System and method for forecasting effect of pathogen on population

Abstract

Disclosed is a system including a computing hardware for processing input data defining parameters describing a population and a pathogen affecting or potentially affecting the population, wherein the computing hardware is configured to apply a mathematical simulation model to the 5 input data to generate output data providing a forecast indicative of an effect of the pathogen on the population, wherein the mathematical simulation model is configured to use at least a Susceptible-Infected- Recovered model (SIR model), characterized in that the system is configured initially to apply the SIR model to an historical part of the input 10 data to generate an initial iteration of the output data, and then to train the SIR model during a training period on a recently-acquired part of the input data to improve an accuracy of the SIR model when generating a subsequent iteration of the output data.

Description

[0001] SYSTEM AND METHOD FOR. FORECASTING EFFECT OF PATHOGEN ON

[0002] POPULATION

[0003] TECHNICAL FIELD

[0004] The present disclosure relates to systems for forecasting effects of pathogens on population. The present disclosure also relates to methods for forecasting effects of pathogens on population. The present disclosure further relates to software products for forecasting effects of pathogens on population.

[0005] BACKGROUND

[0006] Infectious diseases are a significant global health concern, which affects millions of people each year and poses challenges to healthcare departments worldwide. Accurate monitoring and prediction of a spread of the infectious diseases are crucial for effective public health management which allows countries to allocate healthcare resources effectively, develop prevention strategies, and prepare for potential outbreaks. Existing epidemiological models (such as Susceptible- Infectious-Recovered (SIR) model, machine-learning (ML) model, or the like) are utilized to predict the spread of the infectious diseases.

[0007] However, such exiting epidemiological models face significant limitations associated therewith. Firstly, the SIR model is incapable and insufficient in terms of capturing complex dynamics of disease spread, especially for diseases with varying transmission rates or irregular patterns. Additionally, the SIR model is good for diseases (for example, such as flu) with regular, repetitive patterns, but is ineffective for some diseases (for example, such as COVID-19) having irregular infection trends. Secondly, the ML model is effective for short-term disease spread forecasting but is unsuitable and unreliable for long-term forecasting of the impact of disease interventions. For example, the machine learning model is accurate and effective for about a 4-week forecast and struggles to forecast further.

[0008] Moreover, an accuracy of the existing epidemiological model is highly dependent on an availability and a quality of training data, which may be challenging, especially in resource-constrained environment settings. Furthermore, the existing epidemiological models require extensive adjustments for each disease, which limits their flexibility and adaptability for various other diseases. Moreover, such epidemiological models are usually developed by academics (for example, government, pharma companies, and the like) to model effects of vaccines on the disease spread. However, such models are usually highly specialised, which only works in specific scenarios, and as they perform poorly when applied broadly.

[0009] Therefore, in light of the foregoing discussion, there exists a need to overcome the aforementioned drawbacks.

[0010] SUMMARY

[0011] The present disclosure seeks to provide a system, a method, and a software product to generate highly accurate and reliable forecast indicative of an effect of a pathogen on a population, by way of employing an epidemiological model comprising more than three different compartments (namely, susceptible, incubating, infectious, hospitalised, recovered, and deceased compartments) as compared to a Susceptible- Infected-Recovered (SIR) model of the prior art. The aim of the present disclosure is achieved by a system, a method, and a software product for forecasting an effect of a pathogen on a population, as defined in the appended independent claims to which reference is made to. Advantageous features are set out in the appended dependent claims. Throughout the description and claims of this specification, the words "comprise" , "include", "have", and "contain" and variations of these words, for example "comprising" and "comprises" , mean "including but not limited to", and do not exclude other components, items, integers or steps not explicitly disclosed also to be present. Moreover, the singular encompasses the plural unless the context otherwise requires. In particular, where the indefinite article is used, the specification is to be understood as contemplating plurality as well as singularity, unless the context requires otherwise.

[0012] BRIEF DESCRIPTION OF THE DRAWINGS

[0013] FIG. 1 illustrates a block diagram of a system for forecasting an effect of a pathogen on a population, in accordance with an embodiment of the present disclosure;

[0014] FIG. 2 illustrates steps of a method for operating a system for forecasting an effect of a pathogen on a population;

[0015] FIG. 3 illustrates an architecture of a Susceptible-Infected-Recovered model (SIR) model, in accordance with an embodiment of the present disclosure;

[0016] FIG. 4 illustrates an exemplary scenario of processing input data by utilising a Susceptible-Infected-Recovered model (SIR) model, in accordance with an embodiment of the present disclosure;

[0017] FIG. 5 illustrates an exemplary graphical representation of tracking, forecasting and simulating infectious disease outcomes for short-term future and medium-term future, in accordance with an embodiment of the present disclosure;

[0018] FIG. 6 illustrates a comparison between a known forecast and a fitted forecast provided by implementing a system of the present disclosure; FIG. 7 illustrates a comparison between a forecast of a parameter by implementing a system of the present disclosure and an actual reported trend of the parameter;

[0019] FIGs. 8A and 8C illustrate two different graphical representations of an initial version of a forecast before performing a fitting process, while FIGs. 8B and 8D illustrate corresponding graphical representations of an aligned version of the forecast upon performing the fitting process in FIGs. 8A and 8C, respectively; and

[0020] FIGs. 9A-9H illustrate various graphical representations of comparisons between forecasts of different parameters and actual report data of said different parameters, wherein the forecasts of the different parameters are provided by a system for forecasting an effect of a pathogen on a population according to a first aspect.

[0021] DETAILED DESCRIPTION OF EMBODIMENTS

[0022] The following detailed description illustrates embodiments of the present disclosure and ways in which they can be implemented. Although some modes of carrying out the present disclosure have been disclosed, those skilled in the art would recognize that other embodiments for carrying out or practising the present disclosure are also possible.

[0023] In a first aspect, an embodiment of the present disclosure provides a system including a computing hardware for processing input data defining parameters describing a population of people and a pathogen affecting or potentially affecting the population, wherein the computing hardware is configured to apply a mathematical simulation model to the input data to generate output data providing a forecast that is indicative of an effect of the pathogen on the population, wherein the mathematical simulation model is configured to use at least a Susceptible-Infected-Recovered model (SIR model), characterized in that the system is configured initially to apply the SIR. model to an historical part of the input data to generate an initial iteration of the output data, and then to train the SIR model during a training period on a recently- acquired part of the input data to improve an accuracy of the SIR model when generating a subsequent iteration of the output data.

