<< 1 >>
Rating: 
Summary: well written text on time series and longitudinal data
Review: Diggle covers most of the standard approaches to time series analysis in both the time and frequency domains. The first two chapters provide a gentle introduction with the important theory covered in Chapters 3 and 4. What makes it different is the treatment of repeated measures data which comes up in clinical trials and other forms of medical research. The data are time series since they are collected over time. However in engineering and the physical sciences, it is common to deal with the analysis of a single but long time series. In the medical field the time series is often a series of repeated measurements on a patient taken over a short amount of time. Typically, the series may consist of only 3 to 5 time points. However inference is still possible because there are many patients being studied under similar conditions. So whereas in the engineering applications we only have a partial (although long) realization of a single time series in biomedical applications we have many partial (short) realizations of many time series. Some of the methods of analysis are therefore different. The approach to repeated measures is given in Chapter 5 which is authoritative and useful but does not do justice to the subject. Fortunately Diggle, Zeger and Liang have written an entire book on longitudinal analysis that came out in 1994, four years after this book. Other good books have also been written subsequently.The latter chapters deal with case studies, model fitting and diagnostic checking and an interesting chapter on bivariate time series analysis.
<< 1 >>
| 
