private static void TestSinglePrediction(MLContext mlContext)
{
ITransformer loadedModel;
using (var stream = new FileStream(_modelPath, FileMode.Open, FileAccess.Read, FileShare.Read))
{
loadedModel = mlContext.Model.Load(stream);
}
// PredictionEngine is a "wrapper" around a model that allows predictions on individual examples (most common scenario in production for consumer-scale applications? Or is batching more common?
var predictionFunction = loadedModel.CreatePredictionEngine <TaxiTrip, TaxiTripFarePrediction>(mlContext);
// TODO: Parameterise
var taxiTripSample = new TaxiTrip()
{
VendorId = "VTS",
RateCode = "1",
PassengerCount = 1,
TripTime = 1140,
TripDistance = 3.75f,
PaymentType = "CRD",
FareAmount = 0 // To predict. Actual/Observed = 15.5
};
var prediction = predictionFunction.Predict(taxiTripSample);
//TODO: PrintBanner again
Console.WriteLine($"**********************************************************************");
Console.WriteLine($"Predicted fare: {prediction.FareAmount:0.####}, actual fare: 15.5");
Console.WriteLine($"**********************************************************************");
}
private static void TestSinglePrediction(MLContext mlContext, ITransformer model)
{
//Prediction test
// Create prediction function and make prediction.
var predictionFunction = mlContext.Model.CreatePredictionEngine <TaxiTrip, TaxiTripFarePrediction>(model);
//Sample:
//vendor_id,rate_code,passenger_count,trip_time_in_secs,trip_distance,payment_type,fare_amount
//VTS,1,1,1140,3.75,CRD,15.5
var taxiTripSample = new TaxiTrip()
{
VendorId = "VTS",
RateCode = "1",
PassengerCount = 1,
TripTime = 1140,
TripDistance = 3.75f,
PaymentType = "CRD",
FareAmount = 0 // To predict. Actual/Observed = 15.5
};
var prediction = predictionFunction.Predict(taxiTripSample);
Console.WriteLine($"**********************************************************************");
Console.WriteLine($"Predicted fare: {prediction.FareAmount:0.####}, actual fare: 15.5");
Console.WriteLine($"**********************************************************************");
}
private static void TestSinglePrediction(MLContext mlContext)
{
ITransformer loadedModel;
Microsoft.ML.DataViewSchema inputSchema;
using (var stream = new FileStream(_modelPath, FileMode.Open, FileAccess.Read, FileShare.Read))
{
loadedModel = mlContext.Model.Load(stream, out inputSchema);
}
var predictionFunction = mlContext.Model.CreatePredictionEngine <TaxiTrip, TaxiTripFarePrediction>(loadedModel);
var taxiTripSample = new TaxiTrip()
{
VendorId = "VTS",
RateCode = "1",
PassengerCount = 1,
TripTime = 1140,
TripDistance = 3.75f,
PaymentType = "CRD",
FareAmount = 0 // To predict. Actual/Observed = 15.5
};
var prediction = predictionFunction.Predict(taxiTripSample);
Console.WriteLine($"**********************************************************************");
Console.WriteLine($"Predicted fare: {prediction.FareAmount:0.####}, actual fare: 15.5");
Console.WriteLine($"**********************************************************************");
}
static void Main(string[] args)
{
MLContext mainContext = new MLContext();
// RegressionContext regressionContext = new RegressionContext(mainContext);
EasyML MLConstruction = new EasyML("RegressionModel", mainContext);
IEstimator <ITransformer> pipeline = mainContext.Transforms.CopyColumns("FareAmount", "Label")
.Append(mainContext.Transforms.Categorical.OneHotEncoding("VendorId"))
.Append(mainContext.Transforms.Categorical.OneHotEncoding("PaymentType"))
.Append(mainContext.Transforms.Concatenate("Features", "VendorId", "RateCode", "PassengerCount", "TripTimeInSecs", "TripDistance", "PaymentType"))
.Append(mainContext.Regression.Trainers.FastTree());
var model = MLConstruction.CreateFittedModel(pipeline);
// MLConstruction.SaveModelAsFile(model);
TaxiTrip TestInstance = new TaxiTrip()
