Intelligent analysis method and system for epidemiology of zoonosis

Abstract

The invention belongs to the field of disease prediction, and provides a zoonosis epidemiology intelligent analysis method and system.The method comprises the steps that in a preset time period, samples are collected, pathogen detection is conducted, and an infection data sequence is obtained; determining the change rate of the number of people-livestock cross-infected people according to the infection data sequence; determining the human-animal contact rate as a hidden state, and constructing a hidden Markov model; training the hidden Markov model by taking the change rate of the number of people-livestock cross infection as an observation sequence, and determining state distribution, a state transition probability matrix and observation probability distribution parameters of the hidden Markov model; deducing the current human-animal contact rate according to the current human-animal cross infection number change rate and the trained hidden Markov model; and establishing a propagation model according to the human-animal contact rate, and predicting infection data according to the propagation model.

Description

Technical Field

[0001] The present invention belongs to the field of disease prediction, and particularly relates to an intelligent analysis method and system for zoonosis epidemiology. Background Art

[0002] Zoonosis refers to an infectious disease caused by pathogens and capable of natural transmission between humans and animals. It has a wide variety of types and complex transmission routes, posing a relatively high public health risk. Typical zoonoses include avian influenza, brucellosis, campylobacteriosis, rabies, and psittacosis, etc. With the large-scale development of the breeding industry and the expansion of the scope of human activities, scientific modeling of the zoonosis transmission process and prediction of the infection trend have become important technical ways to prevent and control major population health risks.

[0003] Currently, most mainstream epidemiological analysis methods are based on SEIR-like kinetic models and rely on parameters such as the transmission rate to simulate and predict the development trend of the epidemic. In such models, the human-animal contact rate is a key variable affecting the transmission efficiency between humans and animals, and is usually used as a core component of the transmission rate to describe the frequency of effective transmissible contacts between human individuals and animal individuals per unit time.

[0004] However, in practical applications, the change of the human-animal contact rate shows high volatility and uncertainty, and is significantly affected by various factors. For example, seasonal activities, the trading frequency of farmers' markets, environmental control policies, and human behavior patterns, etc. will all cause a significant change in the human-animal contact intensity in a short time. In addition, due to the lack of systematic contact monitoring means, the human-animal contact rate cannot be obtained through direct observation or real-time measurement, and needs to rely on indirect data for estimation.

[0005] Traditional methods often regard the human-animal contact rate as a fixed value or use empirical estimation, which fails to fully reflect the dynamic change law of the contact rate, resulting in deviation in the calculation of the transmission rate, and further affecting the accurate prediction of the change trend of the number of infected people by the model, reducing the credibility and practicality of the epidemiological analysis results. Summary of the Invention

[0006] In order to solve the problems in the prior art, the present invention provides an intelligent analysis method for zoonosis epidemiology, including the following steps:

[0007] Collect samples within a preset time period and perform pathogen detection to obtain an infection data sequence;

[0008] Determine the change rate of the number of human-animal cross-infected people according to the infection data sequence;

[0009] Determine the human-animal contact rate as a hidden state and construct a hidden Markov model;

[0010] Training the hidden Markov model with the change rate of the number of human-animal cross-infection as the observation sequence to determine the state distribution, state transition probability matrix and observation probability distribution parameters of the hidden Markov model;

[0011] Inferring the current human-animal contact rate according to the current change rate of the number of human-animal cross-infection and the trained hidden Markov model;

[0012] Establishing a transmission model according to the human-animal contact rate and predicting the infection data according to the transmission model.

[0013] On the other hand, the present invention also provides an intelligent analysis system for zoonosis epidemiology, including the following modules:

[0014] A data acquisition module, configured to collect samples and perform pathogen detection within a preset time period to obtain an infection data sequence;

[0015] A calculation module, configured to determine the change rate of the number of human-animal cross-infection according to the infection data sequence;

[0016] A modeling module, configured to determine the human-animal contact rate as a hidden state and construct a hidden Markov model;

[0017] A training module, configured to train the hidden Markov model with the change rate of the number of human-animal cross-infection as the observation sequence to determine the state distribution, state transition probability matrix and observation probability distribution parameters of the hidden Markov model;

