/// <summary>
///
/// </summary>
/// <param name="problem"></param>
/// <param name="parameter"></param>
/// <param name="nFolds"></param>
/// <param name="target"></param>
public static void CrossValidation(SVMProblem problem, SVMParameter parameter, int nFolds, out double[] target)
{
if (problem == null)
{
throw new ArgumentNullException("problem");
}
if (parameter == null)
{
throw new ArgumentNullException("parameter");
}
if (nFolds < 2)
{
throw new ArgumentOutOfRangeException("nFolds");
}
IntPtr ptr_target = Marshal.AllocHGlobal(Marshal.SizeOf(typeof(double)) * problem.Length);
IntPtr ptr_problem = SVMProblem.Allocate(problem);
IntPtr ptr_parameter = SVMParameter.Allocate(parameter);
libsvm.svm_cross_validation(ptr_problem, ptr_parameter, nFolds, ptr_target);
target = new double[problem.Length];
Marshal.Copy(ptr_target, target, 0, target.Length);
SVMProblem.Free(ptr_problem);
SVMParameter.Free(ptr_parameter);
Marshal.FreeHGlobal(ptr_target);
ptr_target = IntPtr.Zero;
}
/*public Classification(string training, string test)
{
problem = SVMProblemHelper.Load(training);
testProblem = SVMProblemHelper.Load(test);
}*/
public void ClasificationHand()
{
System.Diagnostics.Debug.WriteLine("comienza a clasificar");
parameter = new SVMParameter();
parameter.Type = SVMType.C_SVC;
parameter.Kernel = SVMKernelType.RBF;
parameter.C = 1;
parameter.Gamma = 1;
System.Diagnostics.Debug.WriteLine("cargo los valores");
// System.Windows.MessageBox.Show("Algo real");
SVMModel model = SVM.Train(problem, parameter);
//System.Windows.MessageBox.Show("Oscar");
System.Diagnostics.Debug.WriteLine("creo el modelo");
double[] target;
target = new double[testProblem.Length];
int a = 3;
for (int i = 0; i < testProblem.Length; i++)
target[i] = SVM.Predict(model, testProblem.X[i]);
System.Diagnostics.Debug.WriteLine("aqui ya clasifico");
double accuracy = SVMHelper.EvaluateClassificationProblem(testProblem, target);
System.Diagnostics.Debug.WriteLine("termino");
}
/// <summary>
///
/// </summary>
/// <param name="problem"></param>
/// <param name="parameter"></param>
/// <returns></returns>
public static string CheckParameter(SVMProblem problem, SVMParameter parameter)
{
if (problem == null)
{
throw new ArgumentNullException("problem");
}
if (parameter == null)
{
throw new ArgumentNullException("parameter");
}
IntPtr ptr_problem = SVMProblem.Allocate(problem);
IntPtr ptr_parameter = SVMParameter.Allocate(parameter);
IntPtr ptr_output = libsvm.svm_check_parameter(ptr_problem, ptr_parameter);
SVMProblem.Free(ptr_problem);
SVMParameter.Free(ptr_parameter);
string output = Marshal.PtrToStringAnsi(ptr_output);
Marshal.FreeHGlobal(ptr_output);
ptr_output = IntPtr.Zero;
return(output);
}
static void Main(string[] args)
{
// Load the datasets: In this example I use the same datasets for training and testing which is not suggested
SVMProblem trainingSet = SVMProblemHelper.Load(@"C:\Users\lozzr\Documents\Visual Studio 2015\Projects\MM\MM\Dataset\rt-1.txt");
SVMProblem testSet = SVMProblemHelper.Load(@"C:\Users\lozzr\Documents\Visual Studio 2015\Projects\MM\MM\Dataset\rt-1.txt");
// Normalize the datasets if you want: L2 Norm => x / ||x||
trainingSet = trainingSet.Normalize(SVMNormType.L2);
testSet = testSet.Normalize(SVMNormType.L2);
// Select the parameter set
SVMParameter parameter = new SVMParameter();
parameter.Type = SVMType.ONE_CLASS;
parameter.Kernel = SVMKernelType.RBF;
parameter.C = 8.0;
parameter.Gamma = 0.0078125;
// Do cross validation to check this parameter set is correct for the dataset or not
double[] crossValidationResults; // output labels
int nFold = 5;
trainingSet.CrossValidation(parameter, nFold, out crossValidationResults);
// Evaluate the cross validation result
// If it is not good enough, select the parameter set again
double crossValidationAccuracy = trainingSet.EvaluateClassificationProblem(crossValidationResults);
// Train the model, If your parameter set gives good result on cross validation
SVMModel model = trainingSet.Train(parameter);
// Save the model
SVM.SaveModel(model, @"C:\Users\lozzr\Documents\Visual Studio 2015\Projects\MM\MM\Model\rt-1_model.txt");
// Predict the instances in the test set
double[] testResults = testSet.Predict(model);
// Evaluate the test results
// int[,] confusionMatrix;
//double testAccuracy = testSet.EvaluateClassificationProblem(testResults, model.Labels, out confusionMatrix);
// Print the resutls
Console.WriteLine("\n\nCross validation accuracy: " + crossValidationAccuracy);
// Console.WriteLine("\nTest accuracy: " + testAccuracy);
//Console.WriteLine("\nConfusion matrix:\n");
// Print formatted confusion matrix
//Console.Write(String.Format("{0,6}", ""));
/*for (int i = 0; i < model.Labels.Length; i++)
Console.Write(String.Format("{0,5}", "(" + model.Labels[i] + ")"));
Console.WriteLine();
/* for (int i = 0; i < confusionMatrix.GetLength(0); i++)
{
Console.Write(String.Format("{0,5}", "(" + model.Labels[i] + ")"));
for (int j = 0; j < confusionMatrix.GetLength(1); j++)
Console.Write(String.Format("{0,5}", confusionMatrix[i, j]));
Console.WriteLine();
}*/
Console.WriteLine("\n\nPress any key to quit...");
Console.ReadLine();
}
public static SVMParameter Convert(svm_parameter x)
{
SVMParameter y = new SVMParameter();
y.Type = (SVMType)x.svm_type;
y.Kernel = (SVMKernelType)x.kernel_type;
y.Degree = x.degree;
y.Gamma = x.gamma;
y.Coef0 = x.coef0;
y.CacheSize = x.cache_size;
y.Eps = x.eps;
y.C = x.C;
y.Nu = x.nu;
y.P = x.p;
y.Shrinking = x.shrinking != 0;
y.Probability = x.probability != 0;
int length = x.nr_weight;
y.WeightLabels = new int[length];
if (length > 0)
{
Marshal.Copy(x.weight_label, y.WeightLabels, 0, length);
}
y.Weights = new double[length];
if (length > 0)
{
Marshal.Copy(x.weight, y.Weights, 0, length);
}
return(y);
}
public static SVMParameter Convert(IntPtr ptr)
{
if (ptr == IntPtr.Zero)
{
return(null);
}
svm_parameter x = (svm_parameter)Marshal.PtrToStructure(ptr, typeof(svm_parameter));
return(SVMParameter.Convert(x));
}
public PredictionResult(double[,] ConfusionMatrix, List<double> Recalls, List<double> Precisions, List<double> Fscores, SVMParameter SVMParam, List<Feature> Features, List<int> Answers, List<int> Guesses)
{
confusionMatrix = ConfusionMatrix;
recalls = Recalls;
precisions = Precisions;
fscores = Fscores;
svmParams = SVMParam;
features = Features;
guesses = Guesses;
correct = Answers;
}
/// <summary>
/// This function constructs and returns an SVM model according to the given training data and parameters.
