STATISTICAL MODELING OF EXTREME EVENTS AND HEALTH RISKS

OUR RESEARCH TEAM IS A CONSOLIDATED GROUP OF 7 DOCTORS, INCLUDING THE IP AND FOUR RESEARCHERS, TWO OF WHICH WERE IP'S FORMER PHD STUDENTS. OUR WORKING TEAM CONSISTS OF THREE IP'S CURRENT PHD STUDENTS. WE TOTAL ELEVEN RESEARCHERS. THE FOLLOWING THREE MAIN TOPICS, OR KEY GOALS, WILL BE THE FOCUS OF OUR PROJECT, AS WELL AS THE ASSOCIATED MATHEMATICAL METHODS THAT WILL RESULT IN METHODOLOGICAL DEVELOPMENTS: REACHING HIDDEN AND UNDERREPORTED POPULATIONS WITH COVARIATES, ANALYSING ANOMALOUS AND CATASTROPHIC EVENTS, AND IMPROVING BIODOSIMETRY MODELS. THESE GOALS ARE DIVIDED INTO THE FOLLOWING 6 SUB-OBJECTIVES: RHUPC1: UNDERREPORTED TIME SERIES WITH INTERVENTIONS; RHUPC2: ESTIMATION OF THE SIZE OF HIDDEN POPULATIONS; ACAE1: ANALYSING ANOMALOUS BEHAVIOR OF CONTROL SOFTWARE; ACAE2: ANALYSING COMPLEX SYSTEMS; IBM1: IMPROVING MODEL SELECTION AND IBM2: BIVARIATE COUNT DISTRIBUTIONS IN BIODOSIMETRY. WITHIN THE SUB-OBJECTIVE RHUPC1, THE MAIN GOAL WILL BE TO STUDY AND PROPOSE NEW METHODOLOGICAL APPROACHES TO ASSESS THE ACTUAL IMPACT OF INTERVENTIONS CARRIED OUT IN ORDER TO MODIFY THE EVOLUTION OF THE NUMBER OF CASES OF DISEASES THAT ARE CURRENTLY A MATTER OF CONCERN FOR THE PUBLIC HEALTH COMMUNITY AND WHICH MIGHT BE MISREPORTED DUE TO THE DISEASE CHARACTERISTICS, AS THE INCREASE IN SEXUALLY TRANSMITTED INFECTIONS (STI) THAT HAS BEEN OBSERVED IN THE LAST YEARS. SUB-OBJECTIVE RHUPC2, WILL ALLOW US TO ESTIMATE THE NUMBER OF UNDISCOVERED NEAR EARTH OBJECTS (NEOS) ACCORDING TO THEIR SIZES, ALBEDOS, AND OTHER COVARIATES. NEOS ARE ASTEROIDS AND COMETS WHOSE TRAJECTORY MAY APPROACH EARTH’S ORBIT. DEPENDING ON THEIR SIZE AND THEIR DISTANCE TO THE EARTH’S ORBIT, SOME NEOS MAY REPRESENT AN IMPORTANT HAZARD BOTH AT LOCAL SCALES AND, POTENTIALLY, AT THE GLOBAL SCALE. ACAE1 PRESENTS TWO MAIN CHALLENGES: ENHANCING NEURAL NETWORKS WITH A HYBRID APPROACH USING EVT AND THE PERFORMANCE OF COMPREHENSIVE CHARACTERIZATION OF THE OUTLIER CONCEPT FOR SAFETY-CRITICAL SYSTEMS. THE FIRST IS GOING TO BE DEVELOPED FOR EXPLORING EVT TECHNIQUES MIXED WITH NEURAL NETWORK CAPABILITIES TO TACKLE CRITICAL SCENARIOS SUCH AS MEDICAL DEVICES WHERE ARTIFICIAL INTELLIGENCE PROCESSES ARE INVOLVED. THE SECOND IS GOING TO CONSIDER EXISTING OUTLIER DETECTION METHODS SUCH AS CONFORMAL ANOMALY DETECTION TO POTENTIALLY COMBINE THEM WITH OTHER LITERATURE METHODS. ACAE2 PROPOSES ADDITIONAL INNOVATIVE RESEARCH TO ELUCIDATE THE UNRESOLVED DEBATE BY LEVERAGING KNOWLEDGE FROM COMPLEX SYSTEM THEORY, INVESTIGATING HOW LAND USE AND LAND COVER (LULC) DATA INFORMATION IS DISTRIBUTED WITHIN BUFFER AREAS AND ASSESSING THE DISTRIBUTION PATTERN IN TERMS OF ENTROPY, CONSIDERING THE INFORMATION WITHIN AND BETWEEN BUFFERS, AND EVALUATING THE IMPACT OF USING DIFFERENT BUFFER RADII IN STATISTICAL HEALTH MODELING. IBM1 PROPOSES TO DEVELOP AN EXACT (NON-ASYMPTOTIC) TEST THAT DISTINGUISHES BETWEEN PARTIAL BODY IRRADIATION (ZERO-INFLATED POISSON DISTRIBUTION) AND HETEROGENEOUS EXPOSURES (MIXED POISSON DISTRIBUTION). MOREOVER, IT ALSO PROPOSES TO SEARCH FOR A MATHEMATICAL MODEL EXPLAINING THE FORMATION OF THE FOCI. THIS WILL BE USEFUL TO BE ABLE TO ESTIMATE THE RADIATION DOSE RECEIVED BY THE PATIENT THROUGH A STATISTICAL ANALYSIS OF THE FOCI OBSERVED IN A SAMPLE OF LYMPHOCYTES. IBM2 INTENDS TO DEVELOP A METHODOLOGY THAT ALLOWS ESTIMATING THE RADIATION DOSE THAT THE PATIENT HAS RECEIVED USING THE INFORMATION FROM TWO BIOMARKERS AT THE SAME TIME, WHICH ARE THE DICENTRICS AND THE FOCI.