[0018] An inference module, configured to infer the current human-animal contact rate according to the current change rate of the number of human-animal cross-infection and the trained hidden Markov model;

[0019] By constructing a hidden Markov model with the change rate of the number of human-animal cross-infection as the observation sequence, the present invention converts the traditional unobservable human-animal contact rate into an inferable hidden state, solves the problem that the human-animal contact rate cannot be monitored in real time, and realizes the quantification and dynamic estimation of the change of the contact rate. Using the dynamically updated human-animal contact rate parameters, the human-animal transmission rate is adjusted in real time, which effectively reflects the impact of contact behavior changes on the epidemic transmission, significantly improves the fitting degree of the transmission model to the infection data and the prediction ability of the trend, and overcomes the problem of inaccurate prediction of the fixed transmission rate model. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0021] Figure 1 is the flowchart of the method of the present invention;

[0022] Figure 2 is the system block diagram of the present invention. Detailed implementation manners

[0023] Next, with reference to the accompanying drawings and specific implementation manners, a preferred description of the invention will be given.

[0024] This embodiment solves the above problems through the following steps:

[0025] In one embodiment, referring to Figure 1 , the present invention provides an intelligent analysis method for zoonotic disease epidemiology, which realizes quantitative analysis of the transmission association between humans and animals through dynamic modeling and parameter estimation, and an intelligent analysis method for predicting the transmission trend of zoonotic diseases based on multi-source data fusion and intelligent algorithms, and is applicable to a technical solution for real-time assessment and early warning of the transmission path, infection risk and epidemic situation of various zoonotic diseases.

[0026] In the present invention, zoonotic diseases refer to infectious diseases caused by pathogens such as bacteria, viruses, fungi, parasites, etc., which can be naturally transmitted between humans and vertebrates, including but not limited to avian influenza, brucellosis, campylobacteriosis, psittacosis, etc. This term specifically refers to the disease type with a dual-host transmission chain between humans and animals, and is the core scope of the analysis object of the present invention.

[0027] Epidemiology refers to the scientific methods and theoretical systems for studying the occurrence, distribution and determinants of diseases or health events in specific populations. Epidemiology focuses on the transmission mechanism, influencing factors and control strategies of infectious diseases, and is the theoretical basis for realizing the monitoring, modeling and prediction of the transmission process of zoonotic diseases. The present invention establishes epidemiology through subsequent steps to perform intelligent analysis on infection data.

[0028] Step S10, within a preset time period, collect samples and perform pathogen detection to obtain an infection data sequence.

[0029] Within a preset time period, that is, the detection time period set according to the epidemiological surveillance plan, such as daily, weekly or monthly, epidemiological samples are collected from human individuals, animal individuals and related environmental media within the target area according to standardized sampling specifications. The target area may include farms, slaughterhouses, farmers' markets, wild animal habitats and surrounding residential areas. The human samples include but are not limited to human respiratory swabs, venous blood, and fecal samples. The animal samples include feathers, feces, nasopharyngeal swabs and whole blood samples of poultry (such as chickens, ducks, pigeons) and livestock (such as pigs, cows, sheep, dogs). The environmental samples include airborne dust particles, water bodies, animal fecal contamination contact surfaces, feed residues and cage surface smear samples in the animal breeding environment, etc.

[0030] The collected samples are preliminarily processed on site, including sub-packaging, labeling, cold-chain transportation, and then sent to the laboratory for sample numbering and information entry according to the established procedures. High-sensitivity pathogen detection methods are used to detect pathogens in the samples, specifically including fluorescence quantitative polymerase chain reaction (qPCR) to detect the nucleic acid sequences of specific pathogens, or metagenomic sequencing technology for non-targeted pathogen detection, or a combination of ELISA, immunochromatography and other means to quantitatively detect the antigens or antibodies of pathogens. The pathogens may include but are not limited to common zoonotic pathogens such as Chlamydia psittaci, avian influenza virus, Brucella, and Campylobacter.

