primary

Type Members

1. case class ARMA(phi: Array[Double] = Array(), theta: Array[Double] = Array(), std: Double = 1, c: Double = 0, seed: Long = Random.nextLong) extends Product with Serializable

   

   An auto-regressive, moving-average model.

   An auto-regressive, moving-average model.

   It provides a weakly stationary stochastic process as a sum of two polynomials - one for the auto-regression - one for the moving average

   The time series is built as follows:

   X_t = c + epsilon_t + ∑(phi_i * X_(t-1)) + ∑(theta_i * epsilon_(t-1))

   Where epsilon is generated from a white noise of a specified standard deviation.

   phi

   the parameters used to characterize the autoregression part of the model

   theta

   the parameters used to characterize the moving average part of the model

   std

   the standard deviation used to characterize the generated white noise

   c

   a constant

   seed

   the seed used to generate the white noise. For a given seed, the process is deterministic

2. case class ConstantTimeSeries(value: Double) extends IndependantTimeSeries[Double] with Product with Serializable

   

   A time series made of a constant value.

3. case class DailyTimeSeries(controlPoints: Map[LocalTime, Double]) extends IndependantTimeSeries[Double] with Product with Serializable

   

   Represents cyclic variation of a time series on a daily basis.

   Represents cyclic variation of a time series on a daily basis.

   controlPoints

   The value a time series must pass by at a given time.

4. case class MonthlyTimeSeries(controlPoints: Map[Int, Double]) extends IndependantTimeSeries[Double] with Product with Serializable

   

   Represents cyclic variation of a time series on a monthly basis.

   Represents cyclic variation of a time series on a monthly basis.

   controlPoints

   The value a time series must pass by at a given time.

5. case class RandomWalkTimeSeries(arma: ARMA, timeStep: Duration) extends TimeSeries[Double] with Product with Serializable

   

   A time series based on an ARMA model.

   A time series based on an ARMA model.

   The ARMA model provides a series of discrete values. In order to bind them to a particular date time, a linear regression is used.

   arma

   the ARMA model used to generate a time series based on a random walk.

   timeStep

   the duration between two consecutive steps.

6. class UndefinedTimeSeries[T] extends TimeSeries[T]

   

   A time series that only have undefined values.

7. case class WeeklyTimeSeries(controlPoints: Map[Int, Double]) extends IndependantTimeSeries[Double] with Product with Serializable

   

   Represents cyclic variation of a time series on a weekly basis.

   Represents cyclic variation of a time series on a weekly basis.

   controlPoints

   The value a time series must pass by at a given time.

8. case class YearlyTimeSeries(controlPoints: Map[Int, Double]) extends IndependantTimeSeries[Double] with Product with Serializable

   

   Represents cyclic variation of a time series on a yearly basis.

   Represents cyclic variation of a time series on a yearly basis. Since years are on an open scale, a smooth interpolation of values is not waranted.

   controlPoints

   The value a time series must pass by at a given time.