public void Feature_and_LabelIndex_Mapping_Test2()
{
var desc = loadMetaData_with_CategoricFeature();
// Creates learning api object
LearningApi api = new LearningApi(desc);
//Real dataset must be defined as object type, because data can be numeric, binary and classification
api.UseActionModule <object[][], object[][]>((input, ctx) =>
{
return(loadRealDataSample());
});
//this call must be first in the pipeline
api.UseDefaultDataMapper();
//
var result = api.Run() as double[][];
var featureNumber = result[0].Count() - 1;//- labelcolumn
//string column is not counted because it is ignored column
//there is one category column in features so the number of features is increased with (calssCount-1)
// featureCount = columnCount- strinCoulumnCount-LabelColumn + calssCount-1
var featureCount = 4 - 1 - 1 + (3 - 1);
Assert.Equal(featureNumber, featureCount);
}
public void SobelFilterTestForPngImageFileType()
{
LearningApi api = new LearningApi();
api.UseActionModule <double[, , ], double[, , ]>((input, ctx) =>
{
string baseDirectory = AppDomain.CurrentDomain.BaseDirectory;
string path = Path.Combine(baseDirectory, "TestInputimages\\TIM4.png"); //path to bin directory of project
Bitmap bitmap = new Bitmap(path);
double[,,] data = helper.ConvertBitmapToDouble(bitmap); // convert bitmap to double
return(data);
});
api.AddModule(new SobelConvolutionFilter());
double[,,] result = api.Run() as double[, , ];
Bitmap bitresult = helper.ConvertDoubleToBitmap(result);// convert double to bitmap
string baseDirectory2 = AppDomain.CurrentDomain.BaseDirectory;
string outpath = baseDirectory2 + "\\TestOutputImages\\";
if (!Directory.Exists(outpath))
{
Directory.CreateDirectory(outpath);
}
bitresult.Save(outpath + "Output4.png");
}
public void NumericTransformation_test2()
{
// Creates learning api object
LearningApi api = new LearningApi(loadMetaData_with_CategoricFeature());
//Real dataset must be defined as object type, because data can be numeric, binary and classification
api.UseActionModule <object[][], object[][]>((input, ctx) =>
{
return(loadRealDataSample());
});
//this call must be first in the pipeline
api.UseDefaultDataMapper();
//
var result = api.Run() as double[][];
//Test result for normalization
var expected = transformedData_with_CategoricFeature();
for (int i = 0; i < expected.Length; i++)
{
for (int j = 0; j < expected[0].Length; j++)
{
Assert.Equal(result[i][j], expected[i][j]);
}
}
//
return;
}
public void Test_LoadSave()
{
string moduleName = "test-action";
double[][] clusterCentars = new double[3][];
clusterCentars[0] = new double[] { 5.0, 5.0 };
clusterCentars[1] = new double[] { 15.0, 15.0 };
clusterCentars[2] = new double[] { 30.0, 30.0 };
string[] attributes = new string[] { "Height", "Weight" };
int numAttributes = attributes.Length; // 2 in this demo (height,weight)
int numClusters = 3; // vary this to experiment (must be between 2 and number data tuples)
int maxCount = 300; // trial and error
ClusteringSettings settings = new ClusteringSettings(maxCount, numClusters, numAttributes, KmeansAlgorithm: 1);
// Creates learning api object
LearningApi api = new LearningApi(loadDescriptor());
//
// Defines action method, which will generate training data.
api.UseActionModule <object, double[][]>((data, ctx) =>
{
var rawData = Helpers.CreateSampleData(clusterCentars, 2, 10000, 0.5);
return(rawData);
}, moduleName);
api.UseKMeans(settings);
var resp = api.Run() as KMeansScore;
Assert.True(resp.Model.Clusters != null);
Assert.True(resp.Model.Clusters.Length == clusterCentars.Length);
var result = api.Algorithm.Predict(clusterCentars, api.Context) as KMeansResult;
Assert.True(result.PredictedClusters[0] == 0);
Assert.True(result.PredictedClusters[1] == 1);
Assert.True(result.PredictedClusters[2] == 2);
// This is where trained model is saved.
api.Save(nameof(TestLoadSave));
// Loads the saved model.
var loadedApi = LearningApi.Load(nameof(TestLoadSave));
//
// Because we have used action method in the LearningApi, we will have to setup it again.
// This is not required because API design limitation. It is restriction of .NET framework. It cannot persist code.