/// </summary>
/// <param name="problem">Training data.</param>
/// <param name="parameter">Parameter set.</param>
/// <returns>SVM model according to the given training data and parameters.</returns>
public static SVMModel Train(SVMProblem problem, SVMParameter parameter)
{
IntPtr ptr_problem = SVMProblem.Allocate(problem);
IntPtr ptr_parameter = SVMParameter.Allocate(parameter);
IntPtr ptr_model = libsvm.svm_train(ptr_problem, ptr_parameter);
SVMModel model = SVMModel.Convert(ptr_model);
SVMProblem.Free(ptr_problem);
SVMParameter.Free(ptr_parameter);
libsvm.svm_free_model_content(ptr_model);
return model;
}
/// <summary>
/// Create a model from the training data, has to be called before predicting
/// </summary>
/// <param name="paramater">The SvmParameter is converted to OneClass type</param>
/// <returns>returns true if the model was created correctly</returns>
public bool CreateModel(SVMParameter parameter)
{
parameter.Type = SVMType.ONE_CLASS;
try
{
_model = _trainingData.Train(parameter);
return true;
}
catch
{
return false;
}
}
/// <summary>
/// This function constructs and returns an SVM model according to the given training data and parameters.
/// </summary>
/// <param name="problem">Training data.</param>
/// <param name="parameter">Parameter set.</param>
/// <returns>SVM model according to the given training data and parameters.</returns>
public static SVMModel Train(SVMProblem problem, SVMParameter parameter)
{
IntPtr ptr_problem = SVMProblem.Allocate(problem);
IntPtr ptr_parameter = SVMParameter.Allocate(parameter);
IntPtr ptr_model = libsvm.svm_train(ptr_problem, ptr_parameter);
SVMModel model = SVMModel.Convert(ptr_model);
SVMProblem.Free(ptr_problem);
SVMParameter.Free(ptr_parameter);
libsvm.svm_free_model_content(ptr_model);
return(model);
}
public static void Free(svm_model x)
{
Marshal.FreeHGlobal(x.rho);
x.rho = IntPtr.Zero;
Marshal.FreeHGlobal(x.probA);
x.probA = IntPtr.Zero;
Marshal.FreeHGlobal(x.probB);
x.probB = IntPtr.Zero;
Marshal.FreeHGlobal(x.sv_indices);
x.sv_indices = IntPtr.Zero;
Marshal.FreeHGlobal(x.label);
x.label = IntPtr.Zero;
Marshal.FreeHGlobal(x.nSV);
x.nSV = IntPtr.Zero;
SVMParameter.Free(x.param);
IntPtr i_ptr_sv = x.SV;
for (int i = 0; i < x.l; i++)
{
IntPtr ptr_nodes = (IntPtr)Marshal.PtrToStructure(i_ptr_sv, typeof(IntPtr));
SVMNode.Free(ptr_nodes);
i_ptr_sv = IntPtr.Add(i_ptr_sv, Marshal.SizeOf(typeof(IntPtr)));
}
Marshal.FreeHGlobal(x.SV);
x.SV = IntPtr.Zero;
IntPtr i_ptr_svcoef = x.sv_coef;
for (int i = 0; i < x.nr_class - 1; i++)
{
IntPtr temp = (IntPtr)Marshal.PtrToStructure(i_ptr_svcoef, typeof(IntPtr));
Marshal.FreeHGlobal(temp);
temp = IntPtr.Zero;
i_ptr_svcoef = IntPtr.Add(i_ptr_svcoef, Marshal.SizeOf(typeof(IntPtr)));
}
Marshal.FreeHGlobal(x.sv_coef);
x.sv_coef = IntPtr.Zero;
}
public static void Free(IntPtr ptr)
{
if (ptr == IntPtr.Zero)
{
return;
}
svm_parameter x = (svm_parameter)Marshal.PtrToStructure(ptr, typeof(svm_parameter));
SVMParameter.Free(x);
Marshal.DestroyStructure(ptr, typeof(svm_parameter));
Marshal.FreeHGlobal(ptr);
ptr = IntPtr.Zero;
}
public SVMModel()
{
Parameter = new SVMParameter();
ClassCount = 0;
TotalSVCount = 0;
SV = null;
SVCoefs = null;
Rho = null;
ProbabilityA = null;
ProbabilityB = null;
SVIndices = null;
Labels = null;
SVCounts = null;
Creation = CreationType.LOAD_MODEL;
}
public SVMModel()
{
Parameter = new SVMParameter();
ClassCount = 0;
TotalSVCount = 0;
SV = null;
SVCoefs = null;
Rho = null;
ProbabilityA = null;
ProbabilityB = null;
SVIndices = null;
Labels = null;
SVCounts = null;
Creation = CreationType.LOAD_MODEL;
}
/// <summary>
///
/// </summary>
/// <param name="problem"></param>
/// <param name="parameter"></param>
/// <param name="nFolds"></param>
/// <param name="target"></param>
public static void CrossValidation(SVMProblem problem, SVMParameter parameter, int nFolds, out double[] target)
{
IntPtr ptr_target = Marshal.AllocHGlobal(Marshal.SizeOf(typeof(double)) * problem.Length);
IntPtr ptr_problem = SVMProblem.Allocate(problem);
IntPtr ptr_parameter = SVMParameter.Allocate(parameter);
libsvm.svm_cross_validation(ptr_problem, ptr_parameter, nFolds, ptr_target);
target = new double[problem.Length];
Marshal.Copy(ptr_target, target, 0, target.Length);
SVMProblem.Free(ptr_problem);
SVMParameter.Free(ptr_parameter);
Marshal.FreeHGlobal(ptr_target);
ptr_target = IntPtr.Zero;
}
/// <summary>
///
/// </summary>
/// <param name="problem"></param>
/// <param name="parameter"></param>
/// <returns></returns>
public static string CheckParameter(SVMProblem problem, SVMParameter parameter)
{
IntPtr ptr_problem = SVMProblem.Allocate(problem);
IntPtr ptr_parameter = SVMParameter.Allocate(parameter);
IntPtr ptr_output = libsvm.svm_check_parameter(ptr_problem, ptr_parameter);
SVMProblem.Free(ptr_problem);
SVMParameter.Free(ptr_parameter);
string output = Marshal.PtrToStringAnsi(ptr_output);
Marshal.FreeHGlobal(ptr_output);
ptr_output = IntPtr.Zero;
return output;
}
/// <summary>
///
/// </summary>
/// <param name="problem"></param>
/// <param name="parameter"></param>
/// <returns></returns>
public static string CheckParameter(SVMProblem problem, SVMParameter parameter)
{
IntPtr ptr_problem = SVMProblem.Allocate(problem);
IntPtr ptr_parameter = SVMParameter.Allocate(parameter);
IntPtr ptr_output = libsvm.svm_check_parameter(ptr_problem, ptr_parameter);
SVMProblem.Free(ptr_problem);
SVMParameter.Free(ptr_parameter);
string output = Marshal.PtrToStringAnsi(ptr_output);
Marshal.FreeHGlobal(ptr_output);
ptr_output = IntPtr.Zero;
return(output);
}
public static IntPtr Allocate(SVMParameter x)
{
if (x == null)
{
return(IntPtr.Zero);
}
svm_parameter y = new svm_parameter();
y.svm_type = (int)x.Type;
y.kernel_type = (int)x.Kernel;
y.degree = x.Degree;
y.gamma = x.Gamma;
y.coef0 = x.Coef0;
y.cache_size = x.CacheSize;
y.eps = x.Eps;
y.C = x.C;
y.nu = x.Nu;
y.p = x.P;
y.shrinking = x.Shrinking ? 1 : 0;
y.probability = x.Probability ? 1 : 0;
y.nr_weight = x.WeightLabels.Length;
y.weight_label = IntPtr.Zero;
if (y.nr_weight > 0)
{
y.weight_label = Marshal.AllocHGlobal(Marshal.SizeOf(typeof(int)) * x.WeightLabels.Length);
Marshal.Copy(x.WeightLabels, 0, y.weight_label, x.WeightLabels.Length);
}
y.weight = IntPtr.Zero;
if (y.nr_weight > 0)
{
y.weight = Marshal.AllocHGlobal(Marshal.SizeOf(typeof(double)) * x.Weights.Length);
Marshal.Copy(x.Weights, 0, y.weight, x.Weights.Length);
}
int size = Marshal.SizeOf(y);
IntPtr ptr = Marshal.AllocHGlobal(size);
Marshal.StructureToPtr(y, ptr, true);
return(ptr);
}
public SVMParameter Clone()
{
SVMParameter y = new SVMParameter();
y.Type = Type;
y.Kernel = Kernel;
y.Degree = Degree;
y.Gamma = Gamma;
y.Coef0 = Coef0;
y.C = C;
y.Nu = Nu;
y.P = P;
y.CacheSize = CacheSize;
y.Eps = Eps;
y.Shrinking = Shrinking;
y.Probability = Probability;
y.WeightLabels = WeightLabels.Select(a => a).ToArray();
y.Weights = Weights.Select(a => a).ToArray();
return(y);
}
/*public Classification(string training, string test)
{