{
VendorId = "VTS",
RateCode = 1,
PassengerCount = 1,
TripTimeInSecs = 1140,
TripDistance = 3.75f,
PaymentType = "CRD",
FareAmount = 0.0f // To predict. Actual/Observed = 15.5
};
var testScore = MLConstruction.TestSinglePrediction <TaxiTrip, TaxiTripFarePrediction>(TestInstance, model);
Console.WriteLine($"{testScore.FareAmount}");
}
private static void TestSinglePrediction(MLContext mlContext)
{
//vendor_id,rate_code,passenger_count,trip_time_in_secs,trip_distance,payment_type,fare_amount
//VTS,1,1,1140,3.75,CRD,15.5
var taxiTripSample = new TaxiTrip
{
VendorId = "VTS",
RateCode = "1",
PassengerCount = 1,
TripTime = 1140,
TripDistance = 3.75f,
PaymentType = "CRD",
FareAmount = 0 // To predict. Actual/Observed = 15.5
};
// Load trained model
var trainedModel = mlContext.Model.Load(ModelPath, out var modelInputSchema);
// Create prediction engine related to the loaded trained model
var predEngine = mlContext.Model.CreatePredictionEngine <TaxiTrip, TaxiTripFarePrediction>(trainedModel);
//Score
var resultprediction = predEngine.Predict(taxiTripSample);
Console.WriteLine($"**********************************************************************");
Console.WriteLine($"Predicted fare: {resultprediction.FareAmount:0.#####}, actual fare: 15.5");
Console.WriteLine($"**********************************************************************");
}
public static async Task <TaxiTripFarePrediction> Predict(TaxiTrip trip)
{
PredictionModel <TaxiTrip, TaxiTripFarePrediction> _model = await PredictionModel.ReadAsync <TaxiTrip, TaxiTripFarePrediction>(_modelpath);
var prediction = _model.Predict(trip);
return(prediction);
}
/// <summary>
/// 使用模型來進行預測
/// TestSinglePrediction 方法會執行下列工作：
/// * 建立單一評論的測試資料。
/// * 根據測試資料預測費用金額。
/// * 合併測試資料和預測來進行報告。
/// * 顯示預測的結果。
/// </summary>
/// <param name="mlContext"></param>
/// <param name="model"></param>
private static void TestSinglePrediction(MLContext mlContext, ITransformer model)
{
var predictionFunction = mlContext.Model.CreatePredictionEngine <TaxiTrip, TaxiTripFarePrediction>(model);
//測試用的計程車行程
var taxiTripSample = new TaxiTrip()
{
VendorId = "VTS",
RateCode = "1",
PassengerCount = 1,
TripTime = 1140,
TripDistance = 3.75f,
PaymentType = "CRD",
FareAmount = 0 // To predict. Actual/Observed = 15.5
};
var taxiTripSample2 = new TaxiTrip()
{
VendorId = "VTS",
RateCode = "1",
PassengerCount = 1,
TripTime = 1260,
TripDistance = 10.33f,
PaymentType = "CSH",
FareAmount = 0 // To predict. Actual/Observed = 29.5
};
var taxiTripSample3 = new TaxiTrip()
{
VendorId = "VTS",
RateCode = "1",
PassengerCount = 3,
TripTime = 480,
TripDistance = 1.9f,
PaymentType = "CRD",
FareAmount = 0 // To predict. Actual/Observed = 8.5
};
var prediction = predictionFunction.Predict(taxiTripSample);
var prediction2 = predictionFunction.Predict(taxiTripSample2);
var prediction3 = predictionFunction.Predict(taxiTripSample3);
Console.WriteLine($"**********************************************************************");
Console.WriteLine($"預測: {prediction.FareAmount:0.####}, 實際: 15.5, 誤差 " + Math.Round((Math.Abs(15.5 - prediction.FareAmount)) / 15.5, 4) + "%");
Console.WriteLine($"預測: {prediction2.FareAmount:0.####}, 實際: 29.5, 誤差 " + Math.Round((Math.Abs(29.5 - prediction.FareAmount)) / 29.5, 4) + "%");
Console.WriteLine($"預測: {prediction3.FareAmount:0.####}, 實際: 8.5, 誤差 " + Math.Round(Math.Abs(8.5 - prediction.FareAmount) / 8.5, 4) + "%");
Console.WriteLine($"**********************************************************************");
}
private static void TestTwoPredictions(MLContext mlContext)
{
//load the model
ITransformer loadedModel;
using (var stream = new FileStream(_modelPath, FileMode.Open, FileAccess.Read, FileShare.Read))
{
loadedModel = mlContext.Model.Load(stream);
}
//Prediction test
// Create prediction function and make prediction.