loadedApi.ReplaceActionModule <object, double[][]>(moduleName, (data, ctx) =>
{
var rawData = Helpers.CreateSampleData(clusterCentars, 2, 10000, 0.5);
return(rawData);
});
loadedApi.Run();
}
public void calculateCorrelation_test1()
{
//
LearningApi api = new LearningApi(null);
// Initialize data provider
api.UseCsvDataProvider(@"CorrelationMatrix/corellation_data.csv", ',', true, 0);
//Custom action of dataset
api.UseActionModule <object[][], double[][]>((input, ctx) =>
{
return(toColumnVector(input));
});
// api.UseMinMaxNormalizer();
var data = api.Run() as double[][];
var prov = api.GetModule <CsvDataProvider>("CsvDataProvider");
var strData = new List <string>();
var hed = prov.Header.ToList();
hed.Insert(0, "");
strData.Add(string.Join(",", hed.ToArray()));
for (int i = 0; i < data.Length; i++)
{
var lst = new List <string>();
lst.Add(prov.Header[i]);
for (int k = 0; k < i; k++)
{
lst.Add(" ");
}
for (int j = i; j < data.Length; j++)
{
var corValue = data[i].CorrCoeffOf(data[j]);
if (double.IsNaN(corValue))
{
continue;
}
lst.Add(corValue.ToString("n5", CultureInfo.InvariantCulture));
}
strData.Add(string.Join(",", lst));
}
Assert.True("Col1,1.00000,0.16892,0.99111,0.75077,-0.82354,-0.85164" == strData[1]);
System.IO.File.WriteAllLines(@"CorrelationMatrix/strCorrlation.txt", strData);
//
return;
}
public void SimpleSequenceWithGapsTest()
{
LearningApi api = new LearningApi();
api.UseActionModule <object, double[][]>((notUsed, ctx) =>
{
const int maxSamples = 1000000;
ctx.DataDescriptor = getDescriptor();
double[][] data = new double[maxSamples / 3][];
//
// We generate following input vectors:
// IN Val - Expected OUT Val
// Every 3th number is given
for (int i = 0; i < maxSamples / 3; i++)
{
data[i] = new double[2];
data[i][0] = i * 3;
if ((i * 3) > (maxSamples / 2))
{
data[i][1] = 1;
}
else
{
data[i][1] = 0;
}
}
return(data);
});
api.UsePerceptron(0.05, 1000);
IScore score = api.Run() as IScore;
double[][] testData = new double[6][];
testData[0] = new double[] { 2.0, 0.0 };
testData[1] = new double[] { 1, 0.0 };
testData[2] = new double[] { 3, 0.0 };
testData[3] = new double[] { 3002, 0.0 };
testData[4] = new double[] { 6002.0, 0.0 };
testData[5] = new double[] { 9005, 0.0 };
var result = api.Algorithm.Predict(testData, api.Context) as PerceptronResult;
// TODO... THUS TEST iS NOT COMPLETED
//Assert.True(result.PredictedValues[0] == 0);
//Assert.True(result.PredictedValues[1] == 0);
//Assert.True(result.PredictedValues[2] == 0);
//Assert.True(result.PredictedValues[3] == 1);
//Assert.True(result.PredictedValues[4] == 1);
//Assert.True(result.PredictedValues[5] == 1);
}
public void LogisticRegression_Test_iterations_20_learningrate_015()
{
var desc = loadMetaData();
LearningApi api = new LearningApi(desc);
//Real dataset must be defined as object type, because data can be numeric, binary and classification
api.UseActionModule <object[][], object[][]>((input, ctx) =>
{
return(loadRealDataSample());
});
// Use mapper for data, which will extract (map) required columns
api.UseDefaultDataMapper();
api.UseMinMaxNormalizer();
//run logistic regression for 10 iteration with learningRate=0.15
api.UseLogisticRegression(0.15, 20);
api.Run();
IScore score = api.GetScore();
//Errors during each iteration
Assert.Equal(Math.Round(score.Errors[0], 5), 0.24236);
Assert.Equal(Math.Round(score.Errors[1], 5), 0.23707);
Assert.Equal(Math.Round(score.Errors[2], 5), 0.23358);
Assert.Equal(Math.Round(score.Errors[3], 5), 0.23001);
Assert.Equal(Math.Round(score.Errors[4], 5), 0.22806);
Assert.Equal(Math.Round(score.Errors[5], 5), 0.22506);
Assert.Equal(Math.Round(score.Errors[6], 5), 0.22409);
Assert.Equal(Math.Round(score.Errors[7], 5), 0.22134);
Assert.Equal(Math.Round(score.Errors[8], 5), 0.22105);
Assert.Equal(Math.Round(score.Errors[9], 5), 0.21840);
Assert.Equal(Math.Round(score.Errors[10], 5), 0.21857);
Assert.Equal(Math.Round(score.Errors[11], 5), 0.21595);
Assert.Equal(Math.Round(score.Errors[12], 5), 0.21640);
Assert.Equal(Math.Round(score.Errors[13], 5), 0.21381);
Assert.Equal(Math.Round(score.Errors[14], 5), 0.21439);
Assert.Equal(Math.Round(score.Errors[15], 5), 0.21189);
Assert.Equal(Math.Round(score.Errors[16], 5), 0.21251);
Assert.Equal(Math.Round(score.Errors[17], 5), 0.21015);
Assert.Equal(Math.Round(score.Errors[18], 5), 0.21076);
Assert.Equal(Math.Round(score.Errors[19], 5), 0.20860);
//LG Model Best Found model in 20 iteration
Assert.Equal(Math.Round(score.Weights[0], 5), 0.28363);
Assert.Equal(Math.Round(score.Weights[1], 5), 0.37424);
Assert.Equal(Math.Round(score.Weights[2], 5), 1.41890);
Assert.Equal(Math.Round(score.Weights[3], 5), 1.01207);
Assert.Equal(Math.Round(score.Weights[4], 5), -0.33841);
Assert.Equal(Math.Round(score.Weights[5], 5), -0.33841);
Assert.Equal(Math.Round(score.Weights[6], 5), -1.62489);
}
public void LogisticRegression_Test_Real_Example()
{
string m_binary_data_path = @"SampleData\binary\admit_binary.csv";
var binary_path = System.IO.Path.Combine(Directory.GetCurrentDirectory(), m_binary_data_path);
LearningApi api = new LearningApi(loadMetaData1());
api.UseCsvDataProvider(binary_path, ',', false, 1);
// Use mapper for data, which will extract (map) required columns
api.UseDefaultDataMapper();
api.UseMinMaxNormalizer();
//run logistic regression for 10 iteration with learningRate=0.15
