> ML_ALGORITHM // VECTOR-AUTOREGRESSION-VAR_v1.0
Vector Autoregression (VAR & VECM)
Classic econometric model that captures the linear dynamic interdependencies among multiple parallel time series by treating every variable symmetrically.
Multivariate Econometric Modelstime-series-forecastinghigh-intrinsicsmall (<1k)
Back to All AlgorithmsComputational Complexity
Training Complexity:O(n * (k * p)^2 + (k * p)^3) where k is number of series
Inference Complexity:O((k * p)^2)
Hardware Profile
CPU Friendly:Yes
Requires GPU:No
Memory Footprint:low
Interpretability & Data
Interpretability Tier:high-intrinsic
Training Data Needs:small (<1k)
Interpretability Assessment
Directly yields Granger causality test statistics and orthogonalized impulse response functions (IRF).
Suitable Tasks & Supported Modalities
Suitable Tasks:
multivariate econometricsgranger causality analysisimpulse response analysis
Supported Modalities:
time-series
Implementing Libraries
statsmodelsstatsmodels Developers / NumFOCUS · v0.14.4
View Speccaret-r
Foundational Literature
Common Pitfalls & Warnings
- Parameter count explodes as O(k^2 * p); modeling more than 10-15 series simultaneously causes severe degrees-of-freedom exhaustion and overfitting