[0024] In a second aspect, an embodiment of the present disclosure provides a method for operating a system of the first aspect, wherein the method includes: configuring a computing hardware to processing input data defining parameters describing a population of people and a pathogen affecting or potentially affecting the population, wherein the method includes configuring the computing hardware to apply a mathematical simulation model to the input data to generate output data providing a forecast that is indicative of an effect of the pathogen on the population, wherein the mathematical simulation model is configured to use at least a Susceptible-Infected-Recovered model (SIR model), characterized in that method further includes: configuring the system initially to apply the SIR model to an historical part of the input data to generate an initial iteration of the output data, and then training the SIR model during a training period on a recently- acquired part of the input data to improve an accuracy of the SIR model when generating a subsequent iteration of the output data.

[0025] In a third aspect, an embodiment of the present disclosure provides a software product that is executable on a computing hardware of a system of the first aspect, to cause the system to implement the method of the second aspect.

[0026] The present disclosure provides the aforementioned system, the aforementioned method, and the aforementioned software product to generate highly accurate and reliable forecast indicative of an effect of the pathogen on the population, by way of employing an epidemiological model comprising more than three different compartments (namely, susceptible, incubating, infectious, hospitalised, recovered, and deceased compartments) as compared to the SIR. model of the prior art. Said epidemiological model is beneficially used for forecasting a spread of pathogens affecting the population in a highly accurate and reliable manner, by incorporating a time-varying infection rate in the SIR model. For this, the system and the method employ the SIR model in a different manner as compared to the prior art. Beneficially, the system and the method are susceptible for capturing complex dynamics of disease spread, especially for diseases with varying transmission rates or irregular patterns. The system and the method are reliable for long-term forecasting of an impact of disease interventions, and flexible and adaptable for various diseases. The system may utilise a Fourier series to capture dynamic changes in an infection rate, which helps the system to adapt to irregular patterns of the spread of the pathogens, such as of COVID-19. Additionally, the system combines the mathematical simulation model with a machine learning technique, which offers a comprehensive approach to pathogen spread. Furthermore, the system offers a real-time analysis of comparing forecasted data to actual data (i.e., ground-truth data) which could help public health officials make more informed decisions regarding medical resource allocation, disease prevention strategies, and outbreak preparedness, applicability, and effectiveness across different scenarios. The aforementioned system and the aforementioned method are simple, robust, fast, reliable, and can be implemented with ease. Throughout the present disclosure the term "computing hardware" refers to a physical computing system that is capable of processing the input data to generate the output data. Examples of the computing hardware include, but are not limited to, a laptop computer, a desktop computer, a tablet, and a phablet. Optionally, the computing hardware comprises at least a processor and a storage device. The computing hardware may be coupled to an input device (for example, a keyboard, a mouse, or the like) and an output device (for example, such as a monitor). Alternatively, the computer hardware is optionally implemented as a server (such as a cloud server).

[0027] Throughout the present disclosure the term "pathogen" refers to a microorganism that causes a disease. Such a microorganism could, for example, be a bacterium, a virus, a fungi, a prion, and the like. Typically, the pathogen has an ability to enter an organism (namely, a host), evade an immune system of the organism, replicate and cause damage to the organism tissues, leading to an illness or an infection. The pathogens may infect a variety of organisms, such as humans, animals, and plants, and are responsible for a wide range of infectious diseases, from a common cold to more serious illnesses such as tuberculosis, malaria, COVID-19, and the like.

[0028] Throughout the present disclosure the term "input data" refers to information pertaining to the population and the pathogen affecting or potentially affecting said population. Optionally, the input data for the population comprises information pertaining to at least one of: an age, a gender, an ethnicity, a population size, a geographic location, health status, pre-existing health conditions, socioeconomic status, environmental conditions. Additionally, optionally, the input data for the pathogen comprises a transmission rate of the pathogen, an incubation period, an infectious period, a basic reproduction number, a variability in transmission rate, an infectiousness by disease stage, severity and clinical manifestations, variants or mutations, asymptomatic and subclinical infections, an environmental stability, a host range and zoonotic potential, an antigenic drift and shift, a drug resistance. Moreover, the input data may also include at least one of: historical data about the pathogen affecting the population, for example, such as data on past infected cases, hospitalizations, and deaths related to the pathogen; environmental factors such as temperature, humidity, wastewater, and population density; information about interventions taken to control the spread of the pathogen, such as vaccination campaigns or social distancing measures; and any other relevant data that may be used for understanding and predicting an effect of the pathogen, such as travel patterns of the pathogen. It will be appreciated that the input data is served as a basis to the mathematical simulation model, for generating the forecast related to the effect of the pathogen on the population.

[0029] Optionally, the input data is acquired, at least in part, by using sensor arrangements to monitor the population of people. Herein, the term "sensor arrangement" refers to a collection of specialised devices that are employed to obtain the input data for the population. It will be appreciated that the sensor arrangements can be deployed in various places such as homes, workplaces, public spaces, and healthcare facilities to continuously collect data pertaining to the population's health status, behavior, and environmental conditions. It will also be appreciated that at least some of the input data would be acquired by the sensor arrangements, and remaining part of the input data may be acquired by some other means, for example, such as historical data, epidemiological studies, data from public health agencies, and the like. Optionally, a given sensor arrangement comprises at least one sensor. Examples of the at least one sensor include, but are not limited to, an environmental sensor, a biomedical sensor, a proximity sensor, a camera, a location tracking sensor, a temperature sensor, a pressure sensor, a biometric sensor, a heart rate monitor, an accelerometer. Furthermore, the input data acquired by the sensor arrangements may further comprise information pertaining to an ambient temperature (which could be relevant for identifying a seasonal variation of the pathogen), an air quality, a humidity, and other environmental factors which are relevant for understanding a transmission of the pathogen, due to such environmental factors.

[0030] Notably, the computing hardware is configured to apply the mathematical simulation model to the input data to generate output data. Throughout the present disclosure the term "mathematical simulation model" refers to a mathematics-based epidemiological model that is used to provide the forecast that is indicative of the effect of the pathogen on the population (namely, a dynamics of disease transmission within the population due to the pathogen).

[0031] It will be appreciated that an effect of the pathogen may vary depending on various factors such as the pathogen's ability to transmit from a person to another person. This could lead to new infections within the population, and development of illness or symptoms in infected individuals, which may range from mild to severe depending on the pathogen. This typically creates a need for infected individuals to seek medical care, potentially leading to increased hospitalizations and strain on healthcare systems. Furthermore, an overall impact of the pathogen on the population may include at least one of: rates of illness, complications, and death, economic and social impact, including economic disruptions, changes in social behavior, psychological impacts. In this regard, the mathematical simulation model is utilised to identify the effect of the pathogen on the population over a period of time. By running simulations on the input data, the mathematical simulation model can help predict a course of an outbreak, evaluate an effectiveness of control measures, and inform public health decision-making. Moreover, the mathematical simulation model uses at least the SIR. model for generating the forecast which indicates the effect of the pathogen. Typically, the SIR model divides the population into three groups: susceptible individuals (S), infected individuals (I), and recovered individuals (R). The susceptible individuals are those individuals who may become infected, the infected individuals are those individuals who are infected with the pathogen and are capable of transmitting it to the susceptible individuals, and the recovered individuals are those individuals who have recovered from the disease caused by the pathogen and are assumed to have immunity or who have been removed from the population due to death. In this regard, the SIR model describes an interplay between such sub-populations, in which an individual will become infected by the pathogen and possibly go to a hospital, then either will recover and eventually return to the susceptible population or pass away. Specific dynamics of the SIR model are determined by conditions of the sub-populations and a rate at which individuals would transition between the aforesaid groups of individuals. Beneficially, the SIR model facilitates in simulating different scenarios, such as an impact of interventions like vaccination or social distancing, and help in making effective public health strategies at a time of an epidemic. The SIR model is well-known in the art.