api.UseLogisticRegression(0.00012, 200);
var score = api.Run();
///**************PREDICTION AFTER MODEL IS CREATED*********////
/////define data for testing (prediction)
LearningApi apiPrediction = new LearningApi(loadMetaData1());
//Real dataset must be defined as object type, because data can be numeric, binary and classification
apiPrediction.UseActionModule <object[][], object[][]>((input, ctx) =>
{
var data = new object[5][]
{
new object[] { 660, 3.88, 2, 1 },
new object[] { 580, 3.36, 2, 0 },
new object[] { 640, 3.17, 2, 0 },
new object[] { 640, 3.51, 2, 0 },
new object[] { 800, 3.05, 2, 1 },
};
return(data);
});
// Use mapper for data, which will extract (map) required columns
apiPrediction.UseDefaultDataMapper();
apiPrediction.UseMinMaxNormalizer();
var testData = apiPrediction.Run();
//use previous trained model
var result = api.Algorithm.Predict(testData as double[][], api.Context) as LogisticRegressionResult;
//
Assert.Equal(Math.Round(result.PredictedValues[0], 0), 0);
Assert.Equal(Math.Round(result.PredictedValues[1], 0), 0);
Assert.Equal(Math.Round(result.PredictedValues[2], 0), 0);
Assert.Equal(Math.Round(result.PredictedValues[3], 0), 0);
Assert.Equal(Math.Round(result.PredictedValues[3], 0), 0);
}
public void ContinuousTrainData()
{
double[][] initialCentroids = new double[4][];
initialCentroids[0] = new double[] { 0.2, -4.0 };
initialCentroids[1] = new double[] { 0.2, -6.0 };
initialCentroids[2] = new double[] { 0.4, -4.0 };
initialCentroids[3] = new double[] { 0.4, -6.0 };
string[] attributes = new string[] { "x", "y" };
int numAttributes = attributes.Length; // 2 in this demo (x,y)
int numClusters = 4;
int maxCount = 300;
SaveLoadSettings sett;
var resp = SaveLoadSettings.JSON_Settings(@"C:\Data\Function1.json", out sett, true);
AnomalyDetectionAPI kmeanApi = new AnomalyDetectionAPI(null, numClusters);
LearningApi api = new LearningApi(loadMetaData1());
api.UseActionModule <object[][], object[][]>((input, ctx) =>
{
var rawDatalist = getRealDataSample(@"C:\Data\Function1.csv").ToList();
double[][] oldSamples;
var nn = kmeanApi.GetPreviousSamples(sett, out oldSamples);
if (oldSamples != null)
{
foreach (var old in oldSamples)
{
var row = old.Cast <object>().ToArray();
rawDatalist.Add(row);
}
}
return(rawDatalist.ToArray());
});
//this call must be first in the pipeline
api.UseDefaultDataMapper();
api.UseGaussNormalizer();
var rawData = api.Run() as double[][];
Helpers.WriteToCSVFile(rawData);
ClusteringSettings Settings = new ClusteringSettings(rawData, maxCount, numClusters, numAttributes, sett, KmeansAlgorithm: 1, InitialGuess: true, Replace: true);
AnomalyDetectionResponse response = kmeanApi.ImportNewDataForClustering(Settings);
}
private object[][] getData(int cnt)
{
string filePath = $"{Directory.GetCurrentDirectory()}\\DataSet\\Book2.csv";
var isris_path = System.IO.Path.Combine(Directory.GetCurrentDirectory(), filePath);
LearningApi api = new LearningApi(loadMetaData1());
api.UseCsvDataProvider(isris_path, ',', 0);
return(api.Run() as object[][]);
}
public void movieRecommendationTestCRbm(int iterations, double learningRate, int visNodes, int hidNodes)
{
Debug.WriteLine($"{iterations}-{visNodes}-{hidNodes}");
LearningApi api = new LearningApi(getDescriptorForRbm(3898));
// Initialize data provider
api.UseCsvDataProvider(Path.Combine(Directory.GetCurrentDirectory(), @"RestrictedBolzmannMachine2\Data\movieDatasetTrain.csv"), ',', false, 0);
api.UseDefaultDataMapper();
double[] featureVector = new double[] { 0, 0.05, 0.1, 0.15, 0.2, 0.25, 0.3, 0.35, 0.4, 0.45, 0.5, 0.55, 0.6, 0.65, 0.7, 0.75, 0.8, 0.85 };
api.UseCRbm(featureVector, learningRate, iterations, visNodes, hidNodes);
Stopwatch watch = new Stopwatch();
watch.Start();
RbmScore score = api.Run() as RbmScore;
watch.Stop();
var hiddenNodes = score.HiddenValues;
var hiddenWeight = score.HiddenBisases;
double[] learnedFeatures = new double[hidNodes];
double[] hiddenWeights = new double[hidNodes];
for (int i = 0; i < hidNodes; i++)
{
learnedFeatures[i] = hiddenNodes[i];
hiddenWeights[i] = hiddenWeight[i];
}
StreamWriter tw = new StreamWriter($"PredictedDigit_I{iterations}_V{visNodes}_H{hidNodes}_learnedbias.txt");
foreach (var item in score.HiddenBisases)
{
tw.WriteLine(item);
}
tw.Close();
var testData = ReadData(Path.Combine(Directory.GetCurrentDirectory(), @"RestrictedBolzmannMachine2\Data\movieDatasetTest.csv"));
var result = api.Algorithm.Predict(testData, api.Context);
var predictedData = ((RbmResult)result).VisibleNodesPredictions;
var predictedHiddenNodes = ((RbmResult)result).HiddenNodesPredictions;
var acc = testData.GetHammingDistance(predictedData);
WriteDeepResult(iterations, new int[] { visNodes, hidNodes }, learningRate, acc, watch.ElapsedMilliseconds * 1000, predictedHiddenNodes);
WriteOutputMatrix(iterations, new int[] { visNodes, hidNodes }, learningRate, predictedData, testData);
}
private object[][] getRealDataSample(string filePath)
{
//
//iris data file
var isris_path = System.IO.Path.Combine(Directory.GetCurrentDirectory(), filePath);
LearningApi api = new LearningApi(loadMetaData1());
api.UseCsvDataProvider(isris_path, ',', 0);
return(api.Run() as object[][]);
}
public void DeltaRuleLearning_Test_iterations_1000_learningrate_02()
{
loadRealDataSample(); // System coefficients initialization.
var desc = loadMetaData(); // Description of the system.
