public App()
{
int i = 3;
var RawData = DataPreparer.Splicer(i);
Constants.Item item = Constants.Item.GetItem(RawData, 4000);
Constants.RawData raw = Constants.Item.GetRawData();
//series.InitializeComponent();
MainWindow window = new MainWindow();
window.InitializeComponent();
}
private static double StrdErrorSolver(List <double> Quant, List <double> Prediction)
{
double Err = 0;
List <double> deStrd = DataPreparer.DeStrd(Prediction);
for (int i = 0; i < Quant.Count(); i++)
{
var PredErr = Quant.ElementAt(i) - deStrd.ElementAt(i);
Err += Math.Pow(PredErr, 2);
}
return(Err);
}
public static void Main(string[] args)
{
int i = 3;
Console.WriteLine("Operation on dataset: Item ({0})...", Constants.id[i]);
//get initial data from files
var RawData = DataPreparer.Splicer(i);
Console.WriteLine("Getting data from files");
//subset length, total set size is 4*n (buy/sell quant, buy/sell price)
//trim dataset to n before singleton implemented!
int n = 4000;
Console.WriteLine("trimmng datset...");
DataPreparer.Trimmer(RawData, n);
//get item specific data for raw dataset as a singleton class
Console.WriteLine("Getting item...");
Constants.Item item = Constants.Item.GetItem(RawData, n);
Constants.RawData Data = Constants.Item.GetRawData();
//standardised dataset as well as weighted and standardised weighted data
Console.WriteLine("Standardising dataset");
Constants.StrdData strdData = DataPreparer.Z_Score(Data);
//LSR fitting solutions to each set (standardised & non-standardised)
Console.WriteLine("Fitting function parameters");
double[,] FitEq = FittingFunctions.FitLin(Data, n);
double[,] StrdFitEq = FittingFunctions.StrdFitLin(strdData, n);
Constants.DataInfo Info = DataPreparer.DataInfo();
Console.WriteLine("Getting model information");
double[] Corrs = ModelInfo.CorrCoeff(Data, Info);
//Get prediction lists and convert to List<int> for StrdDev function
var _DemPreds = FittingFunctions.Predictor(Data.BuyPrice, FitEq[0, 0], FitEq[1, 0]);
List <int> DemPreds = _DemPreds.Select(x => (int)x).ToList();
var _SupPreds = FittingFunctions.Predictor(Data.SellPrice, FitEq[0, 1], FitEq[1, 1]);
List <int> SupPreds = _SupPreds.Select(x => (int)x).ToList();
//Standard deviations of data, standard deviations of predicted data & correlations
double[] FitData =
{
ModelInfo.StrdDev(Data.SellQuantity, Info.SellQuantInfo, n),
ModelInfo.StrdDev(Data.BuyQuantity, Info.BuyQuantInfo, n),
ModelInfo.StrdDev(DemPreds, Info.BuyQuantInfo, n),
ModelInfo.StrdDev(SupPreds, Info.BuyQuantInfo, n),
ModelInfo.RMSE(FitEq[2, 0], n),ModelInfo.RMSE(FitEq[2,1], n),
Corrs[0], Corrs[1]
};
Console.WriteLine("Printing results to files...");
DataWriter.FileMaker(Info, i, FitEq, StrdFitEq, FitData);
Console.Write("Check the desktop!");
Console.ReadLine();
}
public TimeSeries()
{
InitializeComponent();
Constants.RawData data = Constants.Item.GetRawData();
int n = Constants.Item.GetCount();
var Ewma = DataPreparer.Weighter(data, 0.5);
YFormatter = value => value.ToString("C");
List <double> BuyQ = data.BuyQuantity.Select <int, double>(i => i).ToList();
List <double> BuyP = data.BuyPrice.Select <int, double>(i => i).ToList();
List <double> SellQ = data.SellQuantity.Select <int, double>(i => i).ToList();
List <double> SellP = data.SellPrice.Select <int, double>(i => i).ToList();
for (int i = 0; i < BuyQ.Count(); i++)
{
EWMAPrice.Add(new ObservablePoint(i, Ewma._strdBuyPrice.ElementAt(i)));
