FinM 331/STAT 339 Financial Data Analysis
Winter 2010
Mondays, 6:30-9:30PM, Ryerson 251, Chicago
Mondays, 7:30-10:30PM, UBS, Stamford
Tuesdays, 7:30-10:30PM, Spring, Singapore
Syllabus+ and Webpage
28 December 2009 Edition
Visiting Professor Floyd B. Hanson
Course Outline (tentative)
Applied Statistical Analysis of
Financial Data with Computation in MATLAB
- Introduction (1 Lecture):
Data models and Discretized Diffusion log-returns,
Independent observations and their statistics,
Additive and Multiplicative models,
Law of large numbers and Monte Carlo applications,
Central limit theorem and Extreme tail errors.
- Exploratory Data Analysis (2.5):
Pseudo-Random Number Generators (RNGs: Uniform, Normal,
Exponential, Poisson, Compound Poisson),
Histograms, Cumulative histograms,
Kernel smoothing, Quantile-Quantile (Q-Q) plots,
Discretized Jump-diffusions,
Confidence Intervals and Value at Risk (VaR),
more Normal distribution poor at extremes,
short-fall statistics,
Profit and Loss (P&L) statments,
Cauchy flat tail distribution,
Data estimation,
Order statistics,
Extreme statistics (Fat tails, Pareto distribution,
Peaks Over Thresholds (POTs).
- Multivariate Statistics (1.5L):
Bivariate distribution, Bivariate Kernel smoothing,
Conditional expectation,
Covariance, Correlation Coefficient,
Bivariate 3D-histograms,
Hybrid fat and thin tail models,
Principal Component Analysis (PCA),
Multivariate sample means.
- Parametric Regression (1.5L):
Ordinary Least Squares (OLS),
Multiple linear regression,
MATLAB regression functions,
Maximum Likelihood Estimation (MLE),
MLE for discretized linear diffusions,
compound Poisson processes,
MLE for discretized linear jump-diffusions.
- Non-Parametric Regression (1L):
Nnumerical optimization with derivative-free
MATLAB fminsearch,
Black-Scholes European options,
Market calibration and Implied Volatility (IV),
Risk-Neutral options and IV,
more Kernel smoothing regression.
- General Parameter Estimation (1.5L):
Maximum Likelihood, Moment Methods and Bayesian.
- Hypothesis Testing and Analysis of
Variance (1L): Goodness of fit,
Confidence Intervals.
Prerequisite Knowledge:
- Introductory Probability: see for instance,
- Introductory Statistics: especially linear regression:
- Very Basic MATLAB: including Statistics Toolbox that comes
with the Student Edition
MATLAB Student Version,
http://www.mathworks.com/academia/student_version/.
- see also Nygaard's review session notes previously mentioned.
Some Related Resources of the Instructor:
Web Source (temp.): http://www.math.uchicago.edu/~hanson/finm331/
Email Comments or Questions to
fhanson at math dot uchicago dot edu )