LearningApi api = new LearningApi(desc);
//Real dataset must be defined as object type, because data can be numeric, binary and classification
api.UseActionModule <double[, ], double[, ]>((input, ctx) =>
{
return(loadRealDataSample()); // return actual System coefficients data
});
// run input = UseActionModule output
//run Delta Rule for 1000 iterations with learningRate=0.2
api.UseDeltaRuleLearning(0.2, 1000);
var result = api.Run() as double[];
Debug.WriteLine("************ Output Predictions***********");
for (int i = 0; i < result.Length; i++)
{
if (result[i] != 0)
{
Debug.WriteLine(result[i]);
}
}
using (var fs = File.OpenRead(@"H_Test.csv"))
using (var reader = new StreamReader(fs))
{
while (!reader.EndOfStream)
{
var line = reader.ReadLine();
var values = line.Split(',');
ti = new double[values.Length];
to = new double[values.Length];
for (int i = 0; i < values.Length; i++)
{
var val = values[i].Split(' ');
double.TryParse(val[0], out x);
double.TryParse(val[1], out y);
ti[i] = x;
to[i] = y;
}
}
}
for (int i = 0; i < to.Length; i++)
{
//Testing of Test data with Predicted System model
Assert.Equal(Math.Round(result[i], 4), Math.Round(to[i], 4));
}
}
public void SimpleRBMDeepTest()
{
LearningApi api = new LearningApi();
api.UseActionModule <object, double[][]>((notUsed, ctx) =>
{
const int maxSamples = 12;
ctx.DataDescriptor = getDescriptorForRbm_sample1();
double[][] data = new double[maxSamples][];
data[0] = new double[] { 1, 1, 0, 0, 0, 0 }; // A
data[1] = new double[] { 0, 0, 1, 1, 0, 0 }; // B
data[2] = new double[] { 0, 0, 0, 0, 1, 1 }; // C
data[3] = new double[] { 1, 1, 0, 0, 0, 1 }; // noisy A
data[4] = new double[] { 0, 0, 1, 1, 0, 0 }; // BRt
data[5] = new double[] { 0, 0, 0, 0, 1, 1 }; // C
data[6] = new double[] { 1, 0, 0, 0, 0, 0 }; // weak A
data[7] = new double[] { 0, 0, 1, 0, 0, 0 }; // weak B
data[8] = new double[] { 0, 0, 0, 0, 1, 0 }; // weak C
data[9] = new double[] { 1, 1, 0, 1, 0, 0 }; // noisy A
data[10] = new double[] { 1, 0, 1, 1, 0, 0 }; // noisy B
data[11] = new double[] { 0, 0, 1, 0, 1, 1 }; // noisy C
return(data);
});
api.UseRbm(0.01, 1000, 6, 4);
RbmScore score = api.Run() as RbmScore;
double[][] testData = new double[4][];
Assert.True(score.Loss < 1.0);
testData[0] = new double[] { 1, 1, 0, 0, 0, 0 };
testData[1] = new double[] { 0, 0, 0, 0, 1, 1 };
testData[2] = new double[] { 0, 1, 0, 0, 0, 0 };
testData[3] = new double[] { 0, 0, 0, 0, 1, 0 };
var result = api.Algorithm.Predict(testData, api.Context);
// NOT FINISHED.
//Assert.True(result[0] == 1);
//Assert.True(result[1] == 0);
//Assert.True(result[2] == 0);
//Assert.True(result[3] == 0);
//Assert.True(result[4] == 1);
//Assert.True(result[5] == 0);
}
public void FullDataSetRBMTest()
{
const int bits = 10;
LearningApi api = new LearningApi();
api.UseActionModule <object, double[][]>((notUsed, ctx) =>
{
var maxSamples = (int)Math.Pow(2, bits);
double[][] data = new double[maxSamples][];
for (int i = 0; i < maxSamples; i++)
{
data[i] = new double[bits];
var val = 1;
for (int j = 0; j < bits; j++)
{
if ((val & i) >= 1)
{
data[i][j] = 1;
}
val = val << 1;
}
}
ctx.DataDescriptor = getDescriptorForRbm_sample1();
return(data);
});
api.UseRbm(0.01, 1000, bits, 7);
RbmScore score = api.Run() as RbmScore;
double[][] testData = new double[4][];
testData[0] = new double[] { 1, 1, 0, 0, 0, 0, 0, 0, 0, 0 };
testData[1] = new double[] { 0, 0, 0, 0, 1, 1, 0, 0, 0, 0 };
testData[2] = new double[] { 0, 1, 0, 0, 0, 0, 0, 0, 0, 0 };
testData[3] = new double[] { 0, 0, 0, 0, 1, 0, 0, 0, 0, 1 };
var result = api.Algorithm.Predict(testData, api.Context);
// NOT FINISHED.