var predictionFunction = loadedModel.CreatePredictionEngine <TaxiTrip, TaxiTripFarePrediction>(mlContext);
//Sample:
//vendor_id,rate_code,passenger_count,trip_time_in_secs,trip_distance,payment_type,fare_amount
//VTS,1,1,1140,3.75,CRD,15.5
var taxiTripSample = new TaxiTrip()
{
VendorId = "VTS",
RateCode = "1",
PassengerCount = 4,
TripTime = 1140,
TripDistance = 3.75f,
PaymentType = "CRD",
FareAmount = 0 // To predict. Actual/Observed = 15.5
};
var prediction = predictionFunction.Predict(taxiTripSample);
WriteLine($"**********************************************************************");
WriteLine($"Predicted fare: {prediction.FareAmount:0.####}, actual fare: 15.5");
WriteLine($"**********************************************************************\n");
// VTS 1 2 1140 6.29 CRD 21
taxiTripSample = new TaxiTrip()
{
VendorId = "VTS",
RateCode = "1",
PassengerCount = 2,
TripTime = 1140,
TripDistance = 6.29f,
PaymentType = "CRD",
FareAmount = 0 // To predict. Actual/Observed = 21
};
prediction = predictionFunction.Predict(taxiTripSample);
WriteLine($"**********************************************************************");
WriteLine($"Predicted fare: {prediction.FareAmount:0.####}, actual fare: 21");
WriteLine($"**********************************************************************\n");
}
/// <summary>
/// Extending this with something actually useful
/// </summary>
/// <param name="mLContext">Pass context in because I'm not a fan of state</param>
/// <param name="trip">Trip details for which to predict price</param>
/// <returns></returns>
private static TaxiTripFarePrediction Predict(MLContext mlContext, TaxiTrip trip) // This should not really be in Program.cs, make public and move this and everything else out into an Engine class?
{
// TODO: Cache model? Going from disk every time seems extremely inefficient.
ITransformer loadedModel;
using (var stream = new FileStream(_modelPath, FileMode.Open, FileAccess.Read, FileShare.Read))
{
loadedModel = mlContext.Model.Load(stream);
}
var predictionFunction = loadedModel.CreatePredictionEngine <TaxiTrip, TaxiTripFarePrediction>(mlContext);
var prediction = predictionFunction.Predict(trip);
return(prediction);
}
private static void ReadAndTestRandomRowFromFile(MLContext mlContext)
{
if (File.Exists(_testDataPath))
{
using (StreamReader reader = new StreamReader(_testDataPath))
{
string strFile = reader.ReadToEnd();
Random rand = new Random();
List <string> strArray = new List <string>(strFile.Split("\n"));
List <string> randomTaxiTripSample = new List <string>(strArray[rand.Next(1, strArray.Count)].Split(","));
ITransformer loadedModel;
using (var stream = new FileStream(_modelPath, FileMode.Open, FileAccess.Read, FileShare.Read))
{
loadedModel = mlContext.Model.Load(stream);
}
var predictionFunction = loadedModel.CreatePredictionEngine <TaxiTrip, TaxiTripFarePrediction>(mlContext);
var taxiTripSample = new TaxiTrip()
{
VendorId = randomTaxiTripSample[0],
RateCode = randomTaxiTripSample[1],
PassengerCount = float.Parse(randomTaxiTripSample[2]),
TripTime = float.Parse(randomTaxiTripSample[3]),
TripDistance = float.Parse(randomTaxiTripSample[4]),
PaymentType = randomTaxiTripSample[5],
FareAmount = float.Parse(randomTaxiTripSample[6])
};
var prediction = predictionFunction.Predict(taxiTripSample);
WriteLine($"**********************************************************************");
WriteLine($"VendorId: {taxiTripSample.VendorId}");
WriteLine($"RateCode: {taxiTripSample.RateCode}");
WriteLine($"PassengerCount: {taxiTripSample.PassengerCount}");
WriteLine($"TripTime: {taxiTripSample.TripTime}");
WriteLine($"TripDistance: {taxiTripSample.TripDistance}");
WriteLine($"PaymentType: {taxiTripSample.PaymentType}");
WriteLine($"Predicted fare: {prediction.FareAmount:0.####}, actual fare: {taxiTripSample.FareAmount}");
WriteLine($"**********************************************************************\n");
}
}
else
{
WriteLine($"{_testDataPath} not found.");
}
}
static void Main(string[] args)
{
var mlContext = new MLContext(seed: 0);
_textLoader = mlContext.Data.CreateTextLoader(new TextLoader.Arguments()
{
Separators = new[] { ',' },
HasHeader = true,
Column = new[]
{
new TextLoader.Column("VendorId", DataKind.Text, 0),
new TextLoader.Column("RateCode", DataKind.Text, 1), // Peak, offpeak?