problem = SVMProblemHelper.Load(training);
testProblem = SVMProblemHelper.Load(test);
}*/
public void ClasificationHand()
{
parameter = new SVMParameter();
parameter.Type = SVMType.C_SVC;
parameter.Kernel = SVMKernelType.RBF;
parameter.C = 1;
parameter.Gamma = 1;
// System.Windows.MessageBox.Show("Algo real");
SVMModel model = SVM.Train(problem, parameter);
//System.Windows.MessageBox.Show("Oscar");
double[] target;
target = new double[testProblem.Length];
int a = 3;
for (int i = 0; i < testProblem.Length; i++)
target[i] = SVM.Predict(model, testProblem.X[i]);
double accuracy = SVMHelper.EvaluateClassificationProblem(testProblem, target);
}
/// <summary>
/// This function constructs and returns an SVM model according to the given training data and parameters.
/// </summary>
/// <param name="problem">Training data.</param>
/// <param name="parameter">Parameter set.</param>
/// <returns>SVM model according to the given training data and parameters.</returns>
public static SVMModel Train(SVMProblem problem, SVMParameter parameter)
{
if (problem == null)
{
throw new ArgumentNullException("problem");
}
if (parameter == null)
{
throw new ArgumentNullException("parameter");
}
IntPtr ptr_problem = SVMProblem.Allocate(problem);
IntPtr ptr_parameter = SVMParameter.Allocate(parameter);
IntPtr ptr_model = libsvm.svm_train(ptr_problem, ptr_parameter);
SVMModel model = SVMModel.Convert(ptr_model);
SVMProblem.Free(ptr_problem);
SVMParameter.Free(ptr_parameter);
libsvm.svm_free_model_content(ptr_model);
return(model);
}
/// <summary>
/// This function constructs and returns an SVM model according to the given training data and parameters.
/// </summary>
/// <param name="problem">Training data.</param>
/// <param name="parameter">Parameter set.</param>
/// <returns>SVM model according to the given training data and parameters.</returns>
public static SVMModel Train(SVMProblem problem, SVMParameter parameter)
{
if (problem == null)
{
throw new ArgumentNullException("problem");
}
if (parameter == null)
{
throw new ArgumentNullException("parameter");
}
IntPtr ptr_problem = SVMProblem.Allocate(problem);
IntPtr ptr_parameter = SVMParameter.Allocate(parameter);
IntPtr ptr_model = libsvm.svm_train(ptr_problem, ptr_parameter);
SVMModel model = SVMModel.Convert(ptr_model);
SVMProblem.Free(ptr_problem);
SVMParameter.Free(ptr_parameter);
libsvm.svm_free_model_content(ptr_model);
return model;
}
/// <summary>
///
/// </summary>
/// <param name="problem"></param>
/// <param name="parameter"></param>
/// <param name="nFolds"></param>
/// <param name="target"></param>
public static void CrossValidation(SVMProblem problem, SVMParameter parameter, int nFolds, out double[] target)
{
IntPtr ptr_target = Marshal.AllocHGlobal(Marshal.SizeOf(typeof(double)) * problem.Length);
IntPtr ptr_problem = SVMProblem.Allocate(problem);
IntPtr ptr_parameter = SVMParameter.Allocate(parameter);
libsvm.svm_cross_validation(ptr_problem, ptr_parameter, nFolds, ptr_target);
target = new double[problem.Length];
Marshal.Copy(ptr_target, target, 0, target.Length);
SVMProblem.Free(ptr_problem);
SVMParameter.Free(ptr_parameter);
Marshal.FreeHGlobal(ptr_target);
ptr_target = IntPtr.Zero;
}
public MetaClassifier(string Name, SVMParameter Parameter, SAMData SamData, List<StdClassifier> Classifiers)
: base(Name, Parameter, SamData)
{
standardClassifiers = Classifiers;
}
public Classifier(string Name, SVMParameter Parameter, SAMData samData)
{
this.Name = Name;
this.Parameters = new List<SVMParameter>() { Parameter };
this.samData = samData;
}
// private int numberOfTasks = 10;
private async Task<NoveltyResult> DoNoveltyDetection(SENSOR sensor, int start, int end)
{
string sensorPath = path + "/" + sensor.ToString();
var data = featureVectors[sensor].Select(x => x.Features).ToList();
ConcurrentStack<SVMParameter> svmParams = new ConcurrentStack<SVMParameter>();
//Debug purpose
/*for (int i = 0; i < 10; i++)
{
SVMParameter s = new SVMParameter();
s.C = 100;
s.Gamma = 0.01;
s.Kernel = SVMKernelType.RBF;
s.Type = SVMType.ONE_CLASS;
s.Nu = 0.01;
svmParams.Push(s);
}*/
svmParams.PushRange(GenerateOneClassSVMParameters().ToArray());
SetProgressMax(svmParams.Count + 1);
NoveltyResult bestCoveredResult = null;
Mutex bestResultMu = new Mutex(false, sensor.ToString());
int count = 1;
List<Task> tasks = new List<Task>();
for (int i = 0; i < threadMAX.Value; i++)
{
Task task = Task.Run(() => PredictionThread(ref count, sensor, start, end, ref svmParams, data, svmParams.Count, ref bestCoveredResult, ref bestResultMu));
tasks.Add(task);
}
await Task.WhenAll(tasks);
List<int> integers = Enumerable.Range(1, 100).ToList();
List <double> nus = integers.Select(x => ((double)x) / 100).ToList();
//Cover
foreach (double d in nus)
{
SVMParameter para = new SVMParameter();
para.Gamma = bestCoveredResult.parameter.Gamma;
para.Nu = d;
para.Kernel = bestCoveredResult.parameter.Kernel;
svmParams.Push(para);
}
count = 1;
SetProgressMax(svmParams.Count + 1);
bestResultMu.Dispose();
bestResultMu = new Mutex(false, sensor.ToString());
List<Task> nuCov = new List<Task>();
ConcurrentBag<string> covNu = new ConcurrentBag<string>();
for (int i = 0; i < threadMAX.Value; i++)
{
Task task = Task.Run(() => PredictionNuThread(ref count, sensor, start, end, ref svmParams, data, svmParams.Count, ref bestCoveredResult, ref bestResultMu, ref covNu));
tasks.Add(task);
}
await Task.WhenAll(tasks);
File.WriteAllLines(path + "/Params.txt", new List<string> { bestCoveredResult.parameter.Gamma.ToString(), bestCoveredResult.parameter.Kernel.ToString() });
File.WriteAllLines(path + $"/{sensor.ToString()}CovNu.txt", covNu.ToList());
bestResultMu.Dispose();
return bestCoveredResult;
}
public SVMConfiguration(string Name, SVMParameter Parameters, List<Feature> Features)
: this(Parameters, Features)
{
this.Name = Name;
}
public static IntPtr Allocate(SVMParameter x)
{
if (x == null)
return IntPtr.Zero;
svm_parameter y = new svm_parameter();
y.svm_type = (int)x.Type;
y.kernel_type = (int)x.Kernel;
y.degree = x.Degree;
y.gamma = x.Gamma;
y.coef0 = x.Coef0;
y.cache_size = x.CacheSize;
y.eps = x.Eps;
y.C = x.C;
y.nu = x.Nu;
y.p = x.P;
y.shrinking = x.Shrinking ? 1 : 0;
y.probability = x.Probability ? 1 : 0;
y.nr_weight = x.WeightLabels.Length;
y.weight_label = IntPtr.Zero;
if (y.nr_weight > 0)
{
y.weight_label = Marshal.AllocHGlobal(Marshal.SizeOf(typeof(int)) * x.WeightLabels.Length);
Marshal.Copy(x.WeightLabels, 0, y.weight_label, x.WeightLabels.Length);
}
y.weight = IntPtr.Zero;
if (y.nr_weight > 0)
{
y.weight = Marshal.AllocHGlobal(Marshal.SizeOf(typeof(double)) * x.Weights.Length);
Marshal.Copy(x.Weights, 0, y.weight, x.Weights.Length);
}
int size = Marshal.SizeOf(y);
IntPtr ptr = Marshal.AllocHGlobal(size);
Marshal.StructureToPtr(y, ptr, true);
return ptr;
}
private void button1_Click(object sender, EventArgs e)
{
noveltyChart.ChartAreas.First().BackColor = Color.DarkGray;
FolderBrowserDialog fbd = new FolderBrowserDialog();
if (fbd.ShowDialog() == DialogResult.
