Statistics 361/661: DATA ANALYSIS
Statistics 361/661: Data Analysis
Yale University, Fall 1995
Prof. Nicolas Hengartner, office: 207, 24 Hillhouse.
Analyzing various data sets using the Splus
programming environment, a selection of statistical methods are
studied. The emphasis is on the practical applications of these methods
to data sets. Weekly sessions at the Social Science Statistical
Laboratory will be held. Topics covered include:
- Univariate Data Summary: classical univariate statistics,
density estimation, bootstrap, jackknife and permutation methods.
- Linear Statistical Models: linear regression,
regression diagnostics, predictions, factorial designs.
- Generalized linear models: Binomial data, Poisson data,
two way tables.
- Non-linear regression and general maximum likelihood.
- Modern (non-parametric) regression: Additive models and
scatterplot smoothers, projection-pursuit regression, neural networks.
- Survival Analysis: estimators of the survival function,
parametric models, Cox proportional hazard model.
- Multivariate analysis: graphical methods, cluster analysis,
- Tree-based methods.
- Time series: ARIMA models, seasonality.
- Spatial Statistics: point process analysis,
interpolation and kriging.
Venable, W.N. and Ripley, B.D. (1994),
Modern Applied Statistics with S-plus, Springer-Verlag.
The grades will composed from
- 60% Assignments,
- 40% Take-Home Final.