About this question
Hard · machine_learning · Quant Researcher interview question · cross_validation, time_series, finance, data_processing
Standard K-Fold cross-validation is unsuitable for financial time series due to label overlap and serial correlation, which cause information leakage and optimistic performance estimates. Purged K-Fold Cross-Validation addresses these issues by removing training observations that overlap with the test set's label interval or immediately follow it. This method ensures that the model is trained on data strictly independent of the testing period, mimicking a realistic trading scenario. Task Implem