[0032] Optionally, the system is configured to simulate a spread of a disease arising from the pathogen, wherein an infection rate of the pathogen is allowed to be temporally variable in magnitude. Herein, the term "infection rate" refers to a rate at which the susceptible individuals become infected with the pathogen. Typically, the infection rate is a key parameter in the SIR model which simulates the spread of the disease arising from the pathogen. A high magnitude of the infection rate indicates that the pathogen transmitting rapidly through the population, leading to a faster increase in a number of infected individuals; and conversely, a low magnitude of the infection rate indicates a slower transmission and a slower increase in the number of infected individuals. In this regard, the temporal variability in the magnitude of the infection rate may occur due to various factors, such as changes in a behavior of the population (for example, increased social distancing during a pandemic), interventions (such as vaccination campaigns), seasonal variations in infection rates (for example, flu season), and similar. Beneficially, when the infection rate of the pathogen is temporally variable, the simulation of the spread of the disease arising from the pathogen can be highly accurately performed for accurate forecasts.

[0033] Optionally, the system is configured to receive initial estimates of one or more parameters entered by a user of the system, wherein the one or more parameters are utilised to determine at least temporal and coupling characteristics of differential and / or integral computations performed when executing the SIR. model, optionally wherein the initial estimates relate to parameters: beta (0), epsilon (6), gamma (y), eta (q), rho (p) delta (6). In this regard, the input data is provided to the SIR model along with the one or more parameters, for training the SIR model. An initial estimate may comprise at least one of: a lower bound value, an upper bound value, of a given parameter. It will be appreciated that for the initial iteration of the output data, the SIR model utilises the one or more parameters along with the input data available at that time. Subsequently, during the training period, values of the one or more parameters may be refined / adjusted based on additional input data (namely, more recent input data), to improve a prediction accuracy of the SIR model. The (trained) SIR model with the one or more updated parameters is then used for generating highly accurate and reliable forecasts.

[0034] The one or more parameters comprises an infection rate (represented as the beta), a rate of recovery (represented as the gamma), a rate of exposure (represented as the epsilon), a rate of hospitalisation (represented as the eta), a rate of becoming susceptible (represented as the rho), and a rate of death (represented as the delta). It is to be noted that the rate of recovery (namely, the gamma parameter) could be different for different individuals, depending on whether a given individual is recovering from any one of: an incubation group (E), an infectious group (I), a hospitalised group (H). The infection rate refers to a rate at which susceptible individuals become infected when they come into contact with infectious individuals and / or the pathogen. The recovery rate refers to a rate of how quickly infected individuals recover and become immune. The rate of exposure refers to a rate at which susceptible individuals in the population are exposed to the pathogen and become infected. The rate of hospitalisation refers to a rate at which infected individuals require hospitalisation due to a severity of their symptoms. The rate of becoming susceptible refers to a rate at which recovered individuals lose their immunity and again become susceptible to the pathogen.

[0035] Optionally, the initial estimates of the one or more parameters are provided by the user, for example, by way of using a user interface associated with the computing hardware. The user interface could be a web-based form, a desktop application, or a command-line interface, and the like. Moreover, the user interface may include fields or options for the user to select parameters and their initial estimates. Furthermore, the system may be configured to perform validation checks to ensure that an entered initial estimate of a given parameter is within acceptable ranges or follow specific formatting rules. This may potentially help in preventing errors and ensures that the entered initial estimate is relevant and meaningful. Once the user has entered the initial estimates, there may be an option to save and store these estimates in a database or a local memory (coupled to the computing hardware) for later use. Furthermore, the one or more parameters entered by the user (such as beta(P), epsilon(e), gamma(y), eta(q), rho(p), and delta (6)) facilitates in determining temporal aspects of computations to be performed by the SIR. model, as adjusting said one or more parameters, the user can modulate various aspects of disease transmission and progression within the SIR model. The temporal characteristics may use the time step used in the SIR model (for example, daily, weekly), as the input data to determine how often the SIR model is updated with new data, and how time is discretized in the model. Moreover, the coupling characteristic determines how the one or more parameters interact with each other over time. The coupling characteristics could include how changes in one population (for example, infected individuals) affect other populations (for example susceptible individuals). For example, the parameter beta (P) being the rate of infection may directly influence a rate at which susceptible individuals could become infected per unit of time.

[0036] It will be appreciated that the input data may comprise historic raw data (for example, a past number of cases, hospitalisations, deaths, wastewater, population demographics, and the like), simulated data (for example, a number of new cases over time, forecasted hospitalization rates, new estimated disease transmission rates, lockdowns, modeled demographic-specific infection rates, and the like), and projected epidemic curve and published epidemiological studies and literature data (for example, studies on disease transmission dynamic, reports on the efficacy of vaccines and treatments, and the like). During a training period, a simulated data acquired from an output simulation curve is used in conjunction with the historic raw data and the projected epidemic curve and published epidemiological studies and literature data, for complete parameter fitting, and generates the parameter-fitted curve. Said parameter fitting could be done by using existing well-known function scipy. optimize. Ieast_squares, as this is a machine learning procedure. Furthermore, the input data with adjusted parameters is provided to the SIR. model, and a machine learning model for generating subsequent iteration of the output data. The output data further comprises true epidemic levels (for example, modelled estimates for true disease burden cases, hospitalisations, deaths), scenarios (for example, vaccines and treatments strategies, different lock-down conditions) parameter analysis (for example, country dependence impact of vaccine or age, analysis of case fatality rates and the like), and forecasts (for example, forecasts for cases, hospitalisations and deaths).

