StackedML
Practice
Labs
Questions
Models
Pricing
Sign in
Questions
/
Model Evaluation & Experimentation
/
Validation
/
Time-series CV
← Previous
Next →
409.
K-Fold Inappropriateness for Time Series
easy
Why is standard k-fold cross-validation inappropriate for time series data?
A
K-fold underestimates variance in time series models since temporal patterns reduce effective sample size
B
K-fold requires i.i.d. data but time series violates this only when autocorrelation exceeds a fixed threshold
C
K-fold assumes equal fold sizes which cannot be guaranteed when time series data has irregular sampling intervals
D
Random splits allow future observations into training folds, leaking information about future outcomes
Sign in to verify your answer
← Back to Questions