//Assert.True(result[0] == 1);
//Assert.True(result[1] == 0);
//Assert.True(result[2] == 0);
//Assert.True(result[3] == 0);
//Assert.True(result[4] == 1);
//Assert.True(result[5] == 0);
}
public void Test_LoadSave()
{
string moduleName = "test-action";
double[][] clusterCentars = new double[3][];
clusterCentars[0] = new double[] { 5.0, 5.0 };
clusterCentars[1] = new double[] { 15.0, 15.0 };
clusterCentars[2] = new double[] { 30.0, 30.0 };
string[] attributes = new string[] { "Height", "Weight" };
int numAttributes = attributes.Length; // 2 in this demo (height,weight)
int numClusters = 3; // vary this to experiment (must be between 2 and number data tuples)
int maxCount = 300; // trial and error
ClusteringSettings settings = new ClusteringSettings(maxCount, numClusters, numAttributes, KmeansAlgorithm: 1);
// Creates learning api object
LearningApi api = new LearningApi(loadDescriptor());
api.UseActionModule <object, double[][]>((data, ctx) =>
{
var rawData = Helpers.CreateSampleData(clusterCentars, 2, 10000, 0.5);
return(rawData);
}, moduleName);
api.UseKMeans(settings);
var resp = api.Run() as KMeansScore;
Assert.True(resp.Model.Clusters != null);
Assert.True(resp.Model.Clusters.Length == clusterCentars.Length);
var result = api.Algorithm.Predict(clusterCentars, api.Context) as KMeansResult;
Assert.True(result.PredictedClusters[0] == 0);
Assert.True(result.PredictedClusters[1] == 1);
Assert.True(result.PredictedClusters[2] == 2);
api.Save(nameof(LoadSaveTests));
var loadedApi = LearningApi.Load(nameof(LoadSaveTests));
loadedApi.ReplaceActionModule <object, double[][]>(moduleName, (data, ctx) =>
{
var rawData = Helpers.CreateSampleData(clusterCentars, 2, 10000, 0.5);
return(rawData);
});
loadedApi.Run();
}
public void ActionModuleTest()
{
LearningApi api = new LearningApi(null);
api.UseActionModule <double[], double[]>((input, ctx) =>
{
return(new double[] { 1.1, 2.2, 3.3, 4.4 });
});
var result = api.Run();
Assert.Equal(1.1, ((double[])result)[0]);
Assert.Equal(4.4, ((double[])result)[3]);
}
public void SimpleSequenceTest()
{
LearningApi api = new LearningApi();
api.UseActionModule <object, double[][]>((notUsed, ctx) =>
{
const int maxSamples = 10;
ctx.DataDescriptor = getDescriptor();
double[][] data = new double[maxSamples][];
//
// We generate following input vectors:
// IN Val - Expected OUT Val
// 1 - 0
// 2 - 0,
// ...
// maxSamples / 2 - 1,
// maxSamples / 2 + 1 - 1,
// maxSamples / 2 + 2 - 1,
for (int i = 0; i < maxSamples; i++)
{
data[i] = new double[2];
data[i][0] = i;
data[i][1] = (i > (maxSamples / 2)) ? 1 : 0;
}
return(data);
});
api.UseDeltaLearning(0.2, 1000);
IScore score = api.Run() as IScore;
double[][] testData = new double[4][];
testData[0] = new double[] { 2.0, 0.0 };
testData[1] = new double[] { 4.0, 0.0 };
testData[2] = new double[] { 6.0, 0.0 };
testData[3] = new double[] { 8.0, 0.0 };
var result = api.Algorithm.Predict(testData, api.Context) as DeltaLearningResult;
Assert.True(result.PredictedResults[0] == 0);
Assert.True(result.PredictedResults[1] == 0);
Assert.True(result.PredictedResults[2] == 1);
Assert.True(result.PredictedResults[3] == 1);
}
public void Training()
{
int cnt = 0;
double[][] initialCentroids = new double[4][];
initialCentroids[0] = new double[] { 0.4, 25.0 };
initialCentroids[1] = new double[] { 0.4, 15.0 };
initialCentroids[2] = new double[] { 0.6, 15.0 };
initialCentroids[3] = new double[] { 0.6, 25.0 };
string[] attributes = new string[] { "x", "y" };
int numAttributes = attributes.Length;
int numClusters = 4;
int maxCount = 300;
ClusteringSettings clusterSettings = new ClusteringSettings(maxCount, numClusters, numAttributes, KmeansAlgorithm: 1, Replace: true);
AnomalyDetectionAPI kmeanApi = new AnomalyDetectionAPI(clusterSettings); //AnomalyDetectionAPI(clusterSettings), Constractor should not be null when run Training()
AnomalyDetectionResponse response;
// Creates learning api object
LearningApi api = new LearningApi(loadMetaData1());
api.UseActionModule <object[][], object[][]>((input, ctx) =>
{
var rawDataArray = getData(cnt);
return(rawDataArray);
});
api.UseDefaultDataMapper();
api.UseGaussNormalizer();
//
for (int i = 0; i < 15; i++)
{
cnt = i;
var rawData = api.Run() as double[][];
response = kmeanApi.Training(rawData, initialCentroids);
Helpers.WriteToCSVFile(kmeanApi.GetCentroid(), $"Data\\Centroid{i}.csv");
//response = kmeanApi.Save($"Function{i}.json");
}
}
public void smileyTestCRbm(int iterations, double learningRate, int visNodes, int hidNodes)
{
LearningApi api = new LearningApi(getDescriptorForRbm(1600));
// Initialize data provider
api.UseCsvDataProvider(Path.Combine(Directory.GetCurrentDirectory(), @"RestrictedBolzmannMachine2\Data\Smiley.csv"), ',', false, 0);
api.UseDefaultDataMapper();
double[] featureVector = new double[] { 0.1, 0.2 };
api.UseCRbm(featureVector, learningRate, iterations, visNodes, hidNodes);
Stopwatch watch = new Stopwatch();
watch.Start();
RbmScore score = api.Run() as RbmScore;
watch.Stop();
var hiddenNodes = score.HiddenValues;
var hiddenWeight = score.HiddenBisases;
