Abstract
This dissertation is concerned with a discrete population of particles in a steady state compartmental system. The system is considered to have m compartments and the transitions are stochastic in nature. Such systems may be used to model a variety of problems, and several diverse applications are sketched in this work. Efficient estimation of these transition rate parameters requires the associated distribution theory. This dissertation advances the distribution theory considerably by providing a compact analytic solution to the compartmental problem. Often in practice, individual compartments are inaccessible for observation and instead time series data are available only on the passage of material to the system exterior. The covariance kernel of such observations is derived in this paper and utilized for "efficient" parameter estimation. The recommended estimation procedure is demonstrated on both stimulated and biological data, and in the process the merits of the procedure are clarified. The dissertation concludes by outlining several possible avenues of future research. Each suggested extension would be useful from a practical standpoint.
Matis, James Henry (1971). Stochastic compartmental analysis: model and least squares estimation from time series data. Doctoral dissertation, Texas A&M University. Texas A&M University. Libraries. Available electronically from
https : / /hdl .handle .net /1969 .1 /DISSERTATIONS -178732.