Price.Add(new ObservablePoint(i, BuyP.ElementAt(i)));
}
DataContext = this;
}
/// <summary>
/// Fits a line in a univariate analysis to data
/// </summary>
/// <param name="data">Input</param>
/// <param name="n">Range of Input</param>
/// <returns>Fit Param's, first column is gradients</returns>
public static double[,] FitLin(Constants.RawData data, int n)
{
//S(p)=BuyQuantity = b + m*SellPrice
//D(p)=SellQuantity = b - m*BuyPrice
double[,] Params = new double[3, 2];
var BQuant = data.BuyQuantity;
var BPrice = data.BuyPrice;
var SQuant = data.SellQuantity;
var SPrice = data.SellPrice;
//array location 0, 0 is mean value
var InfoObject = DataPreparer.DataInfo();
var BPMean = (double)InfoObject.BuyPriceInfo[0, 0];
var BQMean = (double)InfoObject.BuyQuantInfo[0, 0];
var SPMean = (double)InfoObject.SellPriceInfo[0, 0];
var SQMean = (double)InfoObject.SellQuantInfo[0, 0];
//return B1/Grad for demand and supply
var GradDem = GradSolver(BPrice, SQuant, n, BPMean, SQMean);
var GradSup = GradSolver(SPrice, BQuant, n, SPMean, BQMean);
//return B0/intercept for demand and supply
var DemInt = IntSolver(GradDem, SQMean, BPMean);
var SupInt = IntSolver(GradSup, BQMean, SPMean);
//create predictions of data and add to predictions lists
var PredDem = Predictor(BPrice, GradDem, DemInt);
var PredSup = Predictor(SPrice, GradSup, SupInt);
//return error on predicitons
var DemErr = ErrorSolver(SQuant, PredDem);
var SupErr = ErrorSolver(BQuant, PredSup);
//add to Params and return
Params[0, 0] = Math.Round(GradDem);
Params[0, 1] = Math.Round(GradSup);
Params[1, 0] = Math.Round(DemInt);
Params[1, 1] = Math.Round(SupInt);
Params[2, 0] = Math.Round(DemErr);
Params[2, 1] = Math.Round(SupErr);
Console.WriteLine(Params[2, 1]);
return(Params);
}
/// <summary>
/// Fit to a standardised data set then return de-standardized values
/// </summary>
/// <param name="strdData"></param>
/// <param name="n">Dataset length</param>
/// <returns></returns>
public static double[,] StrdFitLin(Constants.StrdData strdData, int n)
{
//S(p)=SellQuantity = b + m*BuyPrice
//D(p)=BuyQuantity = b + m*SellPrice
double[,] Params = new double[3, 2];
var BQuant = strdData._strdBuyQuant;
var BPrice = strdData._strdBuyPrice;
var SQuant = strdData._strdSellQuant;
var SPrice = strdData._strdSellPrice;
var InfoObject = DataPreparer.StrdDataInfo(strdData);
//array location 0, 0 is mean value
var BPMean = (double)InfoObject.BuyPriceInfo[0, 0];
var BQMean = (double)InfoObject.BuyQuantInfo[0, 0];
var SPMean = (double)InfoObject.SellPriceInfo[0, 0];
var SQMean = (double)InfoObject.SellQuantInfo[0, 0];
//return B1/Grad for demand and supply
var GradDem = StrdGradSolver(SPrice, BQuant, n, SPMean, BQMean);
var GradSup = StrdGradSolver(BPrice, SQuant, n, BPMean, SQMean);
//return B0/intercept for demand and supply
var DemInt = IntSolver(GradDem, BQMean, SPMean);
var SupInt = IntSolver(GradSup, SQMean, BPMean);
//create predictions of data and add to predictions lists
var PredDem = StrdPredictor(SPrice, GradSup, SupInt);
var PredSup = StrdPredictor(BPrice, GradDem, DemInt);
//return error on predicitons
var DemErr = StrdErrorSolver(BQuant, PredDem);
var SupErr = StrdErrorSolver(SQuant, PredSup);
//add to Params and return
Params[0, 0] = GradDem;
Params[0, 1] = GradSup;
Params[1, 0] = DemInt;
Params[1, 1] = SupInt;
Params[2, 0] = DemErr;
Params[2, 1] = SupErr;
return(Params);
}