double[] learnedFeatures = new double[hidNodes];
double[] hiddenWeights = new double[hidNodes];
for (int i = 0; i < hidNodes; i++)
{
learnedFeatures[i] = hiddenNodes[i];
hiddenWeights[i] = hiddenWeight[i];
}
var testData = ReadData(Path.Combine(Directory.GetCurrentDirectory(), @"RestrictedBolzmannMachine2\Data\SmileyTest.csv"));
var result = api.Algorithm.Predict(testData, api.Context);
var predictedData = ((RbmResult)result).VisibleNodesPredictions;
var predictedHiddenNodes = ((RbmResult)result).HiddenNodesPredictions;
var acc = testData.GetHammingDistance(predictedData);
var ValTest = calcDelta(predictedData, testData);
var lossTest = ValTest / (visNodes);
Debug.WriteLine($"lossTest: {lossTest}");
WriteDeepResult(iterations, new int[] { visNodes, hidNodes }, acc, watch.ElapsedMilliseconds * 1000, predictedHiddenNodes);
WriteOutputMatrix(iterations, new int[] { visNodes, hidNodes }, predictedData, testData);
}
public void ContinuousTrainData2()
{
int cnt = 0;
double[][] initialCentroids = new double[4][];
initialCentroids[0] = new double[] { 40.0, 10.0 };
initialCentroids[1] = new double[] { 20.0, 10.0 };
initialCentroids[2] = new double[] { 40.0, 20.0 };
initialCentroids[3] = new double[] { 20.0, 20.0 };
string[] attributes = new string[] { "x", "y" };
int numAttributes = attributes.Length; // 2 in this demo (x,y)
int numClusters = 4;
int maxCount = 300;
SaveLoadSettings sett;
var resp = SaveLoadSettings.JSON_Settings(@"C:\Data\Function1.json", out sett, true);
AnomalyDetectionAPI kmeanApi = new AnomalyDetectionAPI(null, numClusters, initialCentroids);
LearningApi api = new LearningApi(loadMetaData1());
api.UseActionModule <object[][], object[][]>((input, ctx) =>
{
var rawDatalist = getData(cnt);
return(rawDatalist);
});
//this call must be first in the pipeline
api.UseDefaultDataMapper();
api.UseGaussNormalizer();
for (int i = 0; i < 15; i++)
{
cnt = i;
var rawData = api.Run() as double[][];
ClusteringSettings Settings = new ClusteringSettings(rawData, maxCount, numClusters, numAttributes, sett, KmeansAlgorithm: 1, InitialGuess: true, Replace: true);
AnomalyDetectionResponse response = kmeanApi.ImportNewDataForClustering(Settings);
}
}
public void SimpleSequence2DTest()
{
LearningApi api = new LearningApi();
api.UseActionModule <object, double[][]>((notUsed, ctx) =>
{
const int maxSamples = 10000;
ctx.DataDescriptor = get2DDescriptor();
double[][] data = new double[maxSamples][];
for (int i = 0; i < maxSamples / 2; i++)
{
data[2 * i] = new double[3];
data[2 * i][0] = i;
data[2 * i][1] = 5.0;
data[2 * i][2] = 1.0;
data[2 * i + 1] = new double[3];
data[2 * i + 1][0] = i;
data[2 * i + 1][1] = -5.0;
data[2 * i + 1][2] = 0.0;
}
return(data);
});
api.UseDeltaLearning(0.2, 1000);
IScore score = api.Run() as IScore;
double[][] testData = new double[6][];
testData[0] = new double[] { 2.0, 5.0, 0.0 };
testData[1] = new double[] { 2, -5.0, 0.0 };
testData[2] = new double[] { 100, -5.0, 0.0 };
testData[3] = new double[] { 100, -5.0, 0.0 };
testData[4] = new double[] { 490, 5.0, 0.0 };
testData[5] = new double[] { 490, -5.0, 0.0 };
var result = api.Algorithm.Predict(testData, api.Context) as DeltaLearningResult;
Assert.True(result.PredictedResults[0] == 1);
Assert.True(result.PredictedResults[1] == 0);
Assert.True(result.PredictedResults[2] == 0);
Assert.True(result.PredictedResults[3] == 0);
Assert.True(result.PredictedResults[4] == 1);
Assert.True(result.PredictedResults[5] == 0);
}
public void GaussianAndMean()
{
LearningApi lApi = new LearningApi();
lApi.UseActionModule <double[, , ], double[, , ]>((input, ctx) =>
{
Bitmap myBitmap = new Bitmap($"{appPath}/TestPicture/test.gif");
double[,,] data = new double[myBitmap.Width, myBitmap.Height, 3];
for (int x = 0; x < myBitmap.Width; x++)
{
for (int y = 0; y < myBitmap.Height; y++)
{
Color pixelColor = myBitmap.GetPixel(x, y);
data[x, y, 0] = pixelColor.R;
data[x, y, 1] = pixelColor.G;
data[x, y, 2] = pixelColor.B;
}
}
return(data);
});
lApi.AddModule(new GaussianFilter());
lApi.AddModule(new MeanFilter());
double[,,] result = lApi.Run() as double[, , ];
Assert.True(result != null);
Bitmap blurBitmap = new Bitmap(result.GetLength(0), result.GetLength(1));
for (int x = 0; x < result.GetLength(0); x++)
{
for (int y = 0; y < result.GetLength(1); y++)
{
Color pixelColor = Color.FromArgb((int)result[x, y, 0], (int)result[x, y, 1], (int)result[x, y, 2]);
blurBitmap.SetPixel(x, y, pixelColor);
}
}
SaveImage(blurBitmap, "GaussianAndMean.jpg");
}
public void CRbm_ClassifierTest()
{
var dataPath = System.IO.Path.Combine(Directory.GetCurrentDirectory(), @"RestrictedBolzmannMachine2\Data\rbm_twoclass_sample.csv");
LearningApi api = new LearningApi(this.getDescriptorForRbmTwoClassesClassifier());
// Initialize data provider
api.UseCsvDataProvider(dataPath, ';', false, 1);
api.UseDefaultDataMapper();
double[] featureVector = new double[] { 0.1, 0.2, 0.3 };
api.UseCRbm(featureVector, 0.01, 1000, 10, 2);
RbmResult score = api.Run() as RbmResult;
double[][] testData = new double[5][];
//
// This test data contains two patterns. One is grouped at left and one at almost right.
