-using Accord.Neuro.Learning;
 
 namespace ClassificationSample
 {
@@ -45,88 +45,56 @@ static void Main(string[] args)
             network(inputs, outputs);
         }
 
-        private static void network(double[][] inputs, int[] outputs)
+        private static void naiveBayes(double[][] inputs, int[] outputs)
         {
-            // Since we would like to learn binary outputs in the form
-            // [-1,+1], we can use a bipolar sigmoid activation function
-            IActivationFunction function = new BipolarSigmoidFunction();
-
-            // In our problem, we have 2 inputs (x, y pairs), and we will 
-            // be creating a network with 5 hidden neurons and 1 output:
-            //
-            var network = new ActivationNetwork(function,
-                inputsCount: 2, neuronsCount: new[] { 5, 1 });
-
-            // Create a Levenberg-Marquardt algorithm
-            var teacher = new LevenbergMarquardtLearning(network)
-            {
-                UseRegularization = true
-            };
+            // In our problem, we have 2 classes (samples can be either
+            // positive or negative), and 2 inputs (x and y coordinates).
 
+            var nb = new NaiveBayes<NormalDistribution>(classes: 2,
+                inputs: 2, initial: new NormalDistribution());
 
-            // Because the network is expecting multiple outputs,
-            // we have to convert our single variable into arrays
+            // The Naive Bayes expects the class labels to 
+            // range from 0 to k, so we convert -1 to be 0:
             //
-            var y = outputs.ToDouble().ToArray();
-
-            // Iterate until stop criteria is met
-            double error = double.PositiveInfinity;
-            double previous;
-
-            do
-            {
-                previous = error;
-
-                // Compute one learning iteration
-                error = teacher.RunEpoch(inputs, y);
-
-            } while (Math.Abs(previous - error) < 1e-10 * previous);
+            outputs = outputs.Apply(x => x < 0 ? 0 : x);
 
+            // Estimate the Naive Bayes
+            double error = nb.Estimate(inputs, outputs);
 
             // Classify the samples using the model
-            int[] answers = inputs.Apply(network.Compute).GetColumn(0).ToInt32();
+            int[] answers = inputs.Apply(nb.Compute);
 
             // Plot the results
             ScatterplotBox.Show("Expected results", inputs, outputs);
-            ScatterplotBox.Show("Network results", inputs, answers)
+            ScatterplotBox.Show("Naive Bayes results", inputs, answers)
                 .Hold();
         }
 
-        private static void logistic(double[][] inputs, int[] outputs)
+        private static void decisionTree(double[][] inputs, int[] outputs)
         {
-            // In our problem, we have 2 inputs (x, y pairs)
-            var logistic = new LogisticRegression(inputs: 2);
-
-            // Create a iterative re-weighted least squares algorithm
-            var teacher = new IterativeReweightedLeastSquares(logistic);
+            // In our problem, we have 2 classes (samples can be either
+            // positive or negative), and 2 continuous-valued inputs.
+            DecisionTree tree = new DecisionTree(inputs: new[] 
+            {
+                DecisionVariable.Continuous("X"),
+                DecisionVariable.Continuous("Y")
+            }, classes: 2);
 
+            C45Learning teacher = new C45Learning(tree);
 
-            // Logistic Regression expects the output labels 
-            // to range from 0 to k, so we convert -1 to be 0:
+            // The C4.5 algorithm expects the class labels to
+            // range from 0 to k, so we convert -1 to be zero:
             //
             outputs = outputs.Apply(x => x < 0 ? 0 : x);
 
-
-            // Iterate until stop criteria is met
-            double error = double.PositiveInfinity;
-            double previous;
-
-            do
-            {
-                previous = error;
-
-                // Compute one learning iteration
-                error = teacher.Run(inputs, outputs);
-
-            } while (Math.Abs(previous - error) < 1e-10 * previous);
-
+            double error = teacher.Run(inputs, outputs);
 