new TextLoader.Column("PassengerCount", DataKind.R4, 2),
new TextLoader.Column("TripTime", DataKind.R4, 3), // Duration in seconds
new TextLoader.Column("TripDistance", DataKind.R4, 4), // Decimal miles
new TextLoader.Column("PaymentType", DataKind.Text, 5),
new TextLoader.Column("FareAmount", DataKind.R4, 6) // Decimal dollars
}
});
// Don't regenerate and reevaluate the model every time, takes too long
// TODO: Make it regenerate the model if the training data has changed? (Would need to persist a hash of the training data between runs)
if (!File.Exists(_modelPath))
{
var model = Train(mlContext, _trainDataPath); // Generate model from training dataset - TODO: No reason to do this every time!
Evaluate(mlContext, model); // Evaluate model performance against the test data (TODO: Parameterise to allow evaluation against different datasets without modifying source)
}
//TestSinglePrediction(mlContext);
// Let's actually use this as if it was a real service
// TODO: Make it possible to specify this trip from command line
var culture = new CultureInfo("en-US");
var trip = new TaxiTrip()
{
VendorId = "VTS",
RateCode = "1",
PassengerCount = int.Parse(args[0], culture), //1,
TripTime = int.Parse(args[1], culture), //1540,
TripDistance = float.Parse(args[2], culture), // 5.70f,
PaymentType = "CRD",
FareAmount = 0 // To predict.
};
var result = Predict(mlContext, trip);
Console.WriteLine($"Predicted fare: ${result.FareAmount:#.##}");
}
private static void TestSinglePrediction(MLContext mLContext, ITransformer model) {
var predictionFunction = mLContext.Model.CreatePredictionEngine<TaxiTrip, TaxiTripFarePrediction>(model);
var taxiTripSample = new TaxiTrip() {
VendorId = "VTS",
RateCode = "1",
PassengerCount = 1,
TripTime = 1140,
TripDistance = 3.75f,
PaymentType = "CRD",
FareAmount = 0 // To predict. Actual/Observed = 15.5
};
var prediction = predictionFunction.Predict(taxiTripSample);
Console.WriteLine($"**********************************************************************");
Console.WriteLine($"Predicted fare: {prediction.FareAmount:0.####}, actual fare: 15.5");
Console.WriteLine($"**********************************************************************");
}
private static void TestSinglePrediction(TaxiFarePredictionEngine.PredictionEngine predictionEngine)
{
TaxiTrip taxiTripSample = new TaxiTrip()
{
VendorId = "VTS",
RateCode = "1",
PassengerCount = 1,
TripTime = 1140,
TripDistance = 3.75f,
PaymentType = "CRD",
FareAmount = 0
};
TaxiTripFarePrediction prediction = predictionEngine.GetFarePrediction(taxiTripSample);
Console.WriteLine($"**********************************************************************");
Console.WriteLine($"Predicted fare: {prediction.FareAmount:0.####}, actual fare: 15.5");
Console.WriteLine($"**********************************************************************");
}
private static void TestSinglePrediction(MLContext mlContext)
{
//load the model
// <Snippet21>
ITransformer loadedModel;
using (var stream = new FileStream(_modelPath, FileMode.Open, FileAccess.Read, FileShare.Read))
{
loadedModel = mlContext.Model.Load(stream);
}
// </Snippet21>
//Prediction test
// Create prediction function and make prediction.