[0037] It will also be appreciated that the temporal characteristics are used to describe how the output data changes over time. For the SIR model, the temporal characteristics may include rates of infection, recovery, and transmission over time. The coupling characteristics are used to describe how different components or variables in the input data are interconnected or influence each other, to generate a particular output data. For example, in the SIR model, the coupling characteristics may involve how a number of susceptible individuals affects the rate of infection, or how a number of recovered individuals affects a rate of immunity or transmission. The differential computations may involve using differential equations which describe how, in the SIR model, a number of susceptible, infectious, and recovered individuals change over time based on the one or more parameters. The differential equations may be solved numerically by using well-known numerical analysis techniques, for example, such as Euler's technique or Runge-Kutta technique. It is to be noted that in order to solve said different equations, initial values for each parameter and each type of population (namely, compartment) are pre-known and / or are estimated from (actual) reported data. Some differential equations that are employed in the SIR model are as follows: wherein / ?t refers to an infection rate, e refers to a rate of exposure, y refers to a rate of recovery, p refers to a rate of hospitalisation, p refers to a rate of becoming susceptible, cp refers to a rate of death, YErefers to a rate of recovery of exposed population, yi refers to a rate of recovery of infected population, yn refers to a rate of recovery of hospitalized population, and 5H refers to a rate of death of hospitalized population; and

[0038] S refers to a susceptible population, E refers to an exposed population, I refers to an infectious population, H refers to a hospitalised population, R. refers to a recovered population, D refers to a deceased population, T refers to a total population, 12 refers to a cumulative infections, H2 refers to a cumulative hospitalisations, and L is the cumulative number of discharged patients. The equations above describe the SIR. model for a late-stage COVID-19 pandemic. It is to be noted that it is referred to as a "late-stage" pandemic model as it assumed that everyone who dies of COVID-19 passes away in hospital, which was not true in an early pandemic when a non-negligible number of people passed away outside of hospital, notably in care-homes. However, there is the option to include a channel for deaths outside of hospital (between groups I to D), if an earlier timeperiod is considered. There are also three non-conventional subpopulations included, , H2 and L. These have been added to find the cumulative infections, hospitalisations, and discharges from hospital (or leaving hospital), which can in turn be used to find the daily new cases, new hospital admissions and daily hospital discharges, all of which help fit the model. It may be noted that a sum of the equations may equal to zero, meaning the population is conserved (including those who have passed away), and it does not account for births. The SIR model may have several parameters that are hard to measure and may be rarely reported. Despite similar naming conventions like fatality rates or hospitalisation rates that are often reported, they may not be equivalent as model parameters have 'per time' units whereas those reported values are often ratios (i.e. deaths per hospitalisation). Due to this fact the SIR model parameters are found using a parameter fitting procedure that fits the model outcomes to reported data (as discussed later).

[0039] Further, the integral computations may involve integrating solutions of the aforesaid differential equations to obtain cumulative values, and then to analyse a behavior of the SIR model over certain time intervals. For example, the integral computations could be used to calculate metrics, such as a total number of infections, a duration of an outbreak, and the like.

[0040] Notably, the system is configured to apply the SIR model to the historical part of the input data to generate the initial iteration of the output data. Herein, the term "output data" refers to information (comprising the forecast indicative of the effect of the pathogen on the population) generated by the SIR. model upon processing the input data. The forecast may pertain to at least one of: hospitalisation cases, deaths, vaccines and treatment strategies, different lockdown conditions, forecasts for scenarios based on changing parameters, an impact of vaccine or age, an analysis of case fatality rate (CFR), case hospitalization rate (CHR), hospitalization fatality rate (HFR), an economic impact, a burden on healthcare resources (such as including ICU bed occupancy, ventilator usage, and availability of medical supplies), public health interventions (such as testing strategies, contact tracing, and quarantine protocols), social behaviors, mental health impacts (such as stress levels, anxiety disorders, and rates of depression), education disruption, supply chain disruptions.

[0041] It will be appreciated that the initial iteration serves as a starting point for the SIR model, for further refinement and improvement of the SIR model. After generating the initial iteration of the output data, the SIR model is trained on the recently-acquired part of the input data during the training period. The recently-acquired part of the input data is new data which includes information that has been acquired or collected recently, and was not included in the historical data. The recently acquired part of the input data may include new case counts, new hospitalizations, new number of deaths, or other new relevant information that has been reported after collecting the historical part.

[0042] Optionally, the system is configured to use the forecast to apply one or more vaccines to the population for mitigating the effect of the pathogen on the population. In this regard, since the forecast generated by the SIR model may likely comprise information pertaining to an expected number of infected cases, hospitalisations, deaths, and the like, said information could be beneficially utilised for determining an optimal strategy for applying the one or more vaccines to the population, for example, by taking into account factors such as a vaccine efficacy, a coverage rate, and a potential for achieving herd immunity. The one or more vaccines may help in preventing individuals of the population from becoming infected, or it may reduce a severity of the disease when said individuals become infected. Furthermore, the forecast may also recommend vaccination strategies to mitigate the effect of the pathogen on the population by identifying priority groups in the population for applying vaccination, determining a timing and a dosage of the one or more vaccines, and monitoring a progress and an effectiveness of a vaccination campaign. It will be appreciated that by analyzing the forecast, it may be ensured that limited vaccine supplies are efficiently allocated to maximize impact and minimize overall disease burden.

[0043] The technical benefit of using the forecast to apply the one or more vaccines to the population is that it allows for a proactive and targeted approach to mitigate the effect of the pathogen on the population. By using the forecast to determine a timing and distribution of the one or more vaccines, requisite resources may be allocated efficiently, and a spread of the disease could be effectively controlled. This may potentially lead to a reduction in an actual number of cases, hospitalisations, and deaths, thereby reducing an overall burden of the disease (namely, the effect of the pathogen) on the population.

[0044] Optionally, during the training period, the SIR. model is trained on at least one of:

[0045] (i) a number and characteristics of the people designated as being patients;

[0046] (ii) an occurrence of new people being affected by the pathogen; (iii) a record of people afflicted by the pathogen being admitted to one or more hospital establishments;

[0047] (iv) a record of people who have died as a result of the effect of the pathogen.

[0048] In this regard, the recently-acquired part of the input data would comprise at least one of the aforesaid parameters, and when the recently- acquired part is used for training the SIR. model, the one or more parameters may be suitably adjusted to improve an accuracy of the SIR model when generating the subsequent iteration of the output data. For example, the SIR model may adjust the one or more parameters such as the infection rate and the recovery rate, based on the recently-acquired part of the input data. In this way, the SIR model can beneficially dynamically adapt to changing conditions of disease transmission dynamics (for example, when the infection rate of the pathogen is dynamically changing in a time period). By incorporating real-time data, the SIR model may provide accurate and up-to-date predictions of the disease transmission dynamics, enhancing the SIR model's reliability and usefulness for informing public health interventions and decision-making processes.

