Epidemiological forecast by state-of-the-art Gaidai multimodal bio-reliability scheme, combined with extrapolation by self-deconvolution

Presented case study advocates state-of-the-art spatiotemporal multivariate prognostics approach for epidemic/pandemic outbreak risk/hazard assessment. Novel statistical methodology has been applied to clinical unfiltered datasets. To provide robust and reliable long-term prognostics of flu-type outbreak future risks, current research advocates bio-reliability prognostic approach, suitable for multi-regional bio, environmental, public (national) health systems, clinically monitored across representative periods. Novel non-parametric deconvolution extrapolation scheme was employed. The current study utilized clinically daily reported patient COVID-19 or SARS-COV-2 related counts, throughout major administrative locations in the Netherlands. It is seen from Fig. 7 that if epidemic would last 20 years instead of 1 year, then daily case number global maximum would increase less than twice. A novel non-parametric deconvolution scheme was employed for extrapolation towards design return periods. The primary advantage of a non-parametric extrapolation scheme over existing parametric schemes lies within its numerical stability and accuracy. Key objective was to benchmark proposed multimodal bio-reliability and risk assessment method, based on underlying recorded raw (clinical) patient data, accounting for territorial mapping. The latter has critical significance for early epidemiological prognostics. This investigation aims at further development of efficient multivariate bio-reliability approaches, that are suitable for high-dimensional biosystems, facilitating epidemiological dynamics prognostics. The primary goal of this investigation was to provide a case study, manifesting the capabilities of the novel multivariate bio-reliability methodology. The analyzed public health dynamic system was modelled as 350-Dimensional (350D), according to the number of spatial locations. Cross-validation versus 4-parameter Weibull-type fit had been done. The proposed multivariate spatiotemporal reliability approach may be utilized for a range of biomedical/bioinformatical applications – as well as for general/common welfare.