// <Snippet22>
var predictionFunction = loadedModel.MakePredictionFunction <TaxiTrip, TaxiTripFarePrediction>(mlContext);
// </Snippet22>
//Sample:
//vendor_id,rate_code,passenger_count,trip_time_in_secs,trip_distance,payment_type,fare_amount
//VTS,1,1,1140,3.75,CRD,15.5
// <Snippet23>
var taxiTripSample = new TaxiTrip()
{
VendorId = "VTS",
RateCode = "1",
PassengerCount = 1,
TripTime = 1140,
TripDistance = 3.75f,
PaymentType = "CRD",
FareAmount = 0 // To predict. Actual/Observed = 15.5
};
// </Snippet23>
// <Snippet24>
var prediction = predictionFunction.Predict(taxiTripSample);
// </Snippet24>
// <Snippet25>
Console.WriteLine($"**********************************************************************");
Console.WriteLine($"Predicted fare: {prediction.FareAmount:0.####}, actual fare: 15.5");
Console.WriteLine($"**********************************************************************");
// </Snippet25>
}
/// <summary>
/// The main application entry point.
/// </summary>
/// <param name="args">The command line arguments.</param>
static void Main(string[] args)
{
// create the machine learning context
var mlContext = new MLContext();
// set up the text loader
var textLoader = mlContext.Data.CreateTextLoader(
new TextLoader.Options()
{
Separators = new[] { ',' },
HasHeader = true,
Columns = new[]
{
new TextLoader.Column("VendorId", DataKind.String, 0),
new TextLoader.Column("RateCode", DataKind.String, 5),
new TextLoader.Column("PassengerCount", DataKind.Single, 3),
new TextLoader.Column("TripDistance", DataKind.Single, 4),
new TextLoader.Column("PaymentType", DataKind.String, 9),
new TextLoader.Column("FareAmount", DataKind.Single, 10)
}
}
);
// load the data
Console.Write("Loading training data....");
var dataView = textLoader.Load(dataPath);
Console.WriteLine("done");
// split into a training and test partition
var partitions = mlContext.Data.TrainTestSplit(dataView, testFraction: 0.2);
// set up a learning pipeline
var pipeline = mlContext.Transforms.CopyColumns(
inputColumnName: "FareAmount",
outputColumnName: "Label")
// one-hot encode all text features
.Append(mlContext.Transforms.Categorical.OneHotEncoding("VendorId"))
.Append(mlContext.Transforms.Categorical.OneHotEncoding("RateCode"))
.Append(mlContext.Transforms.Categorical.OneHotEncoding("PaymentType"))
// combine all input features into a single column
.Append(mlContext.Transforms.Concatenate(
"Features",
"VendorId",
"RateCode",
"PassengerCount",
"TripDistance",
"PaymentType"))
// cache the data to speed up training
.AppendCacheCheckpoint(mlContext)
// use the fast tree learner
.Append(mlContext.Regression.Trainers.FastTree());
// train the model
Console.Write("Training the model....");
var model = pipeline.Fit(partitions.TrainSet);
Console.WriteLine("done");
// get a set of predictions
Console.Write("Evaluating the model....");
var predictions = model.Transform(partitions.TestSet);
// get regression metrics to score the model
var metrics = mlContext.Regression.Evaluate(predictions, "Label", "Score");
Console.WriteLine("done");
// show the metrics
Console.WriteLine();
Console.WriteLine($"Model metrics:");
Console.WriteLine($" RMSE:{metrics.RootMeanSquaredError:#.##}");
Console.WriteLine($" MSE: {metrics.MeanSquaredError:#.##}");
Console.WriteLine($" MAE: {metrics.MeanAbsoluteError:#.##}");
Console.WriteLine();
// create a prediction engine for one single prediction
var predictionFunction = mlContext.Model.CreatePredictionEngine <TaxiTrip, TaxiTripFarePrediction>(model);
// prep a single taxi trip
var taxiTripSample = new TaxiTrip()
{
VendorId = "2",
RateCode = "1",
PassengerCount = 1,
TripDistance = 3.75f,
PaymentType = "1",
FareAmount = 0 // the model will predict the actual fare for this trip
};
// make the prediction
var prediction = predictionFunction.Predict(taxiTripSample);
// sho the prediction
Console.WriteLine($"Single prediction:");