[0049] Optionally, the mathematical simulation model is configured to generate a Fourier series of coefficients that are fitted to estimated time-varying beta values generated by the SIR model to capture dynamic changes in the infection rate of the pathogen. In this regard, the Fourier series is used to capture periodic or repetitive patterns in the infection rate of the pathogen. By generating the Fourier series of coefficients fitted to the estimated time-varying beta values, the system effectively captures underlying patterns and fluctuations in the infection rate, which may vary due to factors such as seasonality, interventions, or changes in population behavior. Initially, the coefficients of the Fourier series may be set to some initial values by taking into account a given infection rate. These initial values are then adjusted during a fitting process to improve a fit of the Fourier series to the estimated time-varying beta values. It is to be noted that the fitting process is iterative, with the coefficients of the Fourier series being adjusted multiple times to improve the fitting. Said fitting process could be done by using existing well-known function scipy. optimize. Ieast_squares, as this is a machine learning procedure. Thus, the fitting process would continue until a satisfactory fit is achieved, where the Fourier series is closely matched to the estimated time-varying beta values. Once the fitting process is complete, the (fitted) Fourier series is validated by comparing its forecast corresponding to the recently-acquired portion of the input data that is not used in the fitting process. Beneficially, this potentially ensures that the SIR. model is effectively taking into account the dynamic changes in the infection rate of the pathogen over time, for highly accurately and reliably generating the subsequent iteration of the output data.

[0050] It will be appreciated that the initial estimates of the one or more parameters and the recently-acquired portion of the input data are further used for fitting the coefficients of the Fourier series to the estimated time-varying beta values. In this regard, a tool for minimizing a sum of squares of nonlinear functions (for example, such as a scipy. optimize. Ieast_squares function) is utilised to calculate / determine a set of parameters that minimise a residual (i.e., difference) between the output data generated by the SIR model and ground-truth data (namely, an actual data). Furthermore, according to the calculated residual, the one or more parameters are adjusted to minimize a difference between the output data generated by the SIR model and the ground-truth data. Beneficially, this may allow the one or more parameters to be adjusted in a manner that facilitate an optimum fitting of the aforesaid coefficients to capture the dynamic changes in the infection rate. Optionally, the system is configured to maintain beta values constant during the training period, for initiating a series of computational results at a start of a forecast period, aligning with a transition point in a computation of the mathematical simulation model. In this regard, maintaining the beta values constant during the training period may be necessary to establish a baseline understanding of a dynamics of the pathogen, and to ensure a stability of the SIR. model during its initial learning phase. When the beta values are constant, the SIR model could focus on learning underlying patterns and dynamics of the spread of the pathogen, without being influenced by fluctuations in infection rates that may occur due to various factors such as interventions, seasonal effects, or changes in population behaviour. This may allow the SIR model to accurately determine relationships between different compartments of the SIR model (namely, susceptible, infectious, and recovered) before attempting to forecast dynamic changes in the infection rate. Furthermore, aligning with the transition point is required to ensure that the SIR model is initialised with accurate and up-to-date information for forecasting. The transition point may represent a point at which the SIR model shifts from using the historical data part of the input for its training to using the recently-acquired part of the input data for generating subsequent forecast. By aligning the start of the forecast period with the transition point, the SIR model provides more accurate and consistent forecasts.

[0051] Optionally, the system is configured to execute the SIR model with found static beta values and forecasted beta values, starting at a start date of the training period and ending at a final date of the forecast period provided in the output data. In this regard, the system executes the SIR model, which simulates the spread of the disease through the population over time, using the found static beta values and the forecasted beta values. The found static beta values were determined during the training period. In addition to the static beta values, the system also uses the forecasted beta values which are generated based on the fitted Fourier series to capture the dynamic changes in the infection rate. Furthermore, the execution of the SIR. model starts at the beginning of the training period and continues until the final date of the forecast period, which beneficially allows the SIR model to simulate the spread of the disease from the historical data into a future, based on the forecasted infection rate values (namely, beta values). By executing the SIR model with the found static beta values and the forecasted beta values, the system can potentially generate predictions for the spread of the disease in the future. This enhances the SIR model's reliability and usefulness for informing public health interventions and decision-making processes.

[0052] Optionally, the SIR model is configured to provide the forecast indicative of changes in epidemiological burden as a result of using one or more vaccines on at least a portion of the population, wherein the system is configured to: receive a choice of country and a date range of forecast for training the SIR model; load the input data from a data storage arrangement, wherein the loaded input data is selected for training the SIR model; choose initial estimates of parameters indictive of the population, the pathogen, the one or more vaccines and spatial distribution of the pathogen; process reported data by applying at least one algorithm using a least squares computation to generate a model that described predicted and reported data regarding an effect of the pathogen on the population; apply a Fourier series computation to the model to represent the model via Fourier parameters; and use the Fourier parameters in the SIR. model for use in the forecast indicative of the changes in the epidemiological burden.

[0053] In this regard, the term "epidemiological burden" refers to an impact and consequences of the pathogen on the population. The epidemiological burden encompasses various aspects such as disease transmission rates, morbidity (namely, illness), mortality, healthcare resource utilisation, and societal disruptions caused by the pathogen. The epidemiological burden quantifies an extent of the pathogen's impact on individuals, communities, and healthcare systems, thereby providing insights for public health planning and intervention strategies at a time of a pandemic. For example, in a case of a pandemic like COVID-19, the epidemiological burden may comprise measures such as a number of confirmed cases, hospitalizations, deaths, and a strain on healthcare infrastructure. It will be appreciated that understanding and forecasting changes in the epidemiological burden is essential for informing vaccination strategies, implementing control measures, and mitigating the spread of the pathogen within the population.

[0054] The system utilises the SIR model to predict the impact of vaccine deployment on the epidemiological burden. For instance, the SIR model may forecast changes in disease transmission rates, hospitalization cases, or mortality rates following vaccination campaigns. By simulating different vaccination scenarios, the system aids in strategic decisionmaking for public health interventions, such as prioritizing vaccine distribution or adjusting vaccination strategies, based on population needs and disease dynamics.

[0055] The system allows users to specify the country and a forecast period for training the SIR model. By selecting a specific country, the users can tailor the forecasts to local epidemiological conditions and healthcare infrastructure. Additionally, choosing the date range for training enables the SIR model to learn from the initial iteration of the output data and capture temporal variations in disease dynamics, enhancing an accuracy of the subsequent iteration of the output data.

