kats.models.var module

VAR forecasting Model

VAR model is a multivariate extension of the univariate autoregressive (AR) model. It captures the linear interdependencies between multiple variables using a system of equations. Each variable depends not only on its own lagged values but also on the lagged values of other variables. We use the implementation in statsmodels and re-write the API to adapt Kats development style.

Typical usage example:

params = VARParams() m = VARModel(data=TSData_multi, params=params) m.fit() res = m.predict(steps=30) m.plot()

class kats.models.var.VARModel(data: kats.consts.TimeSeriesData, params: kats.models.var.VARParams)

Bases: Generic[kats.models.model.ParamsType]

Model class for VAR

This class provides fit, predict, and plot methods for VAR model

data

the input time series data as kats.consts.TimeSeriesData

params

the parameter class defined with VARParams

fit(**kwargs) → None

Fit VAR model

static get_parameter_search_space() → List[Dict[str, Any]]

Provide a parameter space for VAR model

Move the implementation of get_parameter_search_space() out of var to avoid the massive dependencies of var and huge build size.

plot() → None

Plot forecasted results from VAR model

predict(steps: int, include_history: bool = False, **kwargs) → Dict[str, kats.consts.TimeSeriesData]

Predict with the fitted VAR model

Parameters

• steps – Number of time steps to forecast

• include_history – optional, A boolearn to specify whether to include historical data. Default is False.

• freq – optional, frequency of timeseries data. Defaults to automatically inferring from time index.

• alpha – optional, significance level of confidence interval. Defaults to 0.05

Returns

Disctionary of predicted results for each metric. Each metric result has following columns: time, fcst, fcst_lower, and fcst_upper

class kats.models.var.VARParams(**kwargs)

Bases: kats.consts.Params

Parameter class for VAR model

This is the parameter class for VAR forecasting model which stands for Vector Autoregressive Model.

maxlags

Maximum number of lags to check for order selection, Defaults to 12 * (nobs/100.)**(1./4)

method

Estimation method to use Defaults to OLS

ic

Information criterion to use for VAR order selection Defaults to None

trend

“c” - add constant (Default), “ct” - constant and trend, “ctt” - constant, linear and quadratic trend, “n”/“nc” - no constant, no trend

validate_params()

Validate the parameters for VAR model