Console.WriteLine($" Predicted fare: {prediction.FareAmount:0.####}");
}
//static TextLoader _textLoader; not used
static void Main(string[] args)
{
// STEP 1: Common data loading configuration
MLContext mlContext = new MLContext(seed: 0);
IDataView baseTrainingDataView = mlContext.Data.LoadFromTextFile <TaxiTrip>(_trainDataPath, hasHeader: true, separatorChar: ','); // ReadFromTextFile is deprecated
IDataView testDataView = mlContext.Data.LoadFromTextFile <TaxiTrip>(_testDataPath, hasHeader: true, separatorChar: ',');
//Sample code of removing extreme data like "outliers" for FareAmounts higher than $150 and lower than $1 which can be error-data
var cnt = baseTrainingDataView.GetColumn <float>(mlContext, nameof(TaxiTrip.FareAmount)).Count();
IDataView trainingDataView = mlContext.Data.FilterRowsByColumn(baseTrainingDataView, nameof(TaxiTrip.FareAmount), lowerBound: 1, upperBound: 150);
// STEP 2: Common data process configuration with pipeline data transformations
var dataProcessPipeline = mlContext.Transforms.CopyColumns(outputColumnName: DefaultColumnNames.Label, inputColumnName: nameof(TaxiTrip.FareAmount))
.Append(mlContext.Transforms.Categorical.OneHotEncoding(outputColumnName: "VendorIdEncoded", inputColumnName: nameof(TaxiTrip.VendorId)))
.Append(mlContext.Transforms.Categorical.OneHotEncoding(outputColumnName: "RateCodeEncoded", inputColumnName: nameof(TaxiTrip.RateCode)))
.Append(mlContext.Transforms.Categorical.OneHotEncoding(outputColumnName: "PaymentTypeEncoded", inputColumnName: nameof(TaxiTrip.PaymentType)))
.Append(mlContext.Transforms.Normalize(outputColumnName: nameof(TaxiTrip.PassengerCount), mode: NormalizerMode.MeanVariance))
.Append(mlContext.Transforms.Normalize(outputColumnName: nameof(TaxiTrip.TripTime), mode: NormalizerMode.MeanVariance))
.Append(mlContext.Transforms.Normalize(outputColumnName: nameof(TaxiTrip.TripDistance), mode: NormalizerMode.MeanVariance))
.Append(mlContext.Transforms.Concatenate(DefaultColumnNames.Features, "VendorIdEncoded", "RateCodeEncoded", "PaymentTypeEncoded", nameof(TaxiTrip.PassengerCount)
, nameof(TaxiTrip.TripTime), nameof(TaxiTrip.TripDistance)));
// STEP 3: Set the training algorithm, then create and config the modelBuilder - Selected Trainer (SDCA Regression algorithm)
var trainer = mlContext.Regression.Trainers.StochasticDualCoordinateAscent(labelColumnName: DefaultColumnNames.Label, featureColumnName: DefaultColumnNames.Features);
var trainingPipeline = dataProcessPipeline.Append(trainer);
//STEP 4: Train the model fitting to the DataSet
//The pipeline is trained on the dataset that has been loaded and transformed.
ITransformer trainedModel = trainingPipeline.Fit(trainingDataView); // changed var trainedModel to ITransformer trainedModel
// STEP 5: Evaluate the model and show accuracy stats
// create a copy of the testDataView in IDataView format since IDataView is immutable?
IDataView predictions = trainedModel.Transform(testDataView);
// assign label column (original regression value) and score column (predictive regression values) to columns in data i.e. predictions
// apply metrics e.g. root mean squared error on data
var metrics = mlContext.Regression.Evaluate(predictions, label: DefaultColumnNames.Label, score: DefaultColumnNames.Score);
// print the results of the metrics
Common.ConsoleHelper.PrintRegressionMetrics(trainer.ToString(), metrics);
// STEP 6: Save/persist the trained model to a .ZIP file
using (var fs = File.Create(_modelPath))
trainedModel.SaveTo(mlContext, fs);
Console.WriteLine("The model is saved to {0}", _modelPath);
// STEP 7: Prediction
//Sample:
//vendor_id,rate_code,passenger_count,trip_time_in_secs,trip_distance,payment_type,fare_amount
//VTS,1,1,1140,3.75,CRD,15.5
var taxiTripSample = new TaxiTrip()
{
VendorId = "VTS",
RateCode = "1",
PassengerCount = 1,
TripTime = 1140,
