/// <summary>
/// Trains a model with a collection of emails and a flag to indicate invalidity.
/// </summary>
/// <param name="emailsAndValidities">A collection of emails and whether this email is invalid</param>
/// <param name="iterations">The number of iterations for the training</param>
/// <param name="cut">The cut for the training</param>
/// <returns>The trained GisModel</returns>
public static GisModel TrainModel(IEnumerable <EmailAndValidity> emailsAndValidities, int iterations, int cut)
{
var trainer = new GisTrainer();
var eventReader = new InvalidEmailDetectionDataEventReader(emailsAndValidities);
trainer.TrainModel(eventReader, iterations, cut);
return(new GisModel(trainer));
}
/// <summary>
/// Trains a tokenizer model from input files well formatted for
/// a token event reader.
/// </summary>
/// <param name="inputFiles">The collection of training input files</param>
/// <param name="iterations">The number of iterations to run when training the model</param>
/// <param name="cut">The minimum nb of occurences for statistical relevancy in the trained model</param>
/// <param name="splitMarker">The character indicating a split in the files</param>
/// <returns>The freshly trained GisModel</returns>
public static GisModel Train(IEnumerable <string> inputFiles, int iterations, int cut, char splitMarker = '|')
{
var trainer = new GisTrainer(0.1);
foreach (var inputFile in inputFiles)
{
var dataReader = new StreamReader(inputFile);
var eventReader = new TokenEventReader(dataReader, splitMarker);
trainer.TrainModel(iterations, new TwoPassDataIndexer(eventReader, cut));
}
return(new GisModel(trainer));
}
public static GisModel TrainModel(IEnumerable <string> filePaths, int iterations, int cut, IEndOfSentenceScanner scanner)
{
var trainer = new GisTrainer();
var readers = filePaths.Select(path => new StreamReader(path)).ToList();
// train the model
ITrainingDataReader <string> dataReader = new MultipleFilesPlainTextByLineDataReader(readers);
ITrainingEventReader eventReader = new SentenceDetectionEventReader(dataReader, scanner);
trainer.TrainModel(eventReader, iterations, cut);
return(new GisModel(trainer));
}
/// <summary>
/// Trains a model with a collection of input files with the following format:
/// [email protected] 0
/// mqsldkqsmlqsmdklqs@sdlsqjd 1
/// ...
/// The first line represents a valid email, the second an invalid.
/// </summary>
/// <param name="filePaths">The collection of file paths</param>
/// <param name="iterations">The number of iterations for the training</param>
/// <param name="cut">The cut for the training</param>
/// <returns>The trained GisModel</returns>
public static GisModel TrainModel(IEnumerable <string> filePaths, int iterations, int cut)
{
var trainer = new GisTrainer();
foreach (var file in filePaths)
{
using (var streamReader = new StreamReader(file))
{
ITrainingDataReader <string> dataReader = new PlainTextByLineDataReader(streamReader);
ITrainingEventReader eventReader = new InvalidEmailDetectionEventReader(dataReader);
trainer.TrainModel(eventReader, iterations, cut);
}
}
return(new GisModel(trainer));
}
public static GisModel TrainModel(IEnumerable <string> files, int iterations, int cut, IEndOfSentenceScanner scanner)
{
var trainer = new GisTrainer();
foreach (var file in files)
{
using (var streamReader = new StreamReader(file))
{
ITrainingDataReader <string> dataReader = new PlainTextByLineDataReader(streamReader);
ITrainingEventReader eventReader = new SentenceDetectionEventReader(dataReader, scanner);
trainer.TrainModel(eventReader, iterations, cut);
}
}
return(new GisModel(trainer));
}
/// <summary>
/// Trains a tokenizer model from input files well formatted for
/// a token event reader.
