static void Main(string[] args)
{
var normalize = new ValueNormalizer();
var dataTransformer = new DataTransformer();
var dataSupplier = new DataSourceSupplier();
var data = dataSupplier.GetTrainingSet().ToList();
var trainingSet = GetNormalizedData();
var functions = new FunctionsCalculator(trainingSet);
var newHouse = new double[] { 1, 1203, 3 };
double predictedPrice = 0.0;
var actualPrice = 239500;
var theta = new double[] { 89597.9027899082, 504777.895309395, -34952.0620218024 };
for (int i = 1; i <= 2; i++)
{
var featureVector = dataTransformer.GetNthFeatureVector(data, i).ToArray();
var min = featureVector.Min();
var max = featureVector.Max();
newHouse[i] = normalize.FeatureScaling(min, max, newHouse[i]);
}
predictedPrice = functions.Hypothesis(newHouse, theta);
Console.WriteLine($"Actual price: {actualPrice}");
Console.WriteLine($"Predicted price: {predictedPrice}");
Console.Read();
}
public IEnumerable <T> Parse(string source)
{
using (var s = SourceProvider.GetStreamReader(source))
{
return(StreamParser.Parse(s)
.Select(v => ValueNormalizer.Normalize(v)));
}
}
private IEnumerable <TrainingElement> GetNormalizedData()
{
var dataSupplier = new DataSourceSupplier();
var dataTransformer = new DataTransformer();
var data = dataSupplier.GetTrainingSet().ToList();
var valueNormalizer = new ValueNormalizer();
var featuresCount = data.First().Features.Count() - 1;
for (int i = 1; i <= featuresCount; i++)
{
var featureVector = dataTransformer.GetNthFeatureVector(data, i).ToArray();
var min = featureVector.Min();
var max = featureVector.Max();
for (int j = 0; j < featureVector.Length; j++)
{
data[j].Features[i] = valueNormalizer.FeatureScaling(min, max, featureVector[j]);
}
}
return(data);
}
public void NormalizeTest()
{
Assert.AreEqual("", ValueNormalizer.Normalize(null));
Assert.AreEqual("", ValueNormalizer.Normalize("None"));
Assert.AreEqual("dummy", ValueNormalizer.Normalize("dummy"));
}
