Autocorrelation and Time Series Analysis
Nong Ye · 2013
Time series data consist of data observations over time. If data observations are correlated over time, time series data are autocorrelated. Time series analysis was introduced by Box and Jenkins (1976) to model and analyze time series data with autocorrelation. Time series analysis has been applied to real-world data in many fields, including stock prices (e.g., S&P 500 index), airline fares, labor force size, unemployment data, and natural gas price (Yaffee and McGee, 2000). There are stationary and nonstationary time series data that require different statistical inference procedures. In this chapter, autocorrelation is defined. Several types of stationarity and nonstationarity time series are explained. Autoregressive and moving average (ARMA) models of stationary series data are described. Transformations of nonstationary series data into stationary series data are presented, along with autoregressive, integrated, moving average (ARIMA) models. A list of software packages that support time series analysis is provided. Some applications of time series analysis are given with references.