TripDistance = 3.75f,
PaymentType = "CRD",
FareAmount = 0 // To predict. Actual/Observed = 15.5
};
using (var stream = new FileStream(_modelPath, FileMode.Open, FileAccess.Read, FileShare.Read))
{
trainedModel = mlContext.Model.Load(stream);
}
// Create prediction engine related to the loaded trained model
var predEngine = trainedModel.CreatePredictionEngine <TaxiTrip, TaxiTripFarePrediction>(mlContext);
//Score
var resultprediction = predEngine.Predict(taxiTripSample);
Console.WriteLine($"**********************************************************************");
Console.WriteLine($"Predicted fare: {resultprediction.FareAmount:0.####}, actual fare: 15.5");
Console.WriteLine($"**********************************************************************");
}
static async Task Main(string[] args)
{
var regressorTypes = new [] {
typeof(FastForestRegressor),
typeof(FastTreeTweedieRegressor),
typeof(GeneralizedAdditiveModelRegressor),
// typeof(OnlineGradientDescentRegressor),
// typeof(OrdinaryLeastSquaresRegressor),
typeof(PoissonRegressor),
typeof(StochasticDualCoordinateAscentRegressor),
};
var taxiTripTests = File.ReadAllLines(_testdatapath)
.Skip(1)
.Select(x => {
var fields = x.Split(',');
var taxiTrip = new TaxiTrip {
VendorId = fields[0],
RateCode = fields[1],
PassengerCount = float.Parse(fields[2]),
TripTime = float.Parse(fields[3]),
TripDistance = float.Parse(fields[4]),
PaymentType = fields[5],
FareAmount = float.Parse(fields[6])
};
return(taxiTrip, taxiTripFarePrediction: (TaxiTripFarePrediction)null);
})
//.Take(500)
.ToArray();
Func <PredictionModel <TaxiTrip, TaxiTripFarePrediction> > trainFunc = Train <FastForestRegressor>;
foreach (var regressorType in regressorTypes)
{
var methodInfo = trainFunc.Method.GetGenericMethodDefinition().MakeGenericMethod(regressorType);
Console.WriteLine();
Console.WriteLine("--- " + regressorType.Name + " ---");
var model = (PredictionModel <TaxiTrip, TaxiTripFarePrediction>)methodInfo.Invoke(null, null);
Evaluate(model);
var prediction = model.Predict(TestTrips.Trip1);
Console.WriteLine("Predicted fare: {0}, actual fare: 29.5", prediction.FareAmount);
Console.WriteLine();
Console.Write("Begin test...");
var what = model.Predict(taxiTripTests.Select(x => x.taxiTrip)).ToList();
Parallel.ForEach(what, (x, state, i) => taxiTripTests[(int)i].taxiTripFarePrediction = x);
Console.WriteLine("Done.");
var diffs = taxiTripTests.Select(x => Math.Abs(x.taxiTrip.FareAmount - x.taxiTripFarePrediction.FareAmount))
.AsParallel()
.OrderBy(x => x)
.ToList();
var okay = diffs.Take(diffs.Count - 2)
.Where((x, i) => x > 0.75 & x < 450)
.Where((x, i) => i % 1000 == 0)
.Select((x, i) => i + "," + x);
File.WriteAllLines($@".\Data\{regressorType.Name}_ModelErrors.csv", okay);
Console.WriteLine(taxiTripTests.Min(x => Math.Abs(x.taxiTrip.FareAmount - x.taxiTripFarePrediction.FareAmount)));
Console.WriteLine(taxiTripTests.Max(x => Math.Abs(x.taxiTrip.FareAmount - x.taxiTripFarePrediction.FareAmount)));
Console.WriteLine(taxiTripTests.Average(x => Math.Abs(x.taxiTrip.FareAmount - x.taxiTripFarePrediction.FareAmount)));
}
}
private static void TestRandomPrediction(MLContext mlContext)
{
Random rand = new Random();
//load the model
ITransformer loadedModel;
using (var stream = new FileStream(_modelPath, FileMode.Open, FileAccess.Read, FileShare.Read))
{
loadedModel = mlContext.Model.Load(stream);
}
//Prediction test
// Create prediction function and make prediction.
var predictionFunction = loadedModel.CreatePredictionEngine <TaxiTrip, TaxiTripFarePrediction>(mlContext);
//
//Sample:
//vendor_id,rate_code,passenger_count,trip_time_in_secs,trip_distance,payment_type,fare_amount
//VTS,1,1,1140,3.75,CRD,15.5
var taxiTripSample = new TaxiTrip()
{
RateCode = "1", // 1 - metered, 2 - fixed
PassengerCount = rand.Next(1, 6), // 1 - 6 passengers
TripTime = rand.Next(6, 1100) * 10, // nearest 10 secs.
