public static void DisplayTestDataResults(Classifier.MLDotNet.AffirmativeClassifier model, string testDataPath)
{
var dataView = System.IO.File.ReadAllLines(testDataPath);
Console.WriteLine("=============== Displaying Model Results with Test data===============");
foreach (var item in dataView)
{
var elements = item.Split(',');
if (elements[0].ToLower() != "isaffirmative")
{
var prediction = model.Predict(elements[1]);
DisplayPredictionResults(elements[1], prediction.Prediction, prediction.Probability);
}
}
Console.WriteLine();
}
static void Main(string[] args)
{
bool train = true;
Classifier.MLDotNet.AffirmativeClassifier model;
if (train)
{
// Train a new model using training data
model = new Classifier.MLDotNet.AffirmativeClassifier(_trainDataPath, _testDataPath);
// Store the trained model
var savedModel = model.Export();
var encodedModel = Convert.ToBase64String(savedModel, Base64FormattingOptions.None);
File.WriteAllText(_base64ModelPath, encodedModel);
Console.WriteLine($"Model trained using: {model.TrainingMethodName}");
Console.WriteLine($"Encoded model stored in '{_base64ModelPath}'\r\n");
}
else
{
// Load an existing model
byte[] modelData = System.Convert.FromBase64String(_base64EncodedModel);
model = new Classifier.MLDotNet.AffirmativeClassifier(modelData);
}
var evaluationResults = model.Evaluate(_testDataPath);
DisplayEvaluationResults(evaluationResults.Accuracy, evaluationResults.Auc, evaluationResults.F1Score);
DisplayTestDataResults(model, _testDataPath);
string input = string.Empty;
do
{
Console.Write("\n\nUtterance: ");
input = Console.ReadLine();
if (input != string.Empty)
{
var result = model.Predict(input);
DisplayPredictionResults(input, result.Prediction, result.Probability);
}
}while (input != string.Empty);
}
