STATISTICAL TOOLS FOR SPATIAL PROCESSES

DURING THE LAST FEW YEARS THERE HAS EMERGED A CLEAR PERCEPTION OF THE PERVASIVE IMPORTANCE OF SPATIAL STATISTICSTHROUGHOUT MANY SCIENTIFIC AND ENGINEERING PROBLEM AREAS HAVING CONSIDERABLE NATIONAL INTEREST. FOR EXAMPLE, ONE FINDS THAT SPATIAL STATISTICS PLAYS A CRUCIAL ROLE IN SUCH DIVERSE PROBLEMS AS REMOTE SENSING, IMAGE PROCESSING, MINING, MULTIPLE TARGET DETECTION AND TRACKING, ETC. AT THESAME TIME, IT HAS HAPPENED THAT MANY OF THE MAJOR CONTRIBUTIONS TO THE FOUNDATIONS OF SPATIAL STATISTICS HAVE BEEN MADE OUTSIDE THE UNITED STATES. THE RESEARCH PROPOSED HERE FOCUSES ON THREE SPECIFIC TOPICS WITHIN THE FIELD OF SPATIAL STATISTICS: (1) MODELING AND ESTIMATING SPATIO-TEMPORAL STOCHASTIC PROCESSES, (2) JACKKNIFING AND BOOTSTRAPPING, (3) SMOOTHING SPATIAL DATA AND COMPUTING LINEAR PARTS OF 2-DIMENSIONAL ROBUST SMOOTHERS. IN ADDITIONWE INTEND TO EMBED THE METHODOLOGIES DEVELOPED INTO THE ENVIRONMENT OF THE INTERACTIVE STATISTICAL LANGUAGE AND SYSTEM S-PLUS. WE WILL ALSO CARRY OUT INITIAL STUDIES ON THE DEVELOPMENT OF POWERFUL GRAPHICAL DISPLAY TECHNIQUES FORANALYZING SPATIAL PROCESSES IN THE S-PLUS ENVIRONMENT.