Online Monitoring via Streaming Curves
-
- 0 Rating
- 0 Reviews
- 2 Students Enrolled
Online Monitoring via Streaming Curves
This courselet introduces real-time change detection for streaming curve data (e.g., intraday return curves, yield curves, volatility smiles). You’ll learn a practical monitoring workflow that is robust to dependence and volatility, and produces actionable alert rules for regime shifts.
-
- 0 Rating
- 0 Reviews
- 2 Students Enrolled
Requirements
- Comfort with basic time-series ideas (dependence, structural change, rolling/online analysis) Working knowledge of statistics and linear algebra at an applied level Ability to code in R or Python (loops, matrices, basic data cleaning)
General Overview
Description
- What “streaming curves” look like in quant workflows, and why standard pointwise monitoring often fails
- A train-then-monitor setup: build a stable baseline from historical curves, then monitor new curves as they arrive
- Dimension reduction for curves (turn curves into a small set of signals you can track in real time)
- A monitoring statistic designed to be low-tuning (avoids heavy bandwidth/parameter choices common in long-run variance methods)
- How to set thresholds and alert boundaries using precomputed/simulated critical values
- A market-style case study showing how the method flags a major regime transition, plus guidance on how to adapt it to your own data
Recommended for you
Meet the instructors !