             // Classify the samples using the model
-            int[] answers = inputs.Apply(logistic.Compute).Apply(Math.Round).ToInt32();
+            int[] answers = inputs.Apply(tree.Compute);
 
             // Plot the results
             ScatterplotBox.Show("Expected results", inputs, outputs);
-            ScatterplotBox.Show("Logistic Regression results", inputs, answers)
+            ScatterplotBox.Show("Decision Tree results", inputs, answers)
                 .Hold();
         }
 
@@ -197,57 +165,90 @@ private static void kernelSvm(double[][] inputs, int[] outputs)
             ScatterplotBox.Show("Support vectors", sv).Hold();
         }
 
-        private static void decisionTree(double[][] inputs, int[] outputs)
+        private static void network(double[][] inputs, int[] outputs)
         {
-            // In our problem, we have 2 classes (samples can be either
-            // positive or negative), and 2 continuous-valued inputs.
-            DecisionTree tree = new DecisionTree(attributes: new[] 
+            // Since we would like to learn binary outputs in the form
+            // [-1,+1], we can use a bipolar sigmoid activation function
+            IActivationFunction function = new BipolarSigmoidFunction();
+
+            // In our problem, we have 2 inputs (x, y pairs), and we will 
+            // be creating a network with 5 hidden neurons and 1 output:
+            //
+            var network = new ActivationNetwork(function,
+                inputsCount: 2, neuronsCount: new[] { 5, 1 });
+
+            // Create a Levenberg-Marquardt algorithm
+            var teacher = new LevenbergMarquardtLearning(network)
             {
-                DecisionVariable.Continuous("X"),
-                DecisionVariable.Continuous("Y")
-            }, outputClasses: 2);
+                UseRegularization = true
+            };
 
-            C45Learning teacher = new C45Learning(tree);
 
-            // The C4.5 algorithm expects the class labels to
-            // range from 0 to k, so we convert -1 to be zero:
+            // Because the network is expecting multiple outputs,
+            // we have to convert our single variable into arrays
             //
-            outputs = outputs.Apply(x => x < 0 ? 0 : x);
+            var y = outputs.ToDouble().ToArray();
+
+            // Iterate until stop criteria is met
+            double error = double.PositiveInfinity;
+            double previous;
+
+            do
+            {
+                previous = error;
+
+                // Compute one learning iteration
+                error = teacher.RunEpoch(inputs, y);
+
+            } while (Math.Abs(previous - error) < 1e-10 * previous);
 
-            double error = teacher.Run(inputs, outputs);
 
             // Classify the samples using the model
-            int[] answers = inputs.Apply(tree.Compute);
+            int[] answers = inputs.Apply(network.Compute).GetColumn(0).ToInt32();
 
             // Plot the results
             ScatterplotBox.Show("Expected results", inputs, outputs);
-            ScatterplotBox.Show("Decision Tree results", inputs, answers)
+            ScatterplotBox.Show("Network results", inputs, answers)
                 .Hold();
         }
 
-        private static void naiveBayes(double[][] inputs, int[] outputs)
+        private static void logistic(double[][] inputs, int[] outputs)
         {
-            // In our problem, we have 2 classes (samples can be either
-            // positive or negative), and 2 inputs (x and y coordinates).
+            // In our problem, we have 2 inputs (x, y pairs)
+            var logistic = new LogisticRegression(inputs: 2);
 
-            var nb = new NaiveBayes<NormalDistribution>(classes: 2,
-                inputs: 2, prior: new NormalDistribution());
+            // Create a iterative re-weighted least squares algorithm
+            var teacher = new IterativeReweightedLeastSquares(logistic);
 
-            // The Naive Bayes expects the class labels to 
-            // range from 0 to k, so we convert -1 to be 0:
+
+            // Logistic Regression expects the output labels 
+            // to range from 0 to k, so we convert -1 to be 0:
             //
             outputs = outputs.Apply(x => x < 0 ? 0 : x);
 