OK)
{
string path = fbd.SelectedPath;
string testSubjectId = path.Split('\\')[path.Split('\\').Length - 2];
fdNovelty.LoadFromFile(new string[] { path + @"\EEG.dat", path + @"\GSR.dat", path + @"\HR.dat", path + @"\KINECT.dat" }, DateTime.Now, false);
events = File.ReadAllLines(path + @"\SecondTest.dat");
string[] tmpSevents = File.ReadAllLines(path + @"\sam.dat");
foreach (string ev in tmpSevents)
{
sEvents.Add(new samEvents(int.Parse(ev.Split(':')[0]), int.Parse(ev.Split(':')[1]), int.Parse(ev.Split(':')[2])));
}
}
if (events.Length == 0)
{
//if events is not assigned with second test data
return;
}
int start = (useRestInTraining.Checked) ? 180000 : 0;
int trainingEnd = int.Parse(events[2].Split('#')[0]);
int windowSize = 5000;
int stepSize = 100;
int delay = 2000;
//Split into training & prediction set
List<List<double>> featureVectors = new List<List<double>>();
List<int> timeStamps = new List<int>();
for (int i = 0; i < fdNovelty.gsrData.Last().timestamp - fdNovelty.gsrData.First().timestamp - windowSize; i += stepSize)
{
List<double> featureVector = new List<double>();
List<double> data = fdNovelty.gsrData.SkipWhile(x => (x.timestamp - fdNovelty.gsrData.First().timestamp) < i).TakeWhile(x => i + windowSize > (x.timestamp - fdNovelty.gsrData.First().timestamp)).Select(x => (double)x.resistance).ToList();
if (data.Count == 0) continue;
featureVector.Add(data.Average());
featureVector.Add(data.Max());
featureVector.Add(data.Min());
double avg = data.Average();
double sd = Math.Sqrt(data.Average(x => Math.Pow(x - avg, 2)));
featureVector.Add(sd);
featureVectors.Add(featureVector);
timeStamps.Add(i);
}
featureVectors = featureVectors.NormalizeFeatureList<double>(Normalize.OneMinusOne).ToList();
var dataSet = featureVectors.Zip(timeStamps, (first, second) => { return Tuple.Create(first, second); });
var trainingSet = dataSet.SkipWhile(x => x.Item2 < start).TakeWhile(x => x.Item2 < trainingEnd);
var predictionSet = dataSet.SkipWhile(x => x.Item2 < trainingEnd);
int count = predictionSet.Count();
int firstPredcition = predictionSet.First().Item2;
OneClassClassifier occ = new OneClassClassifier(trainingSet.Select(x => x.Item1).ToList());
SVMParameter svmP = new SVMParameter();
svmP.Kernel = SVMKernelType.RBF;
svmP.C = 100;
svmP.Gamma = 0.01;
svmP.Nu = 0.01;
svmP.Type = SVMType.ONE_CLASS;
occ.CreateModel(svmP);
/* List<int> indexes = occ.PredictOutliers(predictionSet.Select(x => x.Item1).ToList());
foreach (int index in indexes)
{
timestampsOutliers.Add(predictionSet.ElementAt(index).Item2 - firstPredcition + 180000 + 4000);
}
*/
updateChart();
int k = 0;
}
private List<SVMParameter> GenerateSVMParameters()
{
List<double> cTypes = new List<double>() { };
List<double> gammaTypes = new List<double>() { };
List<SVMKernelType> kernels = new List<SVMKernelType> { SVMKernelType.LINEAR, SVMKernelType.POLY, SVMKernelType.RBF, SVMKernelType.SIGMOID };
for (int t = -5; t <= 15; t++)
{
cTypes.Add(Math.Pow(2, t));
}
for (int t = -15; t <= 3; t++)
{
gammaTypes.Add(Math.Pow(2, t));
}
//Generate SVMParams
List<SVMParameter> svmParams = new List<SVMParameter>();
foreach (SVMKernelType kernel in kernels)
{
foreach (double c in cTypes)
{
for (int i = 0; (kernel != SVMKernelType.LINEAR) ? i < gammaTypes.Count : i < 1; i++)
{
SVMParameter t = new SVMParameter();
t.Kernel = kernel;
t.C = c;
t.Gamma = gammaTypes[i];
svmParams.Add(t);
}
}
}
return svmParams;
}
//学習ファイルの作成
public void TrainingExec(List<FaceFeature.FeatureValue> FeatureList)
{
//特徴量をMatに移し替える 2個で一つ
//2個のfloat * LISTの大きさの配列
double[] feature_array = new double[2 * FeatureList.Count];
//特徴量をSVMで扱えるように配列に置き換える
SetFeatureListToArray(FeatureList,ref feature_array);
CvPoint2D32f[] feature_points = new CvPoint2D32f[feature_array.Length/2];
int id = 0;
for (int i = 0; i < feature_array.Length / 2; i++)
{
feature_points[id].X = (float)feature_array[i * 2];
feature_points[id].Y = (float)feature_array[i * 2 + 1];
id++;
}
CvMat dataMat = new CvMat(feature_points.Length, 2, MatrixType.F32C1, feature_points, true);
//これがラベル番号
int[] id_array = new int[FeatureList.Count];
for(int i = 0; i < id_array.Length;i++)
{
id_array[i] = FeatureList[i].ID;
}
CvMat resMat = new CvMat(id_array.Length, 1, MatrixType.S32C1, id_array, true);
// dataとresponsesの様子を描画
CvPoint2D32f[] points = new CvPoint2D32f[id_array.Length];
int idx = 0;
for (int i = 0; i < id_array.Length; i++)
{
points[idx].X = (float)feature_array[i * 2];
points[idx].Y = (float)feature_array[i * 2 + 1];
idx++;
}
//学習データを図にする
Debug_DrawInputFeature(points, id_array);
//デバッグ用 学習させる特徴量を出力する
OutPut_FeatureAndID(points, id_array);
//LibSVMのテスト
//学習用のデータの読み込み
SVMProblem problem = SVMProblemHelper.Load(@"wine.txt");
SVMProblem testProblem = SVMProblemHelper.Load(@"wine.txt");
SVMParameter parameter = new SVMParameter();
parameter.Type = LibSVMsharp.SVMType.C_SVC;
parameter.Kernel = LibSVMsharp.SVMKernelType.RBF;
parameter.C = 1;
parameter.Gamma = 1;
SVMModel model = SVM.Train(problem, parameter);
double[] target = new double[testProblem.Length];
for (int i = 0; i < testProblem.Length; i++)
{
target[i] = SVM.Predict(model, testProblem.X[i]);
}
double accuracy = SVMHelper.EvaluateClassificationProblem(testProblem, target);
//SVMの用意
CvTermCriteria criteria = new CvTermCriteria(1000, 0.000001);
CvSVMParams param = new CvSVMParams(
OpenCvSharp.CPlusPlus.SVMType.CSvc,
OpenCvSharp.CPlusPlus.SVMKernelType.Rbf,
10.0, // degree
100.0, // gamma 調整
1.0, // coeff0
10.0, // c 調整
0.5, // nu
0.1, // p
null,
criteria);
//学習実行
svm.Train(dataMat, resMat, null, null, param);
Debug_DispPredict();
}
