static void Main(string[] args)
{
// reading the data
const string dataFilePath = @"E:\training data\testdata.manual.2009.06.14.csv";
var dataTable = DataTable.New.ReadCsv(dataFilePath);
var dataTablePhones = dataTable.Rows.Select(row => row).ToList(); //.Where(row => row["Tweet"].ToString().Contains("Phone")).ToList();
List<string> x = dataTablePhones.Select(row => row["Tweet"]).ToList();
double[] y = dataTablePhones.Select(row => double.Parse(row["Sentiment"])).ToArray();
// creating vocabulary
var vocabulary = x.SelectMany(GetWords).Distinct().OrderBy(word => word).ToList();
// generating the problem
var problemBuilder = new TextClassificationProblemBuilder();
var problem = problemBuilder.CreateProblem(x, y, vocabulary.ToList());
//creating and training SVM model
const int C = 1;
var model = new C_SVC(problem, KernelHelper.LinearKernel(), C);
// prediction
string userInput;
var _predictionDictionary = new Dictionary<int, string> { { 0, "Bad" }, { 4, "Good" }, {2, "Neutral"} };
do
{
userInput = Console.ReadLine();
var newX = TextClassificationProblemBuilder.CreateNode(userInput, vocabulary);
var predictedY = model.Predict(newX);
Console.WriteLine("{0}", _predictionDictionary[(int)predictedY]);
Console.WriteLine(new string('=', 50));
} while (userInput != "quit");
}
public static void ClassifyBySVM(string trainFile, string testFile, string testTarget)
{
string testResultFile = "../../svm/testResult.txt";
var watch = System.Diagnostics.Stopwatch.StartNew();
//STEP 1 : READ DATA
List <Vector> vectorsTrain = new List <Vector>();
var content = FileIO.ReadFileIntoVector(trainFile, out vectorsTrain, true);
var typeClass = Vector.GetDistinctClassTypes(vectorsTrain);
//Get content of document and lable of document
double[] label = GetLableOfDocument(vectorsTrain);
//Get features list
var features = content.SelectMany(GetWords).Distinct().OrderBy(word => word).ToList();
//STEP 2: Generate a problem
var problem = TextClassificationProblemBuilder.CreateProblem(content, label, features.ToList());
//STEP 3: Create and train a SVM model
const int C = 1;
var model = new C_SVC(problem, KernelHelper.LinearKernel(), C);
//STEP 4: Predict
List <string> test = FileIO.ReadFile(testFile);
List <string> resultList = new List <string>();
_predictionDictionary = new Dictionary <int, string>();
List <Vector> targetVector = new List <Vector>();
FileIO.ReadFileIntoVector(testTarget, out targetVector, true);
for (int l = 0; l < typeClass.Count(); l++)
{
_predictionDictionary.Add(l, typeClass.ElementAt(l));
}
for (int i = 0; i < test.Count(); i++)
{
var newX = TextClassificationProblemBuilder.CreateNode(test[i], features);
var predictedY = model.Predict(newX);
var result = _predictionDictionary[(int)predictedY];
resultList.Add(result + " - " + test[i]);
}
FileIO.WriteFile(resultList, testResultFile);
List <Vector> sourceVector = new List <Vector>();
FileIO.ReadFileIntoVector(testResultFile, out sourceVector, true);
double score = 0;
for (int i = 0; i < typeClass.Count(); i++)
{
score = 1.0 * Vector.CountShareSameTypeRecords(typeClass.ElementAt(i), sourceVector, targetVector) / Vector.CountClassElements(typeClass.ElementAt(i), targetVector);
Console.WriteLine("correct label: " + Vector.CountShareSameTypeRecords(typeClass.ElementAt(i), sourceVector, targetVector));
Console.WriteLine("total label: " + Vector.CountClassElements(typeClass.ElementAt(i), targetVector));
Console.WriteLine("SVM score: " + score);
}
Console.WriteLine("The time for SVM: {0} ", watch.ElapsedMilliseconds);
}
