The proposed research involves the following subject areas: survival analysis, classification and non-linear regression, the low-dose extrapolation problem, and statistical deconvolution and regularization. Among the topics planned for consideration are: an investigation of the efficiencies of goodness-of-fit procedures with censored data; a modification of the Gehan test or the Mantel-Haenszel test for the sequential analysis of randomly censored data; the development of a test for trend in survival distributions with randomly cnesored data; a theoretical investigation of questions concerning recursive partitioning schemes raised in previous research; the application of recursive partitioning schemes to the analysis of data sets collected by Dr. D. Morton of UCLA and by Dr. C. Zippin of UCSF; the development of a new technique of effecting the low-dose extrapolation procedure, the technique being related to D.R. Cox's proportional hazards models; a theoretical and empirical investigation of the effectiveness and resolvability of various methods of handling the low-dose extrapolation problem; an investigation of methods for deconvolving estimated probability density functions.
Showing the most recent 10 out of 12 publications