private void btn_CalculateResults_Click(object sender, EventArgs e)
{
FolderBrowserDialog fbd = new FolderBrowserDialog() { Description = "Select folder to load test subjects from" };
if (fbd.ShowDialog() == DialogResult.OK)
{
FolderBrowserDialog sfd = new FolderBrowserDialog() { Description = "Select folder to load excel sheets from" };
if (sfd.ShowDialog() == DialogResult.OK)
{
Excel.Application exc = new Excel.Application() { Visible = false };
List<string> singles = new List<string>()
{
"EEG",
"FACE",
"GSR",
"HR"
};
List<string> fusion = new List<string>()
{
"Stacking",
"Voting"
};
Dictionary<SAMDataPoint.FeelingModel, int> cellOffsets = new Dictionary<SAMDataPoint.FeelingModel, int>()
{
[SAMDataPoint.FeelingModel.Arousal2High] = 11,
[SAMDataPoint.FeelingModel.Arousal2Low] = 24,
[SAMDataPoint.FeelingModel.Arousal3] = 40,
[SAMDataPoint.FeelingModel.Valence2High] = 53,
[SAMDataPoint.FeelingModel.Valence2Low] = 66,
[SAMDataPoint.FeelingModel.Valence3] = 82
};
//test subject[SAMModel[machineType]]
//items: C, Gamma, Kernel
var configs = new Dictionary<string, Dictionary<SAMDataPoint.FeelingModel, Dictionary<string, Tuple<double, double, int>>>>();
foreach (var configBook in singles.Concat(fusion))
{
Log.LogMessage($"Reading {configBook}...");
Excel.Workbook dataBook = exc.Workbooks.Open(sfd.SelectedPath + $"/{configBook}.xlsx", ExcelHandler.missingValue,
false,
ExcelHandler.missingValue,
ExcelHandler.missingValue,
ExcelHandler.missingValue,
true,
ExcelHandler.missingValue,
ExcelHandler.missingValue,
true,
ExcelHandler.missingValue,
ExcelHandler.missingValue,
ExcelHandler.missingValue);
foreach (Excel.Worksheet sheet in dataBook.Sheets)
{
if (sheet.Name == "Overview" || sheet.Name == "First" || sheet.Name == "Last") continue;
if (!configs.ContainsKey(sheet.Name))
{
configs.Add(sheet.Name, new Dictionary<SAMDataPoint.FeelingModel, Dictionary<string, Tuple<double, double, int>>>());
}
if (configs[sheet.Name].Count == 0)
{
foreach (SAMDataPoint.FeelingModel feel in Enum.GetValues(typeof(SAMDataPoint.FeelingModel)))
{
configs[sheet.Name].Add(feel, new Dictionary<string, Tuple<double, double, int>>());
singles.ForEach(x => configs[sheet.Name][feel].Add(x, null));
fusion.ForEach(x => configs[sheet.Name][feel].Add(x, null));
if (feel.ToString().Contains("Valence"))
{
configs[sheet.Name][feel].Remove("GSR");
}
}
}
foreach (SAMDataPoint.FeelingModel feel in Enum.GetValues(typeof(SAMDataPoint.FeelingModel)))
{
if (configBook == "GSR" && feel.ToString().Contains("Valence")) continue;
Excel.Range c = (Excel.Range)sheet.Cells[cellOffsets[feel], 3];
Excel.Range gamma = (Excel.Range)sheet.Cells[cellOffsets[feel] + 1, 3];
Excel.Range kernel = (Excel.Range)sheet.Cells[cellOffsets[feel] + 2, 3];
if (c.Value != null && gamma.Value != null && kernel.Value != null)
{
configs[sheet.Name][feel][configBook] = Tuple.Create(c.Value, gamma.Value, (int)kernel.Value);
}
}
}
dataBook.Close();
}
List<string> skippedList = new List<string>();
foreach (var subject in configs)
{
foreach (var feel in subject.Value)
{
foreach (var machine in feel.Value)
{
if (machine.Value == null)
{
skippedList.Add(subject.Key);
}
}
}
}
foreach (var skipped in skippedList)
{
configs.Remove(skipped);
Log.LogMessage($"Skipped {skipped}");
}
var resultsList = new Dictionary<string, Dictionary<string, List<int>>>();
foreach (string feel in Enum.GetValues(typeof(SAMDataPoint.FeelingModel)))
{
resultsList.Add(feel, new Dictionary<string, List<int>>());
resultsList.Add("SAM", new Dictionary<string, List<int>>());
foreach (var item in singles.Concat(fusion))
{
resultsList[feel].Add(item, new List<int>());
}
}
foreach (var subject in configs)
{
if (LoadData(fbd.SelectedPath + $"/{subject.Key}", _fd))
{
foreach (var feel in subject.Value)
{
resultsList["SAM"][feel.Key.ToString()].AddRange(samData.dataPoints.Select(x => x.ToAVCoordinate(feel.Key)));
foreach (var machine in feel.Value)
{
Log.LogMessage($"Calculating {subject.Key} / {feel} / {machine}");
var param = new SVMParameter()
{
C = machine.Value.Item1,
Gamma = machine.Value.Item2,
Kernel = (SVMKernelType)machine.Value.Item3
};
if (singles.Contains(machine.Key))
{
var classifier = new StdClassifier(new SVMConfiguration(param, FeatureCreator.GetFeatures(machine.Key, feel.Key)), samData);
var results = classifier.OldCrossValidate(feel.Key, 1);
foreach (var res in results[0].guesses)
{
resultsList[feel.Key.ToString()][machine.Key].Add(res);
}
}
else
{
//var classifier = new MetaClassifier(machine.Key, param, samData)
}
}
}
}
else
{
Log.LogMessage($"Skipped {subject.Key} due to bad data");
skippedList.Add(subject.Key);
}
}
Excel.Workbook resultBook = exc.Workbooks.Add(ExcelHandler.missingValue);
Excel.Worksheet metaSheet = resultBook.Sheets.Add();
metaSheet.Name = "Meta";
metaSheet.Cells[1, 1] = "Skipped " + skippedList.Count;
for (int i = 0; i < skippedList.Count; i++)
{
metaSheet.Cells[i + 2, 1] = skippedList[i];
}
foreach (var feel in Enum.GetNames(typeof(SAMDataPoint.FeelingModel)))
{
Excel.Worksheet feelSheet = resultBook.Sheets.Add();
feelSheet.Name = feel;
}
exc.Quit();
}
}
}
private List<SVMParameter> GenerateOneClassSVMParameters()
{
List<double> gammaTypes = new List<double>() { };
List<SVMKernelType> kernels = new List<SVMKernelType> { SVMKernelType.RBF, SVMKernelType.SIGMOID };
for (int t = -14; t <= 2; t += 1)
{
gammaTypes.Add(Math.Pow(2, t));
}
//Generate SVMParams
List<SVMParameter> svmParams = new List<SVMParameter>();
foreach (SVMKernelType kernel in kernels)
{
for (int i = 0; i < gammaTypes.Count; i++)
{
SVMParameter t = new SVMParameter();
t.Kernel = kernel;
