Regression and Time Series Analyses

• Activity Number: 397
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Section for Statistical Programmers and Analysts
• Author(s): Theresa Ngo*+
• Companies: Warner Bros. Entertainment Group
• Keywords: Regression Analysis ; Time Series Analysis ; Autocorrelation Function (ACF) ; Partial Autocorrelation Function (PACF) ; Cross-Correlation ; Box-Jenkins

Abstract

How do we build an appropriate model for a time series data? The modeling depends on the data examination, the number of observations available, and the objective of the analysis. One common forecasting method is the regression model, which is known to many analysts. However, applying regression models to a time series can be dangerous when mistakenly assuming that the random errors are independent. This can lead to biased estimates and poor forecasts. If the random errors are correlated, this indicates a time series model may be more appropriate to fit the data. This paper shows step-by-step instructions - trend and seasonality, assumptions, statistical tests, and diagnostics - on how to execute and validate a regression model and a time series model for different characteristics of the data.