[0056] Further, the system retrieves the input data from the data storage arrangement to train the SIR. model. Optionally, the data storage arrangement comprises at least a data repository. The data repository could be implemented as a memory of the computing hardware, a cloudbased storage, or the like. It will be appreciated that when loading the input data, initial values for each parameter and each type of population (namely, compartment) are also provided to the SIR model for computation purposes. The system automatically selects the initial estimates of the parameters which may include population demographics, disease transmission rates, vaccine efficacy, geographical factors influencing disease spread, and the like. By initializing the SIR model with reasonable estimates, the system expedites a training process of the SIR model, and the accuracy of predictions is improved. The reported data can be understood to be ground-truth data related to the effect of the pathogen on the population. The reported data may include information such as a number of confirmed cases, hospitalizations, deaths, and other relevant epidemiological indicators. By applying the least squares algorithm, the system generates the (predictive) model that describes a relationship between the reported data and an effect of the pathogen on the population. The model captures key characteristics of the disease's spread, such as transmission dynamics, morbidity, and mortality rates, enabling predictions of future trends and outcomes. The generated model provides insights into an expected impact of the pathogen on the population, based on the input data and the reported data. In other words, the generated model describes how the pathogen has affected the population in the past and allows for an extrapolation of a trends to forecast future scenarios. The Fourier series computation decomposes the model into sinusoidal components, allowing for characterization of periodic patterns and temporal variations in the effect of the pathogen. This may enhance the model's ability to capture complex temporal trends and fluctuations in infection rates. The Fourier parameters are used by the SIR. model to facilitate forecasting of changes in the epidemiological burden and improving an accuracy of forecasts and enables proactive planning for public health interventions. For example, the model may accurately predict a timing and intensity of epidemics, thereby guiding vaccination campaigns and healthcare resource allocation. The aforesaid processing steps may be repeated a minimum of 50 times, with a random variation applied to the initial parameters values and every tenth run the ending date of the training period is shifted a day forward, to a achieve a range of predicted results, from which a median output is taken and used for forecasted output and the prediction intervals are calculated.

[0057] It will be appreciated that the SIR model may rely on a comprehensive dataset comprising various epidemiological metrics, which can be reported either on a daily or weekly basis. These metrics may include new cases, new hospitalization admissions, current hospital patients, and new deaths. Data collection may be facilitated through coded scripts accessing national and subnational sources, each with slightly differing data structures. Upon ingestion into a database, the input data may undergo rigorous cleaning and processing procedures to ensure accuracy and consistency. This may include removing duplicate, irrelevant, or erroneous data points, as determined by cross-checking and validation processes. Additionally, data normalization may also ensure uniformity across countries, with standardized formats, column names, and labels for direct comparability. Subsequently, the data may be normalized to consistent units, such as 'new persons per day'. Before training the SIR model, the data may be transformed into daily or weekly intervals, as appropriate, aligned with the time points of interest for the SIR model. Daily data may undergo smoothing using a 7-day running average to enhance precision by mitigating noise. In instances where reporting of metrics ceases, data points beyond the last reported date are replaced with zeros to maintain continuity. Furthermore, a data point at which the reported data ends, is retained and is used to exclude any zeros value from the training process of the SIR. model. Further data manipulation may involve aggregating new cases, hospitalizations, and deaths to derive cumulative metrics. Additionally, the SIR model may estimate daily hospital discharges when data on new hospitalizations, current patients, and new deaths are available, employing a simple formula. Although, a number of daily discharges is not typically a reported metric, it aids in model fitting. While inclusion of all metrics is ideal, it is to be noted that not all are mandatory for the SIR model's operation. An only essential requirement may be an initial value for cases; without this, the SIR model may not simulate new infections. Similarly, an absence of patient data, though utilized for defining an initial number of hospital patients, may not impede the SIR model functionality as it adjusts accordingly.

[0058] The present disclosure also relates to the method as described above. Various embodiments and variants disclosed above, with respect to the aforementioned system, apply mutatis mutandis to the method.

[0059] Optionally, the SIR model is configured to provide the forecast indicative of changes in epidemiological burden as a result of using one or more vaccines on at least a portion of the population, and wherein the method further includes: configuring the system to receive a choice of country and a date range of forecast for training the SIR model; configuring the system to load input data from a data storage arrangement, wherein the loaded input data is selected for training the SIR model; configuring the system to choose initial estimates of parameters indictive of the population, the pathogen, the one or more vaccines and spatial distribution of the pathogen; configuring the system to process reported data by applying at least one algorithm using a least squares computation to generate a model that described predicted and reported data regarding an effect of the pathogen on the population; configuring the system to apply a Fourier series computation to the model to represent the model via Fourier parameters; and configuring the system to use the Fourier parameters in the SIR. model for use in the forecast indicative of the changes in the epidemiological burden.

[0060] Optionally, the method further includes configuring the system to use the forecast to apply one or more vaccines to the population for mitigating the effect of the pathogen on the population.

[0061] The present disclosure also relates to the software product as described above. Various embodiments and variants disclosed above, with respect to the aforementioned first aspect and aforementioned second aspect, apply mutatis mutandis to the software product.

[0062] DETAILED DESCRIPTION OF THE DRAWINGS

[0063] Referring to FIG. 1, illustrated is a block diagram of a system 100 for forecasting an effect of a pathogen on a population, in accordance with an embodiment of the present disclosure. The system 100 comprises a computing hardware 102. The computing hardware 102 is configured to process input data, wherein when processing the input data, the computing hardware 102 is configured to apply a mathematical simulation model 104 to the input data 104 to generate output data. The mathematical simulation model 104 is configured to use a Susceptible- Infected-Recovered model (SIR. model).

[0064] It may be understood by a person skilled in the art that FIG. 1 includes a simplified architecture of the system 100, for sake of clarity, which should not unduly limit the scope of the claims herein. It is to be understood that the specific implementation of the system 100 is provided as an example and is not to be construed as limiting it to specific numbers or types of computing hardware. The person skilled in the art will recognize many variations, alternatives, and modifications of embodiments of the present disclosure.

[0065] Referring to FIG. 2, illustrated are steps of a method for operating a system for forecasting an effect of a pathogen on a population, in accordance with an embodiment of the present disclosure. At step 202, a computing hardware is configured to process input data defining parameters describing a population of people and a pathogen affecting or potentially affecting the population, wherein the computing hardware is configured to apply a mathematical simulation model to the input data to generate output data providing a forecast that is indicative of an effect of the pathogen on the population, wherein the mathematical simulation is configured to use at least a Susceptible-Infected-Recovered model (SIR model). At step 204, the system is configured initially to apply the SIR model to an historical part of the input data to generate an initial iteration of the output data. At step 206, the SIR model is trained during a training period on a recently-acquired part of the input data to improve the accuracy of the SIR model when generating a subsequent iteration of the output data.