t.Nu = 0.05;
t.Gamma = gammaTypes[i];
svmParams.Add(t);
}
}
return svmParams;
}
public static IntPtr Allocate(SVMModel x)
{
if (x == null || x.ClassCount < 1 || x.Labels == null || x.Labels.Length < 1 || x.Parameter == null ||
x.Rho == null || x.Rho.Length < 1 || x.SVCoefs == null || x.SVCoefs.Count < 1 || x.TotalSVCount < 1 ||
x.SVCounts == null || x.SVCounts.Length < 1)
{
return(IntPtr.Zero);
}
svm_model y = new svm_model();
y.nr_class = x.ClassCount;
y.l = x.TotalSVCount;
y.free_sv = (int)x.Creation;
// Allocate model.parameter
IntPtr ptr_param = SVMParameter.Allocate(x.Parameter);
y.param = (svm_parameter)Marshal.PtrToStructure(ptr_param, typeof(svm_parameter));
// Allocate model.rho
y.rho = Marshal.AllocHGlobal(Marshal.SizeOf(typeof(double)) * x.Rho.Length);
Marshal.Copy(x.Rho, 0, y.rho, x.Rho.Length);
// Allocate model.probA
y.probA = IntPtr.Zero;
if (x.ProbabilityA != null)
{
y.probA = Marshal.AllocHGlobal(Marshal.SizeOf(typeof(double)) * x.ProbabilityA.Length);
Marshal.Copy(x.ProbabilityA, 0, y.probA, x.ProbabilityA.Length);
}
// Allocate model.probB
y.probB = IntPtr.Zero;
if (x.ProbabilityB != null)
{
y.probB = Marshal.AllocHGlobal(Marshal.SizeOf(typeof(double)) * x.ProbabilityB.Length);
Marshal.Copy(x.ProbabilityB, 0, y.probB, x.ProbabilityB.Length);
}
// Allocate model.label
y.label = Marshal.AllocHGlobal(Marshal.SizeOf(typeof(int)) * x.Labels.Length);
Marshal.Copy(x.Labels, 0, y.label, x.Labels.Length);
// Allocate model.nSV
y.nSV = Marshal.AllocHGlobal(Marshal.SizeOf(typeof(int)) * x.SVCounts.Length);
Marshal.Copy(x.SVCounts, 0, y.nSV, x.SVCounts.Length);
// Allocate model.sv_coef
y.sv_coef = Marshal.AllocHGlobal(Marshal.SizeOf(typeof(IntPtr)) * x.SVCoefs.Count);
IntPtr i_ptr_svcoef = y.sv_coef;
for (int i = 0; i < x.SVCoefs.Count; i++)
{
IntPtr temp = Marshal.AllocHGlobal(Marshal.SizeOf(typeof(double)) * x.SVCoefs[i].Length);
Marshal.Copy(x.SVCoefs[i], 0, temp, x.SVCoefs[i].Length);
Marshal.StructureToPtr(temp, i_ptr_svcoef, true);
i_ptr_svcoef = IntPtr.Add(i_ptr_svcoef, Marshal.SizeOf(typeof(IntPtr)));
}
// Allocate model.sv_indices
y.sv_indices = IntPtr.Zero;
if (x.SVIndices != null)
{
y.sv_indices = Marshal.AllocHGlobal(Marshal.SizeOf(typeof(int)) * x.SVIndices.Length);
Marshal.Copy(x.SVIndices, 0, y.sv_indices, x.SVIndices.Length);
}
// Allocate model.SV
y.SV = IntPtr.Zero;
if (x.SV != null)
{
y.SV = Marshal.AllocHGlobal(Marshal.SizeOf(typeof(IntPtr)) * x.SV.Count);
IntPtr i_ptr_sv = y.SV;
for (int i = 0; i < x.SV.Count; i++)
{
// Prepare each node array
// 1) All nodes containing zero value is removed
// 2) A node which index is -1 is added to the end
List <SVMNode> temp = x.SV[i].Where(a => a.Value != 0).ToList();
temp.Add(new SVMNode(-1, 0));
SVMNode[] nodes = temp.ToArray();
// Allocate node array
IntPtr ptr_nodes = SVMNode.Allocate(nodes);
Marshal.StructureToPtr(ptr_nodes, i_ptr_sv, true);
i_ptr_sv = IntPtr.Add(i_ptr_sv, Marshal.SizeOf(typeof(IntPtr)));
}
}
// Allocate the model
int size = Marshal.SizeOf(y);
IntPtr ptr = Marshal.AllocHGlobal(size);
Marshal.StructureToPtr(y, ptr, true);
return(ptr);
}
public static void Libsvm(string user)
{
Thread.Sleep(5000);
int Y = 4;
string word = "east";
while (Y > 0)
{
if (Y == 3)
{
word = "Ground";
}
if (Y == 2)
{
word = "sea";
}
if (Y == 1)
{
word = "station";
}
if (File.Exists(@"..\..\..\keyloggerattack\Dataset\SVM_" + user + "_" + word + ".txt"))
{
// Load the datasets: In this example I use the same datasets for training and testing which is not suggested
SVMProblem trainingSet = SVMProblemHelper.Load(@"..\..\..\keyloggerattack\Dataset\SVM_" + user + "_" + word + ".txt");
//SVMProblem testSet = SVMProblemHelper.Load(@"..\..\..\keyloggerattack\Dataset\SVM_" + user + "_east.txt");
//SVMProblem trainingSet = SVMProblemHelper.Load(@"..\..\..\keyloggerattack\Dataset\SVM_temp_east.txt");
SVMProblem testSet = SVMProblemHelper.Load(@"..\..\..\keyloggerattack\Dataset\SVM_" + user + "_" + word + ".txt");
//SVMProblem trainingSet = SVMProblemHelper.Load(@"..\..\..\keyloggerattack\Dataset\1012964-Ground-train.txt");
//SVMProblem testSet = SVMProblemHelper.Load(@"..\..\..\keyloggerattack\Dataset\1012964-Ground-train.txt");
//Normalize the datasets if you want: L2 Norm => x / ||x||
trainingSet = trainingSet.Normalize(SVMNormType.L2);
testSet = testSet.Normalize(SVMNormType.L2);
// Select the parameter set
SVMParameter parameter = new SVMParameter();
parameter.Type = SVMType.ONE_CLASS;
parameter.Kernel = SVMKernelType.RBF;
parameter.C = 8;
parameter.Gamma = 0.0078125;
parameter.Nu = 0.3;
// Do cross validation to check this parameter set is correct for the dataset or not
//double[] crossValidationResults; // output labels
//int nFold = 5;
//trainingSet.CrossValidation(parameter, nFold, out crossValidationResults);
// Evaluate the cross validation result
// If it is not good enough, select the parameter set again
//double crossValidationAccuracy = trainingSet.EvaluateClassificationProblem(crossValidationResults);
// Train the model, If your parameter set gives good result on cross validation
SVMModel model = SVM.Train(trainingSet, parameter);
// Save the model
SVM.SaveModel(model, @"..\..\..\keyloggerattack\Dataset\SVM_" + user + "_" + word + "_model.txt");
//SVM.SaveModel(model, @"..\..\..\keyloggerattack\Dataset\1012964-Ground-train_model.txt");
// Predict the instances in the test set
double[] testResults = testSet.Predict(model);
//for (int i = 0; i < testSet.Length; i++)
// MessageBox.Show("results"+testResults[i]);
double[] target = new double[testSet.Length];
// Evaluate the test results
