phat-tails 0.0.11 documentation
Introduction
Demo
Define
What Are Fat Tails?
the Phat Distribution
Characteristics
Fit
Peak-Over Threshold
Estimating the Tail - Double Bootstrap
Maximum Likelihood
Neural Networks
Forecast
ARMA-GARCH for Time Series
Phat-GARCH
Misc
References
API
.rst
.pdf
repository
open issue
suggest edit
Pareto Hybrids with Asymmetric Tails
¶
Introduction
P.H.A.T. -
P
areto
H
ybrids with
A
symmetric
T
ails
Installation
Quickstart
Dependencies
Enhancements
Demo
Download Data
Fit GARCH and Estimate
\(\alpha\)
in Both Tails
Fit
\(\mu\)
and
\(\sigma\)
with Machine Learning
Compare Fit with Gaussian and T
Generate Garch Forecasts
Define
What Are Fat Tails?
Thin vs. Fat
The Borderline Sigmoid
The Devil is in the Tails
Scale Invariance
Six Sigma is Not Six Sigma
the Phat Distribution
In Search of Two Tails
the Double Pareto
the CarBen Hybrid
Pareto Hybrids with Asymmetric Tails: a Mixture Model
Pitfalls
Characteristics
Density
Phat
Right CarBen
Left CarBen
Cumulative Distribution Function
Phat
Right CarBen
Left CarBen
Quantile Function
Phat
Right CarBen
Left CarBen
Moments
Carben Right
Carben Left
Mean
Variance
Fit
Peak-Over Threshold
Estimating the Tail - PoT
Estimating the Tail - Double Bootstrap
Tail Quest
the Hill Double Bootstrap
Testing
Caveats
Maximum Likelihood
Neural Networks
Fitting a Standard Gaussian
Fitting S&P 500 Daily Returns
Fitting the Phat
Failure of a Standard Loss Function
PhatLoss: A Custom Loss Function
PhatNet
Caution on Scaling
Forecast
ARMA-GARCH for Time Series
Non-Stationarity
ARMA
Importance of Scale
GARCH Forecasting
Projection with Expanding Window
Feed Forward for Out-of-Sample Projections
Fixed Windows
Fixed Window Across Entire Time Series
Garchcaster
Garchcaster Tests
Phat-GARCH
Approach
A Note on Scaling
Forecasting Time Series with Pareto Hybrids
Generate ARMA-GARCH Residuals
Fit Residuals to the Phat Distribution
Some More Notes on Scaling
Comparing Fit
Phatcast
1-Year Price Forecast
Phat-ARGARCH
T v Phat-ARGARCH
Skew T v Phat-ARGARCH
CAVEATS
Misc
References
API
Introduction