[0066] The aforementioned steps are only illustrative and other alternatives can also be provided where one or more steps are added, one or more steps are removed, or one or more steps are provided in a different sequence without departing from the scope of the claims herein. Referring to FIG. 3, illustrated is an architecture of a Susceptible- Infected-Recovered model (SIR) model 300, in accordance with an embodiment of the present disclosure. With reference to FIG. 3, the SIR model 300 describes a total population being divided into various groups of population, namely, a susceptible population (depicted as S), an exposed population (depicted as E), an infectious population (depicted as I), a hospitalised population (depicted as H), a recovered population (depicted as R), a discharged population (L), and a deceased population (depicted as D). Additionally, the SIR model 300 includes a group for cumulative infections (depicted as 12) which track a total number of infections over time, and a group of cumulative hospitalisations (depicted as H2) which track a total number of hospitalizations over time. A flow within the SIR model 300 is shown using arrows. Moreover, a specific dynamic of the SIR model 300 is determined by initial conditions of the groups of population and a rate at which individuals transition between said groups. For example, an arrow from S to E is indicative of individuals moving from the susceptible population to the exposed population. An arrow from E to I is indicative of some individuals moving from the exposed population to the infectious population. An arrow from E to R is indicative of some people moving from the exposed population to the recovered population. An arrow from I to H is indicative of some individuals moving from the infectious population to the hospitalised population. An arrow from I to R is indicative of some individuals moving from the infectious population to the recovered population. An arrow from H to R is indicative of some individuals moving from the hospitalised population to the recovered population. An arrow from H to D is indicative of some individuals moving from the hospitalised population to the deceased population. An arrow from H to L is indicative of some individuals moving from the hospitalised population to the discharged population (comprising patients leaving hospitals). An arrow from R to S is indicative of some individuals moving from the recovered population to the susceptible population. An arrow from E to 12 is indicative of some individuals moving from the exposed population to the group for cumulative infections to the deceased population. An arrow from I to H2 is indicative of some individuals moving from the infectious population to the group of cumulative hospitalisations. Moreover, the SIR. model 300 also utilise various parameters describing a population of individuals and a pathogen affecting or potentially affecting the population. For example, epsilon (depicted as 6) represents a rate at which individuals become exposed to a pathogen, beta (depicted as 0) represents a rate at which the susceptible population become infected, gamma (depicted as y) represents a rate at which the infected individuals recover, eta (depicted as ) represents a rate at which infected individuals require hospitalization, rho (depicted as p) represents a rate at which individuals lose immunity and become susceptible again, and delta (depicted as 5) represents a rate at which infected individuals die from a disease. It is to be noted that the rate of recovery (i.e., the gamma) could be different for different individuals, depending on whether a given individual is recovering from any one of: an incubation group (E), an infectious group (I), a hospitalised group (H).

[0067] Referring to FIG. 4, illustrated is an exemplary scenario of processing input data 402 by utilising a Susceptible-Infected-Recovered (SIR) model 404, in accordance with an embodiment of the present disclosure. With reference to FIG. 4, the SIR model 404 is initially applied to an historical part of the input data 402 to generate an initial iteration of output data 406, and then the SIR model 404 is trained on a recently-acquired part of the input data 402, to improve an accuracy of the SIR model 404 when generating a subsequent iteration of the output data 406. The input data 402 defines parameters describing a population of people and a pathogen affecting or potentially affecting the population. The output data 406 provides a forecast that is indicative of an effect of the pathogen on the population. Referring to FIG. 5, illustrated is an exemplary graphical representation of tracking, forecasting and simulating infectious disease outcomes for short-term future and medium-term future, in accordance with an embodiment of the present disclosure. With reference to FIG. 5, an X- axis represents time (for example, in days), and a Y-axis represents a stratified population of interest (for example, infectious cases, absentee days, hospitalisation cases, mortality cases, and the like). A region 502 of the graphical representation corresponds to forecast corresponding to an historical part of input data for a Susceptible-Infected-Recovered model (SIR model), a region 504 of the graphical representation corresponds to forecast corresponding to an recently-acquired part of the input data, and a region 506 of the graphical representation corresponds to forecast corresponding to future times when the SIR model has been trained.

[0068] FIG. 3, 4, and 5 are merely examples, which should not unduly limit the scope of the claims herein. The person skilled in the art will recognize many variations, alternatives, and modifications of embodiments of the present disclosure.

[0069] FIGs. 6, 7, 8A-8D, and 9A-9H are discussed hereinbelow in the experimental part section.

[0070] EXPERIMENTAL PART

[0071] Hereinbelow, there are provided some graphical representations pertaining to a forecast conducted in Denmark when a system for forecasting an effect of a pathogen on a population is employed.

[0072] Referring to FIG. 6, illustrated is a comparison between a known forecast and a fitted forecast provided by implementing a system of the present disclosure. Herein, a forecast for a total number of infections over a time period of 12 months is graphically illustrated. An X-axis represents time, and a Y-axis represents a number of infected cases reported for each month. Moreover, the known forecast is represented using a dotted line, depicting a forecasted trend of a number of infected cases recorded for each month. The known forecast is provided by a Susceptible-Infected- Recovered model (SIR model) when infection rate is same. The fitted forecast is represented using a solid line, depicting a forecasted trend of a number of infected cases when dynamic changes in an infection rate are covered, by generating a Fourier series of coefficients that are fitted to estimated time-varying beta values generated by the SIM model. A part of the trend of the known forecast and of the fitted forecast to a left- hand-side of a dashed line 602 corresponds to a constant infection rate, while a remaining part of the trend of the known forecast and of the fitted forecast to a right-hand-side of the dashed line 602 corresponds to a dynamic infection rate.

[0073] Referring to FIG. 7, illustrated is a comparison between a forecast (for example, depicted using a dashed line curve) of a parameter by implementing a system of the present disclosure and an actual reported trend (for example, depicted using a solid line curve) of the (same) parameter. Herein, the forecast for a number of hospitalised cases over time is graphically illustrated. An X-axis represents time, while a Y-axis represents a number of hospitalised cases with respect to time.

[0074] With reference to FIGs. 8A and 8C, illustrated are two different graphical representations of an initial version of a forecast before performing a fitting process. Herein, an X-axis represents time, while a Y-axis represents an infection rate. A dotted line in the graphical representations represents a transition point at time which a third-order Fourier series is applied to capture a dynamic change in an infection rate. With reference to FIGs. 8B and 8D, illustrated are corresponding graphical representations of an aligned version of the forecast upon performing the fitting process in FIGs. 8A and 8C, respectively. Herein, an X-axis represents time, while a Y-axis represents an infection rate. Referring to FIGs. 9A-9H, illustrated are various graphical representations of comparisons between forecasts (for example, depicted using dashed line curves) of different parameters and actual report data (for example, depicted using solid line curves) of said different parameters, wherein the forecasts of the different parameters are provided by a system for forecasting an effect of a pathogen on a population according to a first aspect.

[0075] With reference to FIG. 9A, a comparison between a forecast pertaining to cumulative infectious cases and actual reported data for the cumulative infectious cases with respect to time, is graphically illustrated, wherein an X axis represents time, and a Y axis represents the cumulative infectious cases.

