Course Outline

Class 1:  Get to know your data

1. Overview of forecasting: how do you predict the future? What kind of accuracy is possible?
2. Where to obtain data; data sources at Fuqua and on the web
3. How to move data around: useful things you can do with your word processor and spreadsheet
4. What to look for in data: seasonality, inflation, trends, cycles, etc.
5. How to transform data to reveal its structure; deflation, logging, seasonal adjustment
6. Illustration of basic operations in Statgraphics
7. Forecasts and confidence intervals for the simplest case: the "mean" model

Lecture notes:

Famous forecasting quotes
How to shovel data around
Get to know your data
Inflation adjustment (deflation)
Seasonal adjustment
Stationarity and differencing
The logarithm transformation
Mean (constant) model

Class 2:   Introduction to forecasting

1. Forecasting a nonstationary series I: the trend line model
2. Forecasting a nonstationary series II: the random walk ("naive") model
3. How to identify a random walk: differencing and autocorrelation analysis
4. Geometric random walk: the basic stock price model
5. Three types of forecasts: estimation period, validation period, and extrapolations into the future
6. How to evaluate forecast errors and compare models

Videos: #7, 8, 9, 10, 11, 12

Lecture notes:

Linear trend model
Random walk model
Random walk model with drift
Geometric random walk model
Three types of forecasts: estimation period, validation period, and the future

Class 3:  Modeling of seasonality

1. General considerations in working with seasonal data: causes of seasonality, stability of seasonal patterns
2. Seasonal random walk; and seasonal random trend models
3. Seasonal adjustment by the ratio-to-moving-average method
4. Additive versus multiplicative seasonal adjustment
5. Adjustments for holidays and trading days
6. Trend/cycle decomposition of time series

Videos: #13, 14

Lecture notes:  handed out in class

Class 4:   Averaging and smoothing models

1. Simple moving average model
2. Exponential smoothing models
3. Combination of smoothing and seasonal adjustment

Lecture notes:

Averaging and exponential smoothing models
Spreadsheet implementation of seasonal adjustment and exponential smoothing

Class 5:  Regression to mediocrity

1. Indiana Jones and the temple of R-squared
2. Correlation coefficients
3. Fitting simple regression models; interpreting output
4. Using lagged and differenced variables in regression models
5. Confidence intervals for regression forecasts

Videos: #4, 5, 6

Lecture notes:

Introduction to regression analysis
What to look for in regression output
What's the bottom line? How to compare models
Additional notes on regression analysis
Spreadsheet for illustrating regression formulas

Class 6: Time series regression models

1. Economic interpretation of regression coefficients
   
2. Using auto- and cross-correlation plots to identify useful lags of variables
3. Dummy variables
   
4. How to model seasonality with regression
5. Log-log (constant elasticity) models
   

Videos: #15, 16

Class 7:  Regression continued

1. IN-CLASS QUIZ
2. Confidence limits for sums of coefficients
3. Use of the time index as a regressor
   
4. Predicting the future with regression models
   

Class 8:  Advanced regression methods, GLM and ANOVA

1. Stepwise and all-subsets regression
2. Categorical independent variables (ANOVA)
3. General linear models (GLM)
4. Regression models with hold-out samples
   

Videos: #17, 18

Lecture notes:

Testing the assumptions of linear regression
Stepwise and all-possible-regressions

Class 9:  Introduction to ARIMA models

1. Random walk + Autoregressive + Exponential Smoothing = ARIMA
2. Using ACF and PACF plots to determine the "signature" of a time series
3. Fitting non-seasonal ARIMA models
4. The spectrum of ARIMA models

Videos: #19

Lecture notes:

Class 10: ARIMA continued; seasonal models

1. Identification of seasonal models
2. Examples of seasonal model-fitting
3. Spreadsheet implementation

Videos: #20, 21, 22

Lecture notes:

Seasonal differencing
Seasonal random walk
Seasonal random trend
Seasonal ARIMA models
Summary of rules for identifying ARIMA models

Class 11:  ARIMA with regressors, sales of new products

1. Combination of ARIMA and regression models
2. Sales of new products:  the BASS model
   
3. Recap of steps in choosing a forecasting model
4. Review of models: what to use and when

Lecture notes:

 Class 12:  Automatic forecasting;  Political, ethical, and management issues

1. Automatic forecasting software
2. Political and ethical issues in forecasting

Lecture notes:

Automatic forecasting software
Political and ethical issues in forecasting
How to avoid trouble

Last updated May 16, 2005.  Always under construction.