public IPredictImage Predict(IPredict predict, Action <IPredictConfiguration> configuration)
{
PredictConfiguration predictConfiguration;
configuration(predictConfiguration = new PredictConfiguration(predict));
return(predictConfiguration.Predict(_scanImage, predictConfiguration.Alphabet, predictConfiguration.Options));
}
public void AddPredictor(IPredict p, string label)
{
_predicts.Add(p);
_predictReporter.Add(new Abcd()
{
Label = label
});
}
async void showPrediction()
{
IPredict predictions = DependencyService.Get <IPredict>();
string a = await Task.Run(() => predictions.getPrediction(file));
string value = a;
System.Console.WriteLine(value);
}
public PredictorShould()
{
var dataFixture = new DataFixture();
dataFixture.Init();
_predict = new Predict();
_textClassifier =
new TextMatcher(dataFixture.Companies, dataFixture.Localities, dataFixture.Classifiers)
{
UseCache = false
};
_account = dataFixture.GetTestBankAccount();
}
private void buttonGenerate_Click(object sender, EventArgs e)
{
if (_data == null)
{
XtraMessageBox.Show("There is no data create model!", "Uh Oh!", MessageBoxButtons.OK, MessageBoxIcon.Warning);
return;
}
var model = new PerceptronModel <Student>();
_classifier = model.Generate(_data);
XtraMessageBox.Show("Done!", "Success!", MessageBoxButtons.OK, MessageBoxIcon.Information);
}
public async Task PredictAsync(PredictBankStatement2Command command, IPredict predict)
{
var results = await predict.PredictManyAsync(State.PredictionRequests);
var predictionResults = results as PredictionResult[] ?? results.ToArray();
if (predictionResults.Any(x => x == null))
{
throw new ApplicationException();
}
var ev = new BankStatementPredicted2Event
{
PredictionResults = predictionResults
};
Emit(ev);
await Task.CompletedTask;
}
public static void DrawGrid(double[] extent, double resolution, IPredict predict, string fileName)
{
int width = (int)Math.Round((extent[2] - extent[0]) / resolution) + 1, height = (int)Math.Round((extent[3] - extent[1]) / resolution) + 1;
Bitmap bitmap = new Bitmap(width, height);
double lon, lat;
for (int i = 0; i < width; i++)
{
lon = extent[0] + i * resolution;
for (int j = 0; j < height; j++)
{
lat = extent[3] - j * resolution;
int value = (int)Math.Round(predict.Predict(lon, lat));
bitmap.SetPixel(i, j, Color.FromArgb(value / 256, value % 256, 0));
}
}
bitmap.Save(fileName);
bitmap.Dispose();
}
public PredictBankStatement2CommandHandler(IPredict predict)
{
_predict = predict;
}
public PredictConfiguration(IPredict predict)
{
Predict = predict.PredictModel;
}
public PredictionController(IPredict predictor)
{
_predictor = predictor;
}
public void Learn(IEnumerable <T> examples)
{
Examples = examples;
var total = Examples.Count();
// 80% for training
var trainingCount = (int)System.Math.Ceiling(total * .8);
// 20% for testing
var testingSlice = GetTestPoints(total - trainingCount, total)
.ToArray();
var trainingSlice = GetTrainingPoints(testingSlice, total);
// getting data
var data = Converter.Convert <T>(examples);
Matrix x = data.Item1;
Vector y = data.Item2;
TypeDescription description = data.Item3;
// training
var x_t = x.Slice(trainingSlice, VectorType.Row);
var y_t = y.Slice(trainingSlice);
Predictors = new IPredict <T> [Models.Length];
// run in parallel since they all have
// read-only references to the data model
// and update independently to different
// spots
//Parallel.For(0, Models.Length, i =>
//{
for (int i = 0; i < Models.Length; i++)
{
Models[i].X = x_t;
Models[i].Y = y_t;
Models[i].Description = description;
Predictors[i] = Models[i].Generate();
}
//});
// testing
T[] test = GetTestExamples(testingSlice, examples);
Accuracy = Vector.Zeros(Predictors.Length);
for (int i = 0; i < Predictors.Length; i++)
{
Accuracy[i] = 0;
for (int j = 0; j < test.Length; j++)
{
var truth = Conversion.Converter.GetItem <T>(test[j], description.Label);
Predictors[i].Predict(test[j]);
var pred = Conversion.Converter.GetItem <T>(test[j], description.Label);
// need to set it back...
Converter.SetItem <T>(test[j], description.Label, truth);
if (Object.Equals(truth, pred))
{
Accuracy[i] += 1;
}
}
Accuracy[i] /= test.Length;
}
;
}
