Set Specific Lags¶

Example created by Wilson Rocha Lacerda Junior

Looking for more details on NARMAX models? For comprehensive information on models, methods, and a wide range of examples and benchmarks implemented in SysIdentPy, check out our book: Nonlinear System Identification and Forecasting: Theory and Practice With SysIdentPy

This book provides in-depth guidance to support your work with SysIdentPy.

Different ways to set the maximum lag for input and output

pip install sysidentpy

from sysidentpy.model_structure_selection import FROLS
from sysidentpy.basis_function import Polynomial

Setting lags using a range of values¶

If you pass int values for ylag and xlag, the lags are defined as a range from 1-ylag and 1-xlag.

For example: if ylag=4 then the candidate regressors are \(y_{k-1}, y_{k-2}, y_{k-3}, y_{k-4}\)

basis_function = Polynomial(degree=1)

model = FROLS(
    order_selection=True,
    ylag=4,
    xlag=4,
    info_criteria="aic",
    basis_function=basis_function,
)

Setting specific lags using lists¶

If you pass the ylag and xlag as a list, only the lags related to values in the list will be created. \(y_{k-1}, y_{k-4}\), \(x_{k-1}, x_{k-4}\)

model = FROLS(
    order_selection=True,
    ylag=[1, 4],
    xlag=[1, 4],
    info_criteria="aic",
    basis_function=basis_function,
)

Setting lags for Multiple Input Single Output (MISO) models¶

The following example shows how to define specific lags for each input. One should notice that we have to use a nested list in that case.

# The example considers a model with 2 inputs, but you can use the same for any amount of inputs.

model = FROLS(
    order_selection=True,
    ylag=[1, 4],
    xlag=[[1, 2, 3, 4], [1, 7]],
    info_criteria="aic",
    basis_function=basis_function,
)
# The lags defined are:
# x1(k-1), x1(k-2), x(k-3), x(k-4)
# x2(k-1), x1(k-7)