public Task <RegressionPrediction> Predict(RegressionData data)
{
return(Task.Run(() =>
{
return _model.Predict(data);
}));
}
public static void InitRegressionPrediction()
{
// load the mlContext and data path
regressionPredict.LoadData(mlContext, regressionData);
// create the model, boolean to force model retrain
regressionPredict.BuildAndTrainModel();
// evaluate the model against test data
regressionPredict.Evaluate();
// test output with a test method
var sample =
new RegressionData {
VendorId = "VTS",
RateCode = "1",
PassengerCount = 1,
TripTime = 900,
TripDistance = 8.63f,
PaymentType = "CRD",
FareAmount = 0 // To predict. Actual/Observed = 15.5
};
regressionPredict.TestModel(sample);
}
public void Predict(MLContext mlContext, ITransformer model, RegressionData dataSample)
{
var predictionFunction = mlContext.Model.CreatePredictionEngine <RegressionData, RegressionPrediction>(model);
var prediction = predictionFunction.Predict(dataSample);
Console.WriteLine($"**********************************************************************");
// Console.WriteLine($"Predicted fare: {prediction.FareAmount:0.####}, actual fare: 15.5");
Console.WriteLine($"Prediction: {prediction.Result:0.####}");
Console.WriteLine($"**********************************************************************");
}
public void TestModel(RegressionData sample)
{
// var taxiTripSample = new RegressionData()
// {
// Date = ("2/01/2023"),
// VendorId = "VTS",
// RateCode = "1",
// PassengerCount = 1,
// TripTime = 1140,
// TripDistance = tripDistance,
// PaymentType = "CRD",
// FareAmount = 0 // To predict. Actual/Observed = 15.5
//};
var predictionFunction = mlContext.Model.CreatePredictionEngine <RegressionData, RegressionPrediction>(trainedModel);
var prediction = predictionFunction.Predict(sample);
Console.WriteLine($"**********************************************************************");
Console.WriteLine($"Prediction: {prediction.Result:0.####}");
Console.WriteLine($"**********************************************************************");
}
RegressionData calculateRegression(List<double> coords)
{
double xySum = 0;
double xSquaredSum = 0;
double ySquaredSum = 0;
double ySum = 0;
double xSum = 0;
double mean;
for (int i = 1; i < coords.Count; i++)
{
double x = i;
double xSquared = x * x;
double y = coords[i - 1];
double ySquared = y * y;
double xy = x * y;
xSum += x;
ySum += y;
xySum += xy;
xSquaredSum += xSquared;
ySquaredSum += ySquared;
}
double a = ( (ySum * xSquaredSum) - (xSum * xySum) ) / ( (coords.Count * xSquaredSum) - (xSum * xSum) );
double b = ((coords.Count * xySum) - (xSum * ySum) ) / ( ( coords.Count * xSquaredSum) - (xSum * xSum) );
//Determine regression line
// y = a + bx
List<double> yData = new List<double>();
for (int i = 1; i < coords.Count; i++)
{
double x = i;
double y = a + (b * x);
yData.Add(y);
}
RegressionData regData = new RegressionData();
regData.yVals = yData;
//Calculate standard deviation of regression line
ySum = 0;
//First calculate sum
for (int i = 1; i < yData.Count; i++)
{
ySum += yData[i - 1];
}
mean = ySum / yData.Count;
//Next calculate deviations
List<double> deviations = new List<double>();
double devSquaredSum = 0;
for (int i = 1; i < yData.Count; i++)
{
double dev = yData[i - 1] - mean;
double devSquared = dev * dev;
devSquaredSum += devSquared;
}
double standardDev = System.Math.Sqrt(devSquaredSum / (yData.Count - 1));
regData.a = a;
regData.b = b;
regData.standardDev = standardDev;
regData.yVals = yData;
return regData;
}
