SAS/ETS® Programming Training

• Illustrate the features and functions available in SAS/ETS^®
• Demonstrate the use of different models used for time series
• Recommend suggestions based on Autoregressive Integrated Moving Average (ARIMA) model
• Find autoregressive errors
• Elucidate State Space modeling
• Work on periodic data, harmonic frequencies, spectral density for spectral analysis
• Apply Scorecard, Forecasting data model, Goal-Seeking model for data mining and forecasting

• MBA graduates
• Academicians
• PhD scholars
• Survey researcher
• Market research
• Banking operations professionals
• IT professionals/consultants
• Risk analyst
• Forecaster

• Understanding of the statistical concepts and Econometrics
• Familiarity with regression analysis
• Knowledge of Base SAS or at least two years of working experience in SAS environment

1. Overview of Time Series

• Introduction
• Analysis Methods and SAS/ETS Software
  • Options
  • How SAS/ETS Software Procedures Interrelate
• Simple Models: Regression
  • Linear Regression
  • Highly Regular Seasonality
  • Regression with Transformed Data

2.  Simple Models: Autoregression

• Introduction
  • Terminology and Notation
  •  Statistical Background
• Forecasting
  • Forecasting with PROC ARIMA
  • Backshift Notation B for Time Series
  • Yule-Walker Equations for Covariances
•  Fitting an AR Model in PROC REG

3.  General ARIMA Model

• Introduction
  • Statistical Background
  • Terminology and Notation
• Prediction
  • One-Step-Ahead Predictions
  • Future Predictions
• Model Identification
  • Stationarity and Invertibility
  • Time Series Identification
  • Chi-Squared Check of Residuals
  • Summary of Model Identification
• Examples and Instructions
  •  IDENTIFY Statement for Series
  • Example: Iron and Steel Export Analysis
  • Estimation Methods Used in PROC ARIMA
  • ESTIMATE Statement for Series
  • Nonstationary Series
  • Effect of Differencing on Forecasts
  • Examples: Forecasting IBM Series and Silver Series
  • Models for Nonstationary Data
  • Differencing to Remove a Linear Trend
  • Other Identification Techniques

4.  ARIMA Model: Introductory Applications

• Seasonal Time Series
  • Introduction to Seasonal Modeling
  • Model Identification
• Models with Explanatory Variables
  • Case 1: Regression with Time Series Errors
  • Case 1A: Intervention
  • Case 2: Simple Transfer Function
  • Case 3: General Transfer Function
  • Case 3A: Leading Indicators
  • Case 3B: Intervention
• Methodology and Example
  • Case 1: Regression with Time Series Errors
  • Case 2: Simple Transfer Functions
  • Case 3: General Transfer Functions
  • Case 3B: Intervention
• Further Examples
  • North Carolina Retail Sales
  • Construction Series Revisited
  • Milk Scare (Intervention)
  • Terrorist Attack

5.  ARIMA Model: Special Applications

• Regression with Time Series Errors and Unequal Variances
  • Autoregressive Errors
  • Example: Energy Demand at a University
  • Unequal Variances
  • ARCH, GARCH, and IGARCH for Unequal Variances
• Cointegration
  • Introduction
  • Cointegration and Eigenvalues
  • Impulse Response Function
  • Roots in Higher-Order Models
  • Cointegration and Unit Roots
  • An Illustrative Example
  • Estimating the Cointegrating Vector
  • Intercepts and More Lags
  • PROC VARMAX
  • Interpreting the Estimates
  • Diagnostics and Forecasts

6.  State Space Modeling

• Introduction
  • Some Simple Univariate Examples
    • A Simple Multivariate Example
  • Equivalence of Statespace and Vector ARMA Models
• More Examples
  • Some Univariate Examples
  • ARMA (1, 1) of Dimension
• PROC STATESPACE
  • State Vectors Determined from Covariances
  • Canonical Correlations
  • Simulated Example

7.  Spectral Analysis

• Periodic Data: Introduction
• Example: Plant Enzyme Activity
• PROC SPECTRA Introduced
• Testing for White Noise
• Harmonic Frequencies
• Extremely Fast Fluctuations and Aliasing
• The Spectral Density
• Some Mathematical Detail (Optional Reading)
• Estimating the Spectrum: The Smoothed Periodogram
• Cross Spectral Analysis
  • Interpreting Cross-Spectral Quantities
  • Interpreting Cross-Amplitude and Phase Spectra
  • PROC SPECTRA Statements
  • Cross-Spectral Analysis of the Neuse River Data
  • Details on Gain, Phase, and Pure Delay

8.  Data Mining and Forecasting

• Introduction
• Forecasting Data Model
• Time Series Forecasting System
• HPF Procedure
• Scorecard Development
• Business Goal Performance Metrics
• Graphical Displays
• Goal-Seeking Model Development