[0076] With reference to FIG. 9B, a comparison between a forecast pertaining to cumulative hospitalisation cases and actual reported data for the cumulative hospitalisation cases with respect to time, is graphically illustrated, wherein an X axis represents time, and a Y axis represents the cumulative hospitalisation cases.

[0077] With reference to FIG. 9C, a comparison between a forecast pertaining to cumulative death cases and actual reported data for the cumulative death cases with respect to time, is graphically illustrated, wherein an X axis represents time, and a Y axis represents the cumulative death cases.

[0078] With reference to FIG. 9D, a forecast pertaining to an infection rate (namely, beta values) with respect to time is graphically illustrated, wherein an X axis represents time, and a Y axis represents the infection rate.

[0079] With reference to FIG. 9E, a comparison between a forecast pertaining to newly infected cases and actual reported data for the newly infected cases with respect to time, is graphically illustrated, wherein an X axis represents time, and a Y axis represents a number of the newly infected cases.

[0080] With reference to FIG. 9F, a comparison between a forecast pertaining to new hospitalisation cases and actual reported data for the new hospitalisation cases with respect to time, is graphically illustrated, wherein an X axis represents time, and a Y axis represents a number of the new hospitalisation cases.

[0081] With reference to FIG. 9G, a comparison between a forecast pertaining to new death cases and actual reported data for the new death cases with respect to time, is graphically illustrated, wherein an X axis represents time, and a Y axis represents a number of the new death cases.

[0082] With reference to FIG. 9H, a comparison between a forecast pertaining to hospitalised and previously-hospitalised patients and actual reported data for the hospitalised and previously-hospitalised patients with respect to time, is graphically illustrated, wherein an X axis represents time, and a Y axis represents a number of the hospitalised and previously- hospitalised patients.

Claims

CLAIMSWhat is claimed is:

1. A system including a computing hardware for processing input data defining parameters describing a population of people and a pathogen affecting or potentially affecting the population, wherein the computing hardware is configured to apply a mathematical simulation model to the input data to generate output data providing a forecast that is indicative of an effect of the pathogen on the population, wherein the mathematical simulation model is configured to use at least a Susceptible-Infected- Recovered model (SIR model), characterized in that the system is configured initially to apply the SIR model to an historical part of the input data to generate an initial iteration of the output data, and then to train the SIR model during a training period on a recently- acquired part of the input data to improve an accuracy of the SIR model when generating a subsequent iteration of the output data.

2. A system of claim 1, wherein the system is configured to use the forecast to apply one or more vaccines to the population for mitigating the effect of the pathogen on the population.

3. A system of claim 1 or 2, wherein the system is configured to simulate a spread of a disease arising from the pathogen, wherein an infection rate of the pathogen is allowed to be temporally variable in magnitude.

4. A system of any one of the preceding claims, wherein, during the training period, the SIR model is trained on at least one of:(i) a number and characteristics of the people designated as being patients;(ii) an occurrence of new people being affected by the pathogen;(iii) a record of people afflicted by the pathogen being admitted to one or more hospital establishments;(iv) a record of people who have died as a result of the effect of the pathogen.

5. A system of any one of the preceding claims, wherein the system is configured to receive initial estimates of one or more parameters entered by a user of the system, wherein the one or more parameters are utilised to determine at least temporal and coupling characteristics of differential and / or integral computations performed when executing the SIR. model, optionally, wherein the initial estimates relate to parameters: beta (P), epsilon (6), gamma (y), eta (n) , rho (p), delta (5).

6. A system of any one of the preceding claims, wherein the mathematical simulation model is configured to generate a Fourier series of coefficients that are fitted to estimated time-varying beta values generated by the SIM model, to capture dynamic changes in an infection rate of the pathogen.

7. A system of claim 6, wherein the system is configured to maintain beta values constant during the training period, for initiating a series of computational results at a start of a forecast period, aligning with a transition point in a computation of the mathematical simulation model.

8. A system of claim 6 or 7, wherein the system is configured to execute the SIR model with found static beta values and forecasted beta values, starting at a start date of the training period and ending at a final date of the forecast period provided in the output data.

9. A system of any one of the preceding claims, wherein the input data is acquired, at least in part, by using sensor arrangements to monitor the population of people.

10. A system of any one of the preceding claims, wherein the SIR. model is configured to provide the forecast indicative of changes in epidemiological burden as a result of using one or more vaccines on at least a portion of the population, wherein the system is configured to: receive a choice of country and a date range of forecast for training the SIR model; load the input data from a data storage arrangement, wherein the loaded input data is selected for training the SIR model; choose initial estimates of parameters indictive of the population, the pathogen, the one or more vaccines and spatial distribution of the pathogen; process reported data by applying at least one algorithm using a least squares computation to generate a model that described predicted and reported data regarding an effect of the pathogen on the population; apply a Fourier series computation to the model to represent the model via Fourier parameters; and use the Fourier parameters in the SIR model for use in the forecast indicative of the changes in the epidemiological burden.

11. A method for operating a system of any one of the preceding claims, wherein the method includes: configuring a computing hardware to processing input data defining parameters describing a population of people and a pathogen affecting or potentially affecting the population, wherein the method includesconfiguring the computing hardware to apply a mathematical simulation model to the input data to generate output data providing a forecast that is indicative of an effect of the pathogen on the population, wherein the mathematical simulation model is configured to use at least a Susceptible-Infected-Recovered model (SIR. model), characterized in that method further includes: configuring the system initially to apply the SIR model to an historical part of the input data to generate an initial iteration of the output data, and then training the SIR model during a training period on a recently- acquired part of the input data to improve an accuracy of the SIR model when generating a subsequent iteration of the output data.

12. A method of claim 11, wherein the SIR model is configured to provide the forecast indicative of changes in epidemiological burden as a result of using one or more vaccines on at least a portion of the population, and wherein the method further includes: configuring the system to receive a choice of country and a date range of forecast for training the SIR model; configuring the system to load input data from a data storage arrangement, wherein the loaded input data is selected for training the SIR model; configuring the system to choose initial estimates of parameters indictive of the population, the pathogen, the one or more vaccines and spatial distribution of the pathogen; configuring the system to process reported data by applying at least one algorithm using a least squares computation to generate amodel that described predicted and reported data regarding an effect of the pathogen on the population; configuring the system to apply a Fourier series computation to the model to represent the model via Fourier parameters; and configuring the system to use the Fourier parameters in the SIR. model for use in the forecast indicative of the changes in the epidemiological burden.

13. A method of claim 12, wherein the method further includes configuring the system to use the forecast to apply one or more vaccines to the population for mitigating the effect of the pathogen on the population.

14. A software product that is executable on a computing hardware of a system of any one of claims 1 to 10, to cause the system to implement the method of claim 11, 12 or 13.

Images