Learn how to estimate daily volatility from noisy high-frequency data.
This guide explains how observation noise affects price moves and how to extract reliable variance estimates from rapid, tick-by-tick data.
The approach focuses on practical estimation in finance, showing how high-frequency data can reveal day-by-day volatility rather than just a single average. It discusses threats to accuracy, like micro-activities in the market, and presents methods that work even when noise is not perfectly constant or Gaussian.
- Identify what counts as observation noise and how micro-activities distort price signals.
- Learn how the maximum likelihood estimator (MLE) is applied to currencies and futures prices.
- Explore a quadratic estimator that remains robust when variances change over time or noise isn’t Gaussian.
- See when to prefer the quadratic estimator as an initial guess for more complex methods like MLE.
Ideal for readers of financial time series, econometrics, and practitioners seeking practical estimation methods for volatility and covariance in high-frequency data.