Harvard University Press has partnered with De Gruyter to make available for sale worldwide virtually all in-copyright HUP books that had become unavailable since their original publication. The 2,800 titles in the “e-ditions” program can be purchased individually as PDF eBooks or as hardcover reprint (“print-on-demand”) editions via the “Available from De Gruyter” link above. They are also available to institutions in ten separate subject-area packages that reflect the entire spectrum of the Press’s catalog. More about the E-ditions Program »
During the twentieth century, statistical methods for analyzing time series have broadened considerably in both scope and feasibility. Notable among these developments is the spectral approach to time series analysis, which decomposes a time series into frequency components. In this volume, George Fishman presents a detailed account of the underlying theory for univariate, bivariate, and multivariate time series, and describes corresponding methods of statistical inference for spectral quantities of interest. He places special emphasis on the estimation of coefficients in distributed lag models.