public ForecastingModel calculateTrend()
{
this.Trend = new List <double>();
List <Tuple <double, double> > LeastSquaresSerie = new List <Tuple <double, double> >();
for (int i = 0; i < this.SeasonallyAdjustedData.Count; i++)
{
LeastSquaresSerie.Add(new Tuple <double, double>(i + 1, this.SeasonallyAdjustedData.ElementAt(i)));
}
//y = mx + b
LeastSquaresParameters = StatisticsModel.LeastSquares(LeastSquaresSerie);
for (int i = 0; i < this.Serie.Count; i++)
{
this.Trend.Add((i + 1) * LeastSquaresParameters.M + LeastSquaresParameters.B);
}
return(this);
}
public ForecastingModel findSeasonality(Boolean isPearson)
{
double[] arrSource;
double[] arrSource2;
if (isPearson)
{
arrSource = this.Serie.ToArray();
}
else
{
arrSource2 = this.Serie.ToArray();
arrSource = new double[arrSource2.Length - 1];
for (int i = 0; i < arrSource2.Length; i++)
{
arrSource2[i] = Math.Log(arrSource2[i]);
if (i >= 1)
{
arrSource[i - 1] = arrSource2[i] - arrSource2[i - 1];
}
}
}
double pearson;
double max = -1;
for (int shifts = 1; shifts <= 19; shifts++)
{
double[] arr = new double[arrSource.Length];
Array.Copy(arrSource, 0, arr, shifts, arr.Length - shifts);
if (isPearson)
{
pearson = StatisticsModel.Pearson(arrSource, arr);
}
else
{
pearson = RankCorrelation.ComputeRankCorrelation(arrSource, arr);
}
if (pearson > max)
{
seasonalityRate = shifts;
max = pearson;
}
}
Console.WriteLine("pearson: " + max + " seasonalityRate: " + seasonalityRate);
return(this);
}
public ForecastingModel findSeasonality()
{
double[] arrSource = this.Serie.ToArray();
double pearson;
double max = -1;
for (int shifts = 1; shifts <= 20; shifts++)
{
double[] arr = new double[arrSource.Length];
Array.Copy(arrSource, 0, arr, shifts, arr.Length - shifts);
pearson = StatisticsModel.Pearson(arrSource, arr);
if (pearson > max)
{
seasonalityRate = shifts;
max = pearson;
}
}
Console.WriteLine("pearson: " + max + " seasonalityRate: " + seasonalityRate);
return(this);
}