-            // Estimate the Naive Bayes
-            double error = nb.Estimate(inputs, outputs);
+
+            // Iterate until stop criteria is met
+            double error = double.PositiveInfinity;
+            double previous;
+
+            do
+            {
+                previous = error;
+
+                // Compute one learning iteration
+                error = teacher.Run(inputs, outputs);
+
+            } while (Math.Abs(previous - error) < 1e-10 * previous);
+
 
             // Classify the samples using the model
-            int[] answers = inputs.Apply(nb.Compute);
+            int[] answers = inputs.Apply(logistic.Compute).Apply(Math.Round).ToInt32();
 
             // Plot the results
             ScatterplotBox.Show("Expected results", inputs, outputs);
-            ScatterplotBox.Show("Naive Bayes results", inputs, answers)
+            ScatterplotBox.Show("Logistic Regression results", inputs, answers)
                 .Hold();
         }
+
     }
 }
diff --git a/...Classification/Properties/AssemblyInfo.cs → .... Introduction/Properties/AssemblyInfo.cs b/...Classification/Properties/AssemblyInfo.cs → .... Introduction/Properties/AssemblyInfo.cs
diff --git a/...urces/Decision Tree results [on hold].png → ...urces/Decision Tree results [on hold].png b/...urces/Decision Tree results [on hold].png → ...urces/Decision Tree results [on hold].png
diff --git a/...sification/Resources/Expected results.png → ...troduction/Resources/Expected results.png b/...sification/Resources/Expected results.png → ...troduction/Resources/Expected results.png
diff --git a/...es/Gaussian support vectors [on hold].png → ...es/Gaussian support vectors [on hold].png b/...es/Gaussian support vectors [on hold].png → ...es/Gaussian support vectors [on hold].png
diff --git a/...ication/Resources/GaussianSVM results.png → ...duction/Resources/GaussianSVM results.png b/...ication/Resources/GaussianSVM results.png → ...duction/Resources/GaussianSVM results.png
diff --git a/...ification/Resources/LinearSVM results.png → ...roduction/Resources/LinearSVM results.png b/...ification/Resources/LinearSVM results.png → ...roduction/Resources/LinearSVM results.png
diff --git a/...sources/Naive Bayes results [on hold].png → ...sources/Naive Bayes results [on hold].png b/...sources/Naive Bayes results [on hold].png → ...sources/Naive Bayes results [on hold].png
diff --git a/...n/Resources/Support vectors [on hold].png → ...n/Resources/Support vectors [on hold].png b/...n/Resources/Support vectors [on hold].png → ...n/Resources/Support vectors [on hold].png
diff --git a/...fication/Resources/Yin-Yang [on hold].png → ...oduction/Resources/Yin-Yang [on hold].png b/...fication/Resources/Yin-Yang [on hold].png → ...oduction/Resources/Yin-Yang [on hold].png
diff --git a/...cation/Classification/Resources/graph.xls → ...ation/0. Introduction/Resources/graph.xls b/...cation/Classification/Resources/graph.xls → ...ation/0. Introduction/Resources/graph.xls
diff --git a/...ion/Classification/Resources/yin-yang.png → ...on/0. Introduction/Resources/yin-yang.png b/...ion/Classification/Resources/yin-yang.png → ...on/0. Introduction/Resources/yin-yang.png
diff --git a/Samples/Tutorials/Classification/0. Introduction/app.config b/Samples/Tutorials/Classification/0. Introduction/app.config
@@ -0,0 +1,31 @@
+<?xml version="1.0" encoding="utf-8"?>
+<configuration>
+  <runtime>
+    <assemblyBinding xmlns="urn:schemas-microsoft-com:asm.v1">
+      <dependentAssembly>