//int[,] confusionMatrix;
//double testAccuracy = testSet.EvaluateClassificationProblem(testResults, model.Labels, out confusionMatrix);
for (int i = 0; i < testSet.Length; i++)
{
target[i] = SVM.Predict(model, testSet.X[i]);
}
double accuracy = SVMHelper.EvaluateClassificationProblem(testSet, testResults);
// Print the resutls
//Console.WriteLine("\n\nThe accuracy: " + accuracy);
MessageBox.Show("(認知因子)精確率為：" + accuracy);
//MessageBox.Show("test精確率為：" + testAccuracy);
//SVM.Free(trainingSet);
}
if (File.Exists(@"..\..\..\keyloggerattack\Dataset\SVM_Inertia_" + user + "_" + word + ".txt"))
{
// Load the datasets: In this example I use the same datasets for training and testing which is not suggested
SVMProblem trainingSet = SVMProblemHelper.Load(@"..\..\..\keyloggerattack\Dataset\SVM_Inertia_" + user + "_" + word + ".txt");
//SVMProblem testSet = SVMProblemHelper.Load(@"..\..\..\keyloggerattack\Dataset\SVM_" + user + "_east.txt");
//SVMProblem trainingSet = SVMProblemHelper.Load(@"..\..\..\keyloggerattack\Dataset\SVM_temp_east.txt");
SVMProblem testSet = SVMProblemHelper.Load(@"..\..\..\keyloggerattack\Dataset\SVM_Inertia_" + user + "_" + word + ".txt");
//SVMProblem trainingSet = SVMProblemHelper.Load(@"..\..\..\keyloggerattack\Dataset\1012964-Ground-train.txt");
//SVMProblem testSet = SVMProblemHelper.Load(@"..\..\..\keyloggerattack\Dataset\1012964-Ground-train.txt");
//Normalize the datasets if you want: L2 Norm => x / ||x||
trainingSet = trainingSet.Normalize(SVMNormType.L2);
testSet = testSet.Normalize(SVMNormType.L2);
// Select the parameter set
SVMParameter parameter = new SVMParameter();
parameter.Type = SVMType.ONE_CLASS;
parameter.Kernel = SVMKernelType.RBF;
parameter.C = 8;
parameter.Gamma = 0.0078125;
parameter.Nu = 0.3;
// Do cross validation to check this parameter set is correct for the dataset or not
//double[] crossValidationResults; // output labels
//int nFold = 5;
//trainingSet.CrossValidation(parameter, nFold, out crossValidationResults);
// Evaluate the cross validation result
// If it is not good enough, select the parameter set again
//double crossValidationAccuracy = trainingSet.EvaluateClassificationProblem(crossValidationResults);
// Train the model, If your parameter set gives good result on cross validation
SVMModel model = SVM.Train(trainingSet, parameter);
// Save the model
SVM.SaveModel(model, @"..\..\..\keyloggerattack\Dataset\SVM_Inertia_" + user + "_" + word + "_model.txt");
//SVM.SaveModel(model, @"..\..\..\keyloggerattack\Dataset\1012964-Ground-train_model.txt");
// Predict the instances in the test set
double[] testResults = testSet.Predict(model);
//for (int i = 0; i < testSet.Length; i++)
// MessageBox.Show("results"+testResults[i]);
double[] target = new double[testSet.Length];
// Evaluate the test results
//int[,] confusionMatrix;
//double testAccuracy = testSet.EvaluateClassificationProblem(testResults, model.Labels, out confusionMatrix);
for (int i = 0; i < testSet.Length; i++)
{
target[i] = SVM.Predict(model, testSet.X[i]);
}
double accuracy = SVMHelper.EvaluateClassificationProblem(testSet, testResults);
// Print the resutls
//Console.WriteLine("\n\nThe accuracy: " + accuracy);
//MessageBox.Show("(慣性因子)精確率為：" + accuracy);
//MessageBox.Show("test精確率為：" + testAccuracy);
//SVM.Free(trainingSet);
}
Y--;
}
// Console.WriteLine("\n\nCross validation accuracy: " + crossValidationAccuracy);
// Console.WriteLine("\nTest accuracy: " + testAccuracy);
//Console.WriteLine("\nConfusion matrix:\n");
// Print formatted confusion matrix
//Console.Write(String.Format("{0,6}", ""));
/*for (int i = 0; i < model.Labels.Length; i++)
Console.Write(String.Format("{0,5}", "(" + model.Labels[i] + ")"));
Console.WriteLine();
/* for (int i = 0; i < confusionMatrix.GetLength(0); i++)
{
Console.Write(String.Format("{0,5}", "(" + model.Labels[i] + ")"));
for (int j = 0; j < confusionMatrix.GetLength(1); j++)
Console.Write(String.Format("{0,5}", confusionMatrix[i, j]));
Console.WriteLine();
}*/
}
/// <summary>
///
/// </summary>
/// <param name="problem"></param>
/// <param name="parameter"></param>
/// <returns></returns>
public static string CheckParameter(SVMProblem problem, SVMParameter parameter)
{
if (problem == null)
{
throw new ArgumentNullException("problem");
}
if (parameter == null)
{
throw new ArgumentNullException("parameter");
}
IntPtr ptr_problem = SVMProblem.Allocate(problem);
IntPtr ptr_parameter = SVMParameter.Allocate(parameter);
IntPtr ptr_output = libsvm.svm_check_parameter(ptr_problem, ptr_parameter);
SVMProblem.Free(ptr_problem);
SVMParameter.Free(ptr_parameter);
string output = Marshal.PtrToStringAnsi(ptr_output);
Marshal.FreeHGlobal(ptr_output);
ptr_output = IntPtr.Zero;
return output;
}
public static SVMParameter Convert(svm_parameter x)
{
SVMParameter y = new SVMParameter();
y.Type = (SVMType)x.svm_type;
y.Kernel = (SVMKernelType)x.kernel_type;
y.Degree = x.degree;
y.Gamma = x.gamma;
y.Coef0 = x.coef0;
y.CacheSize = x.cache_size;
y.Eps = x.eps;
y.C = x.C;
y.Nu = x.nu;
y.P = x.p;
y.Shrinking = x.shrinking != 0;
y.Probability = x.probability != 0;
int length = x.nr_weight;
y.WeightLabels = new int[length];
if (length > 0)
Marshal.Copy(x.weight_label, y.WeightLabels, 0, length);
y.Weights = new double[length];
if (length > 0)
Marshal.Copy(x.weight, y.Weights, 0, length);
return y;
}
public SVMConfiguration(SVMParameter Parameters, List<Feature> Features)
{
parameters = Parameters;
features = Features;
}
public SVMParameter Clone()
{
SVMParameter y = new SVMParameter();
y.Type = Type;
y.Kernel = Kernel;
y.Degree = Degree;
