A Note on the Finite Sample Bias in Time Series Cross-Validation
By: Amaze Lusompa
It is well known that model selection via cross validation can be biased for time series models. However, many researchers have argued that this bias does not apply when using cross-validation with vector autoregressions (VAR) or with time series models whose errors follow a martingale-like structure. I show that even under these circumstances, performing cross-validation on time series data will still generate bias in general.
Similar Papers
Temporal cross-validation impacts multivariate time series subsequence anomaly detection evaluation
Machine Learning (Stat)
Finds hidden problems in data streams.
The bias of IID resampled backtests for rolling-window mean-variance portfolios
Portfolio Management
Fixes money predictions that use old data.
On Statistical Inference for High-Dimensional Binary Time Series
Methodology
Helps computers understand patterns in yes/no data.