Back to Quantitative trading

Module 06 · Validation and research

Validation and financial machine learning

Information-driven dollar bars, triple-barrier labeling, fractional differentiation memory preservation, and the Deflated Sharpe Ratio.

4 lessons · 4 graded exercises·3.5 hours·advanced·Prerequisite: Portfolio risk and optimal liquidation

Financial data structures and dollar bars

20 min · exercise

Replace arbitrary time bars with information-driven dollar bars to restore statistical normality in financial return distributions.

The triple-barrier labeling method

25 min · exercise

Replace fixed-horizon classification with path-dependent profit-take, stop-loss, and vertical timeout barriers.

Fractional differentiation and memory

25 min · exercise

Make non-stationary price series stationary while preserving long-term predictive memory using fractional differencing.

Multiple testing and the Deflated Sharpe Ratio

25 min · exercise

Correct backtested performance metrics for selection bias, number of trials, non-normal returns, and sample length.