/

This presentation explores:
- Principal Components Analysis (PCA) and Independent Components Analysis (ICA) (including their main characteristics, similarities and differences)
- Extreme events and what drives markets
- Similarities between PCA and ICA and how the two can be blended to enhance risk models and refine portfolio construction techniques

[as pdf]

Slides

1	Extreme events: Blending principal components analysis with independent components analysis
2	Extreme events: Blending principal components analysis with independent components analysis
3	PCA vs. ICA - main characteristics and differences
4	Relevance for risk and portfolio construction modelling
5	Visualising extreme events, i.e. fat tails: single return series
6	Fat-tailed behaviour depends partly on timescale
7	Fat tails and portfolio construction
8	Fat tails – in joint return series
9	Visualisation of joint fat-tailed behaviour
10	Quantile-quantile box plots
11	Principal components analysis (PCA)
12	Applying PCA to sector relatives
13	Principal components - explanation or noise?
14	Independent components analysis
15	E.g. Non-Normality and Projection pursuit
16	Projection pursuit algorithm (1)
17	Projection pursuit algorithm (2)
18	Independent components analysis (ICA)
19	Blending together PCA and ICA
20	Identifying Principal Components one at a time
21	Blending PCA with ICA
22	Extreme events appear to be very important!
23	Limitations
24	Summary
25	Appendix A: Time-varying volatility in single return series
26	Interpretation via Cornish-Fisher asymptotic expansion
27	Flaws in Cornish Fisher (and hence in skew/kurtosis)
28	A better approach?
29	Time varying volatility explains some market index fat tails, particularly on the upside
30	Not just a developed market phenomenon
31	More periods give more scope for extreme events
32	Time-varying volatility remains an important contributor
33	Appendix B: Time-varying volatility in joint return series
34	Applying PCA to sector relatives
35	Adjusting for time-varying volatility in joint return series
36	Longitudinal time-varying volatility adjustment
37	Cross-sectional time-varying volatility adjustment
38	Back-testing time-varying volatility adjustments
39	Other sources of fat tails?
40	Important Information

---

NAVIGATION LINKS
Contents | Next | Library

---