TripDistance = ((float)rand.NextDouble() * 100.0f) + 0.1f,
FareAmount = 0 // To predict. Actual/Observed = 15.5
};
//
// Pick a vendor ID
if (rand.Next() % 2 == 0)
{
taxiTripSample.VendorId = "VTS";
}
else
{
taxiTripSample.VendorId = "CMT";
}
if (rand.Next() % 2 == 0)
{
taxiTripSample.PaymentType = "CSH";
}
else
{
taxiTripSample.PaymentType = "CRD";
}
var prediction = predictionFunction.Predict(taxiTripSample);
WriteLine($"**********************************************************************");
WriteLine($"VendorId: {taxiTripSample.VendorId}");
WriteLine($"RateCode: {taxiTripSample.RateCode}");
WriteLine($"PassengerCount: {taxiTripSample.PassengerCount}");
WriteLine($"TripTime: {taxiTripSample.TripTime}");
WriteLine($"TripDistance: {taxiTripSample.TripDistance}");
WriteLine($"PaymentType: {taxiTripSample.PaymentType}");
WriteLine($"Predicted fare: {prediction.FareAmount:0.####}, actual fare: 15.5");
WriteLine($"**********************************************************************\n");
}
static void Main(string[] args)
{
Helper.PrintLine("创建 MLContext...");
MLContext mlContext = new MLContext(seed: 0);
ITransformer model;
if (File.Exists(ModelPath))
{
Helper.PrintLine("加载神经网络模型...");
model = mlContext.Model.Load(ModelPath, out DataViewSchema inputScema);
}
else
{
// 训练数据集合
IDataView trainingDataView = mlContext.Data.LoadFromTextFile <TaxiTrip>(TrainingDataPath, hasHeader: true, separatorChar: ',');
// 创建神经网络管道
Helper.PrintLine("创建神经网络管道...");
var pipeline = mlContext.Transforms.CopyColumns(outputColumnName: "Label", inputColumnName: "FareAmount")
.Append(mlContext.Transforms.Categorical.OneHotEncoding(outputColumnName: "VendorIdEncoded", inputColumnName: "VendorId"))
.Append(mlContext.Transforms.Categorical.OneHotEncoding(outputColumnName: "RateCodeEncoded", inputColumnName: "RateCode"))
.Append(mlContext.Transforms.Categorical.OneHotEncoding(outputColumnName: "PaymentTypeEncoded", inputColumnName: "PaymentType"))
.Append(mlContext.Transforms.Concatenate("Features", "VendorIdEncoded", "RateCodeEncoded", "PassengerCount", "TripTime", "TripDistance", "PaymentTypeEncoded"))
.Append(mlContext.Regression.Trainers.FastTree());
// 开始训练神经网络
Helper.PrintSplit();
Helper.PrintLine("开始训练神经网络...");
model = pipeline.Fit(trainingDataView);
Helper.PrintLine("训练神经网络完成");
Helper.PrintSplit();
Helper.PrintLine($"导出神经网络模型...");
mlContext.Model.Save(model, trainingDataView.Schema, ModelPath);
}
// 测试
Helper.PrintLine("评估神经网络：");
var testDataView = mlContext.Data.LoadFromTextFile <TaxiTrip>(TestDataPath, hasHeader: true, separatorChar: ',');
var testMetrics = mlContext.Regression.Evaluate(model.Transform(testDataView), "Label", "Score");
Helper.PrintLine($"\t=>R^2: {testMetrics.RSquared:0.###}");
Helper.PrintLine($"\t=>RMS error: {testMetrics.RootMeanSquaredError:0.###}");
// 预测
Helper.PrintLine("预测：");
var predictionEngine = mlContext.Model.CreatePredictionEngine <TaxiTrip, TaxiTripFarePrediction>(model);
var taxiTripSample = new TaxiTrip()
{
VendorId = "VTS",
RateCode = "1",
PassengerCount = 1,
TripTime = 1140,
TripDistance = 3.75f,
PaymentType = "CRD",
FareAmount = 0 // To predict. Actual/Observed = 15.5
};
var prediction = predictionEngine.Predict(taxiTripSample);
Helper.PrintLine($"预测价格: {prediction.FareAmount:0.####}, actual fare: 15.5");
Helper.Exit(0);
}
private static void Predict(PredictionModel <TaxiTrip, TaxiTripFarePrediction> model, TaxiTrip trip1)
{
var prediction = model.Predict(trip1);
Console.WriteLine("Predicted fare: {0}, actual fare: 29.5", prediction.FareAmount);
}