The estimation of a lag between two time series
Kowalski, Kenneth G.
MetadataShow full item record
In this thesis the Box-Jenkins and the R-S array methods of ARMA model identification are reviewed. Several data sets are presented and identified as ARMA processes using these two methods. In studying the data sets of widths of successive growth rings of trees, it is conjectured that a lag structure exists between pairs of tree ring data sets. Two lag models are introduced to explain the behavior of the lag structure between pairs of tree ring data sets. Three methods for estimating a lag between two correlated time series are introduced, and their large sample properties are investigated through simulation. This thesis concludes with a lag analysis of the tree ring data sets and a discussion of the adequacy of the lag model.