ARX_Quad_MV
The ARX_Quad_MV object supports quadratic regression for Time Series data. Multi-horizon forecasting supported via the DIRECT method. Given a response vector y, a predictor matrix x is built that consists of lagged y vectors. Additional future response vectors are built for training. y_t = b dot x where x = [y_{t-1}, y_{t-2}, ... y_{t-lags}]. Matrix x includes constant, linear and generalized quadratic terms (x^pw where pw defaults to 2.0). For example, a lower power such as 1.5 may work better for longer horizons (pw may be tuned).
Attributes
- Graph
-
- Supertypes
-
class Objecttrait Matchableclass Any
- Self type
-
ARX_Quad_MV.type
Members list
Value members
Concrete methods
Create a RegressionMV object from a Time Series response vector y. The input/data matrix x is formed from the lagged y vectors as columns in matrix x. Quadratic terms are added to the model, one for each lag.
Create a RegressionMV object from a Time Series response vector y. The input/data matrix x is formed from the lagged y vectors as columns in matrix x. Quadratic terms are added to the model, one for each lag.
Value parameters
- h
-
the forecasting horizon (1, 2, ... h)
- hparam
-
the hyper-parameters ((use Regression.hp for default)
- intercept
-
whether to add a column of all ones to the matrix (intercept)
- lags
-
the maximum lag included (inclusive)
- pw
-
the power to raise the variables to (x_j ^ pw) defaults to 2.0
- y
-
the original un-expanded output/response vector
Attributes
Create a RegressionMV object from a response vector to fit a quadratic surface to Time Series data. The input/data matrix x is formed from the lagged y vectors as columns in matrix x. In addition, lagged exogenous variables are added.
Create a RegressionMV object from a response vector to fit a quadratic surface to Time Series data. The input/data matrix x is formed from the lagged y vectors as columns in matrix x. In addition, lagged exogenous variables are added.
Value parameters
- elag1
-
the minimum exo lag included (inclusive)
- elag2
-
the maximum exo lag included (inclusive)
- h
-
the forecasting horizon (1, 2, ... h)
- hparam
-
the hyper-parameters (use Regression.hp for default)
- intercept
-
whether to add a column of all ones to the matrix (intercept)
- lags
-
the maximum lag included (inclusive)
- pw
-
the power to raise the variables to (x_j ^ pw) defaults to 2.0
- y
-
the original un-expanded output/response vector
Attributes
Create a RegressionMV object from a response matrix. The input/data matrix x is formed from the lagged y vectors as columns in matrix x. This method provides data rescaling.
Create a RegressionMV object from a response matrix. The input/data matrix x is formed from the lagged y vectors as columns in matrix x. This method provides data rescaling.
Value parameters
- h
-
the forecasting horizon (1, 2, ... h)
- hparam
-
the hyper-parameters (use Regression.hp for default)
- intercept
-
whether to add a column of all ones to the matrix (intercept)
- lags
-
the maximum lag included (inclusive)
- pw
-
the power to raise the variables to (x_j ^ pw) defaults to 2.0
- y
-
the original un-expanded output/response vector