y.Gamma = Gamma;
y.Coef0 = Coef0;
y.C = C;
y.Nu = Nu;
y.P = P;
y.CacheSize = CacheSize;
y.Eps = Eps;
y.Shrinking = Shrinking;
y.Probability = Probability;
y.WeightLabels = WeightLabels.Select(a => a).ToArray();
y.Weights = Weights.Select(a => a).ToArray();
return y;
}
public static SVMModel Convert(svm_model x)
{
SVMModel y = new SVMModel();
y.Creation = (CreationType)x.free_sv;
y.ClassCount = x.nr_class;
y.TotalSVCount = x.l;
if (y.Creation == CreationType.LOAD_MODEL)
{
y.Parameter = new SVMParameter();
y.Parameter.Type = (SVMType)x.param.svm_type;
y.Parameter.Kernel = (SVMKernelType)x.param.kernel_type;
switch (y.Parameter.Kernel)
{
case SVMKernelType.LINEAR:
break;
case SVMKernelType.POLY:
y.Parameter.Gamma = x.param.gamma;
y.Parameter.Coef0 = x.param.coef0;
y.Parameter.Degree = x.param.degree;
break;
case SVMKernelType.RBF:
y.Parameter.Gamma = x.param.gamma;
break;
case SVMKernelType.SIGMOID:
y.Parameter.Gamma = x.param.gamma;
y.Parameter.Coef0 = x.param.coef0;
break;
}
}
else
{
y.Parameter = SVMParameter.Convert(x.param);
}
int problemCount = (int)(y.ClassCount * (y.ClassCount - 1) * 0.5);
y.Rho = new double[problemCount];
Marshal.Copy(x.rho, y.Rho, 0, y.Rho.Length);
y.ProbabilityA = null;
if (x.probA != IntPtr.Zero)
{
y.ProbabilityA = new double[problemCount];
Marshal.Copy(x.probA, y.ProbabilityA, 0, y.ProbabilityA.Length);
}
y.ProbabilityB = null;
if (x.probB != IntPtr.Zero)
{
y.ProbabilityB = new double[problemCount];
Marshal.Copy(x.probB, y.ProbabilityB, 0, y.ProbabilityB.Length);
}
y.SVCounts = new int[y.ClassCount];
Marshal.Copy(x.nSV, y.SVCounts, 0, y.SVCounts.Length);
y.Labels = new int[y.ClassCount];
Marshal.Copy(x.label, y.Labels, 0, y.Labels.Length);
y.SVCoefs = new List <double[]>(y.ClassCount - 1);
IntPtr i_ptr_svcoef = x.sv_coef;
for (int i = 0; i < y.ClassCount - 1; i++)
{
y.SVCoefs.Add(new double[y.TotalSVCount]);
IntPtr coef_ptr = (IntPtr)Marshal.PtrToStructure(i_ptr_svcoef, typeof(IntPtr));
Marshal.Copy(coef_ptr, y.SVCoefs[i], 0, y.SVCoefs[i].Length);
i_ptr_svcoef = IntPtr.Add(i_ptr_svcoef, Marshal.SizeOf(typeof(IntPtr)));
}
y.SVIndices = null;
if (x.sv_indices != IntPtr.Zero)
{
y.SVIndices = new int[y.TotalSVCount];
Marshal.Copy(x.sv_indices, y.SVIndices, 0, y.SVIndices.Length);
}
y.SV = new List <SVMNode[]>();
IntPtr i_ptr_sv = x.SV;
for (int i = 0; i < x.l; i++)
{
IntPtr ptr_nodes = (IntPtr)Marshal.PtrToStructure(i_ptr_sv, typeof(IntPtr));
SVMNode[] nodes = SVMNode.Convert(ptr_nodes);
y.SV.Add(nodes);
i_ptr_sv = IntPtr.Add(i_ptr_sv, Marshal.SizeOf(typeof(IntPtr)));
}
return(y);
}
public StdClassifier(string Name, SVMParameter Parameter, List<Feature> Features, SAMData samData)
: base(Name, Parameter, samData)
{
features = new List<Feature>(Features);
}
private void CreateSVM(SENSOR machine, List<OneClassFV> trainingSet)
{
var data = trainingSet.Select(x => x.Features).ToList();
OneClassClassifier occ = new OneClassClassifier(data);
SVMParameter svmP = new SVMParameter();
svmP.Kernel = SVMKernelType.RBF;
svmP.C = 100;
svmP.Gamma = 0.01;
svmP.Nu = 0.01;
svmP.Type = SVMType.ONE_CLASS;
occ.CreateModel(svmP);
machines.Add(machine, occ);
}
//--------------------------------------------------------------------------------------
// private
//---------------------------------------------------------------------------------------
/// <summary>
/// 辞書ファイルを作成する
/// </summary>
/// <param name="input_learing_file"></param>
/// <param name="gammma"></param>
/// <param name="cost"></param>
private void Training(string input_learing_file, float gammma, float cost)
{
//LibSVMのテスト
//学習用のデータの読み込み
SVMProblem problem = SVMProblemHelper.Load(input_learing_file);
//SVMパラメータ
SVMParameter parameter = new SVMParameter();
parameter.Type = LibSVMsharp.SVMType.C_SVC;
parameter.Kernel = LibSVMsharp.SVMKernelType.RBF;
parameter.C = cost;
parameter.Gamma = gammma;
libSVM_model = SVM.Train(problem, parameter);
//辞書ファイルを出力(xmlファイル)
string xml_name = @"model_" + input_learing_file;
xml_name = xml_name.Replace(@".csv", @".xml");
SVM.SaveModel(libSVM_model, xml_name);
//判定結果をファイルに出してみる
SVMProblem testProblem = SVMProblemHelper.Load(input_learing_file);
double[] target = new double[testProblem.Length];
string debug_file_str = @"debug_" + input_learing_file;
using (StreamWriter w = new StreamWriter(debug_file_str))
{
for (int i = 0; i < testProblem.Length; i++)
{
target[i] = SVM.Predict(libSVM_model, testProblem.X[i]);
w.Write(target[i] + "\n");
Console.Out.WriteLine(@"{0} : {1}", i, target[i]);
}
}
//正解率を出す。
double accuracy = SVMHelper.EvaluateClassificationProblem(testProblem, target);
}
/// <summary>
///
/// </summary>
/// <param name="problem"></param>
/// <param name="parameter"></param>
/// <param name="nFolds"></param>
/// <param name="target"></param>
public static void CrossValidation(SVMProblem problem, SVMParameter parameter, int nFolds, out double[] target)
{
if (problem == null)
{
throw new ArgumentNullException("problem");
}
if (parameter == null)
{
throw new ArgumentNullException("parameter");
}
if (nFolds < 2)
{
throw new ArgumentOutOfRangeException("nFolds");
}
IntPtr ptr_target = Marshal.AllocHGlobal(Marshal.SizeOf(typeof(double)) * problem.Length);
IntPtr ptr_problem = SVMProblem.Allocate(problem);
IntPtr ptr_parameter = SVMParameter.Allocate(parameter);
libsvm.svm_cross_validation(ptr_problem, ptr_parameter, nFolds, ptr_target);
target = new double[problem.Length];
Marshal.Copy(ptr_target, target, 0, target.Length);
SVMProblem.Free(ptr_problem);
SVMParameter.Free(ptr_parameter);
Marshal.FreeHGlobal(ptr_target);
ptr_target = IntPtr.Zero;
}