+        <assemblyIdentity name="AForge" publicKeyToken="c1db6ff4eaa06aeb" culture="neutral" />
+        <bindingRedirect oldVersion="0.0.0.0-2.2.5.0" newVersion="2.2.5.0" />
+      </dependentAssembly>
+      <dependentAssembly>
+        <assemblyIdentity name="AForge.Math" publicKeyToken="abba2e25397ee8c9" culture="neutral" />
+        <bindingRedirect oldVersion="0.0.0.0-2.2.5.0" newVersion="2.2.5.0" />
+      </dependentAssembly>
+      <dependentAssembly>
+        <assemblyIdentity name="Accord" publicKeyToken="fa1a88e29555ccf7" culture="neutral" />
+        <bindingRedirect oldVersion="0.0.0.0-2.14.0.0" newVersion="2.14.0.0" />
+      </dependentAssembly>
+      <dependentAssembly>
+        <assemblyIdentity name="Accord.Math" publicKeyToken="fa1a88e29555ccf7" culture="neutral" />
+        <bindingRedirect oldVersion="0.0.0.0-2.14.0.0" newVersion="2.14.0.0" />
+      </dependentAssembly>
+      <dependentAssembly>
+        <assemblyIdentity name="Accord.MachineLearning" publicKeyToken="fa1a88e29555ccf7" culture="neutral" />
+        <bindingRedirect oldVersion="0.0.0.0-2.14.0.0" newVersion="2.14.0.0" />
+      </dependentAssembly>
+      <dependentAssembly>
+        <assemblyIdentity name="Accord.Statistics" publicKeyToken="fa1a88e29555ccf7" culture="neutral" />
+        <bindingRedirect oldVersion="0.0.0.0-2.14.0.0" newVersion="2.14.0.0" />
+      </dependentAssembly>
+    </assemblyBinding>
+  </runtime>
+</configuration>
diff --git a/...lassification/Classification/examples.xls → ...assification/0. Introduction/examples.xls b/...lassification/Classification/examples.xls → ...assification/0. Introduction/examples.xls
diff --git a/...sification/Classification/packages.config → ...ification/0. Introduction/packages.config b/...sification/Classification/packages.config → ...ification/0. Introduction/packages.config
@@ -1,13 +1,13 @@
 <?xml version="1.0" encoding="utf-8"?>
 <packages>
-  <package id="Accord" version="2.13.1" targetFramework="net40-Client" />
-  <package id="Accord.Controls" version="2.13.1" targetFramework="net40-Client" />
-  <package id="Accord.Imaging" version="2.13.1" targetFramework="net40-Client" />
-  <package id="Accord.IO" version="2.13.1" targetFramework="net40-Client" />
-  <package id="Accord.MachineLearning" version="2.13.1" targetFramework="net40-Client" />
-  <package id="Accord.Math" version="2.13.1" targetFramework="net40-Client" />
-  <package id="Accord.Neuro" version="2.13.1" targetFramework="net40-Client" />
-  <package id="Accord.Statistics" version="2.13.1" targetFramework="net40-Client" />
+  <package id="Accord" version="2.14.0" targetFramework="net40-Client" />
+  <package id="Accord.Controls" version="2.14.0" targetFramework="net40-Client" />
+  <package id="Accord.Imaging" version="2.14.0" targetFramework="net40-Client" />
+  <package id="Accord.IO" version="2.14.0" targetFramework="net40-Client" />
+  <package id="Accord.MachineLearning" version="2.14.0" targetFramework="net40-Client" />
+  <package id="Accord.Math" version="2.14.0" targetFramework="net40-Client" />
+  <package id="Accord.Neuro" version="2.14.0" targetFramework="net40-Client" />
+  <package id="Accord.Statistics" version="2.14.0" targetFramework="net40-Client" />
   <package id="AForge" version="2.2.5" targetFramework="net40-Client" />
   <package id="AForge.Genetic" version="2.2.5" targetFramework="net40-Client" />
   <package id="AForge.Imaging" version="2.2.5" targetFramework="net40-Client" />