//public static void ClassifyAndValidateOne(string trainFile, string sampleClassifiedFile, string classificationFile, string validationFile)
//1412520
public static void ClassifyAndValidateOne()
{
var testsWithValueType = FileIO.ReadFile("../../validation/testTarget.txt");
var testRecords = RemoveLabelOfList(testsWithValueType);
FileIO.WriteFile(testRecords, "../../validation/test.txt");
Bow_tfidf.GenerateTFIDFMatrix("../../validation/train.txt", "../../validation/processed.txt", "../../validation/features.txt", "../../validation/tf_idf.txt");
Model.classifyForValidate();
Validate("../../validation/testTarget.txt", "../../validation/testResult.txt", "../../validation/result.txt");
}
// 1412542
// Date: 08/04/2018
// Cross validation
public static void Validate(string kFoldsFile, string inputFile)
{
List <string> fileNameList = FileUtils.getAllFileNames("../../splited_data");
String trainFile = "../../validation/train.txt";
String processedFile = "../../validation/processed.txt";
String featuresFile = "../../validation/features.txt";
String tfIdfFile = "../../validation/tf_idf.txt";
String testFile = "../../validation/test.txt";
String testTargetFile = "../../validation/testTarget.txt";
String testResultFile = "../../validation/testResult.txt";
// Read the number of subset
int numberOfFolds = int.Parse(FileIO.ReadFile(kFoldsFile)[0]);
/*
* // Read data
* List<string> data = FileIO.ReadFile(inputFile);
*
* // Split data into subsets
* SplitTrainTest(data, numberOfFolds);
*/
// Read subsets
List <List <string> > subset = new List <List <string> >();
foreach (string fileName in fileNameList)
{
subset.Add(FileIO.ReadFile(fileName));
}
//calculate avg Fmicro and avg Fmacro
double sum_Fmicro = 0;
double sum_Fmacro = 0;
// build model and test by k-folds cross validation
for (int i = 0; i < numberOfFolds; i++)
{
var watch = System.Diagnostics.Stopwatch.StartNew();
Console.WriteLine("Fold {0}: ", i);
//subset[i] is test set, others are training sets
List <string> trainingSet = new List <string>();
List <string> testTarget = new List <string>();
for (int subIndex = 0; subIndex < numberOfFolds; subIndex++)
{
if (subIndex != i)
{
trainingSet.AddRange(subset[subIndex]);
}
else
{
testTarget = subset[i].ToList();
}
}
Console.WriteLine("The time to create training and test sets: {0} ", watch.ElapsedMilliseconds);
FileIO.WriteFile(testTarget, testTargetFile);
//Remove label of testSetTarget
List <Vector> targetVector = new List <Vector>();
var testSet = FileIO.ReadFileIntoVector(testTargetFile, out targetVector, true);
//testSet is removed label
//Write testSet and trainSet into file
FileIO.WriteFile(trainingSet, trainFile);
FileIO.WriteFile(testSet, testFile);
Console.WriteLine("The time to write training and test sets: {0} ", watch.ElapsedMilliseconds);
Bow_tfidf.GenerateTFIDFMatrix(trainFile, processedFile, featuresFile, tfIdfFile);
Console.WriteLine("The time to calculate tf_idf: {0} ", watch.ElapsedMilliseconds);
Model.classifyForValidate();
List <Vector> sourceVector = new List <Vector>();
FileIO.ReadFileIntoVector(testResultFile, out sourceVector, true);
Double Rmacro = calculateRmacro(sourceVector, targetVector);
Double Pmacro = CalculatePmacro(sourceVector, targetVector);
sum_Fmacro += calculateFmacroOrFscore(Rmacro, Pmacro);
sum_Fmicro += calculateFmicro(sourceVector, targetVector);
watch.Stop();
var elapsedMs = watch.ElapsedMilliseconds;
Console.WriteLine("Validating the fold {0} takes: {1} ", i, elapsedMs);
}
List <string> F_array = new List <string>();
double avg_Fmacro = (1.0 * sum_Fmicro) / numberOfFolds;
double avg_Fmicro = (1.0 * sum_Fmacro) / numberOfFolds;
F_array.Add(avg_Fmacro.ToString());
F_array.Add(avg_Fmicro.ToString());
//Write avg Fmacro and avg Fmicro into file
FileIO.WriteFile(F_array, "../../validation/result.txt");
}