testData[0] = new double[] { 1, 1, 1, 0, 0, 0, 0, 0, 0, 0 };
testData[1] = new double[] { 1, 0, 1, 0, 0, 0, 0, 0, 0, 0 };
testData[2] = new double[] { 0, 0, 0, 0, 0, 1, 1, 1, 0, 0 };
testData[3] = new double[] { 0, 0, 0, 0, 0, 1, 0, 1, 0, 0 };
// This will be classified as third class.
testData[4] = new double[] { 1, 1, 1, 0, 0, 1, 1, 1, 0, 0 };
var result = api.Algorithm.Predict(testData, api.Context) as RbmResult;
//
// 2 * BIT1 + BIT2 of [0] and [1] should be same.
// We don't know how RBM will classiffy data. We only expect that
// same or similar pattern of data will be assigned to the same class.
// Note, we have here two classes (two hiddne nodes).
// First and second data sample are of same class. Third and fourth are also of same class.
// Here we check first classs.
Assert.True(2 * result.HiddenNodesPredictions[0][0] + result.HiddenNodesPredictions[0][1] ==
2 * result.HiddenNodesPredictions[1][0] + result.HiddenNodesPredictions[1][1]);
// Here is test for second class.
Assert.True(2 * result.HiddenNodesPredictions[2][0] + result.HiddenNodesPredictions[2][1] ==
2 * result.HiddenNodesPredictions[3][0] + result.HiddenNodesPredictions[3][1]);
printVector("Weights", result.Weights);
}
public void LogisticsRegression_Test_iterations_10_learningrate_013()
{
var desc = loadMetaData();
LearningApi api = new LearningApi(desc);
//Real dataset must be defined as object type, because data can be numeric, binary and classification
api.UseActionModule <object[][], object[][]>((input, ctx) =>
{
return(loadRealDataSample());
});
// Use mapper for data, which will extract (map) required columns
api.UseDefaultDataMapper();
api.UseMinMaxNormalizer();
//run logistic regression for 10 iterations with learningRate=0.13
api.UseLogisticRegression(0.13, 10);
api.Run();
IScore score = api.GetScore();
//Errors during each iteration. IF the learningRate is suitable erros is descrising for every next iteration
Assert.Equal(Math.Round(score.Errors[0], 5), 0.24278);
Assert.Equal(Math.Round(score.Errors[1], 5), 0.23749);
Assert.Equal(Math.Round(score.Errors[2], 5), 0.23359);
Assert.Equal(Math.Round(score.Errors[3], 5), 0.23010);
Assert.Equal(Math.Round(score.Errors[4], 5), 0.22740);
Assert.Equal(Math.Round(score.Errors[5], 5), 0.22476);
Assert.Equal(Math.Round(score.Errors[6], 5), 0.22271);
Assert.Equal(Math.Round(score.Errors[7], 5), 0.22065);
Assert.Equal(Math.Round(score.Errors[8], 5), 0.21902);
Assert.Equal(Math.Round(score.Errors[9], 5), 0.21739);
//LG Model Best Found model in 10 iteration
Assert.Equal(Math.Round(score.Weights[0], 5), 0.06494);
Assert.Equal(Math.Round(score.Weights[1], 5), 0.21584);
Assert.Equal(Math.Round(score.Weights[2], 5), 0.89901);
Assert.Equal(Math.Round(score.Weights[3], 5), 0.51497);
Assert.Equal(Math.Round(score.Weights[4], 5), -0.30213);
Assert.Equal(Math.Round(score.Weights[5], 5), -0.30213);
Assert.Equal(Math.Round(score.Weights[6], 5), -0.85624);
}
public void CSVDataProviderTest_IrisData()
{
//
//iris data file
var isris_path = System.IO.Path.Combine(Directory.GetCurrentDirectory(), m_iris_data_path);
LearningApi api = new LearningApi(TestHelpers.GetDescriptor());
api.UseCsvDataProvider(isris_path, ',', false, 1);
var result = api.Run() as object[][];
//get expected result
var expected = GetReal_Iris_DataSet();
for (int i = 0; i < result.Length; i++)
{
for (int j = 0; j < result[0].Length; j++)
{
var col = api.Context.DataDescriptor.Features[j];
if (col.Type == ColumnType.STRING)
{
continue;
}
else if (col.Type == ColumnType.NUMERIC)//numeric column
{
var val1 = double.Parse((string)result[i][j], System.Globalization.NumberStyles.Any, CultureInfo.InvariantCulture);
var val2 = Convert.ToDouble(expected[i][j], CultureInfo.InvariantCulture);
Assert.Equal(val1, val2);
}
else if (col.Type == ColumnType.BINARY)//binary column
{
Assert.Equal(result[i][j].ToString(), expected[i][j].ToString());
}
else if (col.Type == ColumnType.CLASS)//class column
{
Assert.Equal(result[i][j].ToString(), expected[i][j].ToString());
}
}
}
return;
}
public void CSVDataProviderTest_SecomData()
{
//
//iris data file
var isris_path = System.IO.Path.Combine(Directory.GetCurrentDirectory(), m_secom_data_path);
LearningApi api = new LearningApi(TestHelpers.GetDescriptor(m_secom_data_mapper_path));
api.UseCsvDataProvider(isris_path, ',', false, 1);
var result = api.Run() as object[][];
//get expected result
var expected = GetReal_Secom_DataSet();
for (int i = 0; i < result.Length; i++)
{
for (int j = 0; j < result[0].Length; j++)
{
var col = api.Context.DataDescriptor.Features[j];
if (col.Type == ColumnType.STRING)
{
Assert.Equal(result[i][j], expected[i][j]);
}
else if (col.Type == ColumnType.NUMERIC)//numeric column
{
//var val1 = double.Parse(result[i][j].ToString());