/// </summary>
/// <param name="inputFilePaths">The collection of training input files</param>
/// <param name="iterations">The number of iterations to run when training the model</param>
/// <param name="cut">The minimum nb of occurences for statistical relevancy in the trained model</param>
/// <param name="splitMarker">The character indicating a split in the files</param>
/// <returns>The freshly trained GisModel</returns>
public static GisModel Train(IEnumerable <string> inputFilePaths, int iterations, int cut, char splitMarker = '|', bool includeAllCapsExamples = false)
{
var trainer = new GisTrainer(0.1);
var dataReaders = new List <StreamReader>();
foreach (var path in inputFilePaths)
{
var dataReader = new StreamReader(path);
dataReaders.Add(dataReader);
}
// train the model
var eventReader = new MultipleFileTokenEventReader(dataReaders, splitMarker, includeAllCapsExamples);
trainer.TrainModel(iterations, new TwoPassDataIndexer(eventReader, cut));
return(new GisModel(trainer));
}
/// <summary>
/// Trains a tokenizer model from input files well formatted for
/// a token event reader.
/// </summary>
/// <param name="inputFilePaths">The collection of training input files</param>
/// <param name="iterations">The number of iterations to run when training the model</param>
/// <param name="cut">The minimum nb of occurences for statistical relevancy in the trained model</param>
/// <param name="splitMarker">The character indicating a split in the files</param>
/// <returns>The freshly trained GisModel</returns>
public static GisModel Train(IEnumerable<string> inputFilePaths, int iterations, int cut, char splitMarker = '|', bool includeAllCapsExamples = false)
{
var trainer = new GisTrainer(0.1);
var dataReaders = new List<StreamReader>();
foreach (var path in inputFilePaths)
{
var dataReader = new StreamReader(path);
dataReaders.Add(dataReader);
}
// train the model
var eventReader = new MultipleFileTokenEventReader(dataReaders, splitMarker, includeAllCapsExamples);
trainer.TrainModel(iterations, new TwoPassDataIndexer(eventReader, cut));
return new GisModel(trainer);
}
/// <summary>
/// Trains a tokenizer model from input files well formatted for
/// a token event reader.
/// </summary>
/// <param name="inputFiles">The collection of training input files</param>
/// <param name="iterations">The number of iterations to run when training the model</param>
/// <param name="cut">The minimum nb of occurences for statistical relevancy in the trained model</param>
/// <param name="splitMarker">The character indicating a split in the files</param>
/// <returns>The freshly trained GisModel</returns>
public static GisModel Train(IEnumerable<string> inputFiles, int iterations, int cut, char splitMarker = '|')
{
var trainer = new GisTrainer(0.1);
foreach (var inputFile in inputFiles)
{
var dataReader = new StreamReader(inputFile);
var eventReader = new TokenEventReader(dataReader, splitMarker);
trainer.TrainModel(iterations, new TwoPassDataIndexer(eventReader, cut));
}
return new GisModel(trainer);
}
public static GisModel TrainModel(IEnumerable<string> filePaths, int iterations, int cut, IEndOfSentenceScanner scanner)
{
var trainer = new GisTrainer();
var readers = filePaths.Select(path => new StreamReader(path)).ToList();
// train the model
ITrainingDataReader<string> dataReader = new MultipleFilesPlainTextByLineDataReader(readers);
ITrainingEventReader eventReader = new SentenceDetectionEventReader(dataReader, scanner);
trainer.TrainModel(eventReader, iterations, cut);
return new GisModel(trainer);
}
public static GisModel TrainModel(IEnumerable<string> files, int iterations, int cut, IEndOfSentenceScanner scanner)
{
var trainer = new GisTrainer();
foreach (var file in files)
{
using (var streamReader = new StreamReader(file))
{
ITrainingDataReader<string> dataReader = new PlainTextByLineDataReader(streamReader);
ITrainingEventReader eventReader = new SentenceDetectionEventReader(dataReader, scanner);
trainer.TrainModel(eventReader, iterations, cut);
}
}
return new GisModel(trainer);
}
