Statistical Models and Data
LEARNING OBJECTIVES
- Modelling
- Time Series
- Applied Data Mining
- Markov Chains
- RNG and Tests of Hypotheses
- Modelling, components of time series, estimates for trend and seasonal components, quality of the model.
- General forecasting issues.
- IBM SPSS Statistics
- Resources: Groebner et al.: Business Statistics ch. 15
- Additional: Review from chapter 2 (line chart), chapters 14 and 15 (linear and nonlinear regression models).
- Model diagnosis, fitting, and specification.
- Forecasting period, forecasting interval.
- Components of a time series. Trend, season, cycle.
- Random component. Seasonal component, seasonal index. Moving averages.
- Trend-based forecast. Linear trend, least squares equation.
- Errors MSE, MAD, autocorrelation.
- Durbin-Watson statistic, test for autocorrelation. Forecast bias.
- Adjust for seasonality, computing seasonal indexes, seasonality unadjusted forecast.
- Forecast using a single exponential smoothing.
- Link to the Self Study Quiz:
- http://wps.prenhall.com/bp_groebner_busstats_7/64/16436/4207761.cw/index.html
- DM process – description
- Data preparation
- Selected tasks: LDA, Regression, C&RT – principles and case studies, analysis with IBM SPSS Statistics use
- Understanding output information
- Interpret results – deployment
- Predictive quality – Methods and measures
- Regular chains, absorption chains, long – run properties of Markov chains.
- Resources: – Chapter 11. Markov Chains - Grinstead book http://www.dartmouth.edu/~chance/teaching_aids/books_articles/probability_book/book.html
- Random Numbers Generators.
- R(0,1) – uniform distribution and its properties.
- Congruential generators, transformation to N(0,1) random sample (Box – Mueller transformation, Central limit theorem).
- Kolmogorov – Smirnov test, and χ2 test.
- Applicaton of random numbers in DM and modelling.
Related materials