//var val2 = double.Parse(expected[i][j].ToString());
Assert.Equal(result[i][j], expected[i][j]);
}
else if (col.Type == ColumnType.BINARY)//binary column
{
Assert.Equal(result[i][j].ToString(), expected[i][j].ToString());
}
else if (col.Type == ColumnType.CLASS)//class column
{
Assert.Equal(result[i][j].ToString(), expected[i][j].ToString());
}
}
}
return;
}
public void smileyTestDeepRbm(int iterations, double learningRate, int[] layers)
{
LearningApi api = new LearningApi(getDescriptorForRbm(1600));
// Initialize data provider
api.UseCsvDataProvider(Path.Combine(Directory.GetCurrentDirectory(), @"RestrictedBolzmannMachine2\Data\Smiley.csv"), ',', false, 0);
api.UseDefaultDataMapper();
api.UseDeepRbm(learningRate, iterations, layers);
Stopwatch watch = new Stopwatch();
watch.Start();
RbmDeepScore score = api.Run() as RbmDeepScore;
watch.Stop();
var testData = ReadData(Path.Combine(Directory.GetCurrentDirectory(), @"RestrictedBolzmannMachine2\Data\SmileyTest.csv"));
var result = api.Algorithm.Predict(testData, api.Context) as RbmDeepResult;
var accList = new double[result.Results.Count];
var predictions = new double[result.Results.Count][];
var predictedHiddenNodes = new double[result.Results.Count][];
var Time = watch.ElapsedMilliseconds / 1000;
int i = 0;
foreach (var item in result.Results)
{
predictions[i] = item.First().VisibleNodesPredictions;
predictedHiddenNodes[i] = item.Last().HiddenNodesPredictions;
accList[i] = testData[i].GetHammingDistance(predictions[i]);
i++;
}
var ValTest = calcDelta(predictions, testData);
var lossTest = ValTest / (layers.First());
Debug.WriteLine($"lossTest: {lossTest}");
WriteDeepResult(iterations, layers, accList, Time / 60.0, predictedHiddenNodes);
/// write predicted hidden nodes.......
WriteOutputMatrix(iterations, layers, predictions, testData);
}
public void DigitRecognitionDeepTest(int iterations, double learningRate, int[] layers)
{
Debug.WriteLine($"{iterations}-{String.Join("", layers)}");
LearningApi api = new LearningApi(getDescriptorForRbmTwoClassesClassifier(4096));
// Initialize data provider
// TODO: Describe Digit Dataset.
api.UseCsvDataProvider(Path.Combine(Directory.GetCurrentDirectory(), @"RestrictedBolzmannMachine2\Data\DigitDataset.csv"), ',', false, 0);
api.UseDefaultDataMapper();
api.UseDeepRbm(learningRate, iterations, layers);
Stopwatch watch = new Stopwatch();
watch.Start();
RbmDeepScore score = api.Run() as RbmDeepScore;
watch.Stop();
var testData = RbmHandwrittenDigitUnitTests.ReadData(Path.Combine(Directory.GetCurrentDirectory(), @"RestrictedBolzmannMachine2\Data\predictiondigitdata.csv"));
var result = api.Algorithm.Predict(testData, api.Context) as RbmDeepResult;
var accList = new double[result.Results.Count];
var predictions = new double[result.Results.Count][];
var predictedHiddenNodes = new double[result.Results.Count][];
var Time = watch.ElapsedMilliseconds / 1000;
int i = 0;
foreach (var item in result.Results)
{
predictions[i] = item.First().VisibleNodesPredictions;
predictedHiddenNodes[i] = item.Last().HiddenNodesPredictions;
accList[i] = testData[i].GetHammingDistance(predictions[i]);
i++;
}
RbmHandwrittenDigitUnitTests.WriteDeepResult(iterations, layers, accList, Time * 1000, predictedHiddenNodes);
/// write predicted hidden nodes.......
RbmHandwrittenDigitUnitTests.WriteOutputMatrix(iterations, layers, predictions, testData);
}
public void Test_OptimalNumberOfCLusters()
{
// directory to load
string loadDirectory = rootFolder + "Functions\\";
string FunctionName = "SIN X"; //without extension
string savePath = rootFolder + "Optimal Clusters\\" + FunctionName + " Results.csv";
double[][] function = Helpers.LoadFunctionData(loadDirectory + FunctionName + "\\" + FunctionName + ".csv");
function = TestFunctionGenerators.normalizeData(function);
int numAttributes = 2; // 2 in this demo (height,weight)
int numClusters = 0; // vary this to experiment (must be between 2 and number data tuples)
int maxCount = 300; // trial and error
ClusteringSettings settings = new ClusteringSettings(maxCount, numClusters, numAttributes, KmeansAlgorithm: 2);
// Creates learning api object
LearningApi api = new LearningApi();
api.UseActionModule <object, double[][]>((data, ctx) =>
{
return(KMeansAlgorithm.transposeFunction(function));
});
api.UseKMeans(settings);
// train
var resp = api.Run() as KMeansScore;
Assert.True(resp.Model.NumberOfClusters > 1);
double[][] OptimalClustersResults = new double[4][];
OptimalClustersResults[0] = new double[] { 2, 3, 4, 5, 6, 7, 8, 9, 10 };
OptimalClustersResults[1] = resp.Model.D;
OptimalClustersResults[2] = resp.Model.DPrime;
OptimalClustersResults[3] = resp.Model.Fmin;
Helpers.Write2CSVFile(OptimalClustersResults, savePath);
// implement
}