[0031] During the detection process, the detection time, sample source, host category, sampling location and detection results of each sample are recorded, including indicators such as positive or negative determination, pathogen load (such as Ct value, copy number), etc. The above information is sorted and encoded in chronological order to form a sequence of infection data arranged in a time series, which is used to describe the pathogen infection dynamics of a specific area or group within the preset time period. For example, if 50 human samples and 100 animal samples are detected in a certain area within four consecutive weeks, and 5, 8, 12, 10 positive human samples and 10, 15, 20, 18 positive animal samples are found respectively, a cross-time-node human-animal infection quantity change sequence can be formed. This infection data sequence is used as an input to the transmission model to support subsequent prediction of transmission trends and dynamic estimation of human-animal transmission parameters.

[0032] Step S20, determine the change rate of the number of human-animal cross-infections according to the infection data sequence.

[0033] Classify and screen the infection data sequence obtained in step S10. According to the sample source, the data is divided into two categories: human infection data and animal infection data. The human infection data includes the number of human individuals in the state of pathogen infection at each detection time point, and the animal infection data includes the number of animal individuals in the infected state at the corresponding time point. Further, based on the sample collection time and spatial location, determine the time alignment relationship between human and animal samples to ensure the comparability and synchronization of the data.

[0034] Compare the human infection data and animal infection data at consecutive time points. In the areas with close human-animal contact, identify and label human-animal cross-infection events. The cross-infection event is a situation where the number of infected animals in the same geographical area also changes significantly before or after the time node when the number of human infections changes significantly. A set threshold is used for determination, and the threshold can be set as the relative growth rate or absolute growth value of the change in the number of infected people according to the sample size.

[0035] For the identified human-animal cross-infection events, calculate the joint change rate of the number of human infections and the number of animal infections in each detection cycle, specifically:

[0036] Let the number of human infections at time t be I h (t), and the number of animal infections be I a (t). Then the change rate of human-animal cross-infection number △I ha (t) is defined as:

[0037]

[0038] Where N h is the total number of sampled people, N a is the total number of sampled animals, and △I ha (t) represents the change in human-animal infection intensity per unit time, reflecting the dynamics of human-animal cross-infection.

[0039] Furthermore, to reduce the impact of data fluctuations on the calculation results, smooth the change rate using the moving average method or exponential weighting method to obtain a smoothed change rate sequence; and standardize the change rate data to eliminate the influence of different time points or different sample sizes on the results.

[0040] Step S30, determine the human-animal contact rate as the hidden state and construct a hidden Markov model.

[0041] In actual scenarios, human-animal contact behaviors are highly dispersed and diverse, involving multiple links such as breeding, transportation, and market transactions; there is a lack of technical means to accurately record each human-animal contact event, resulting in the difficulty of quantifying the contact frequency. At the same time, under the influence of different times, locations, and policies, the human-animal contact frequency shows non-linear fluctuations; holidays, breeding cycles, and market transaction activities can all cause significant changes and cannot be represented by a fixed value.

[0042] Therefore, according to the human-animal contact characteristics in the transmission process of zoonotic diseases, the human-animal contact rate is set as the hidden state of the hidden Markov model. The human-animal contact rate represents the average effective contact frequency between human individuals and animal individuals per unit time and cannot be obtained through direct observation.

[0043] To realize model construction, the following hidden state set is set:

[0044] S = {s1, s2, s3} = {L, M, H}

[0045] Where:

[0046] L represents the low contact rate state;

[0047] M represents the medium contact rate state;

[0048] H represents the high contact rate state.

[0049] Each state corresponds to a specific contact rate value or interval. For example:

[0050] C ha (t) = C0 × 0.5 corresponds to s1 = L;

[0051] C ha (t) = C0 × 1.0 corresponds to s2 = M

[0052] C ha (t) = C0 × 1.5 corresponds to S3 = H.

[0053] Where C ha (t) represents the human-animal contact rate at time t, and C0 is the basic contact rate constant.

[0054] The change rate of the number of human-animal cross-infections calculated based on the infection data sequence is used as the observation variable of the hidden Markov model to form an observation sequence:

[0055] O = {o1, o2, …, o T}

[0056] Where:

[0057] o t = △I ha(t) represents the change rate of the number of human-animal cross-infections at time t;

[0058] T is the total number of time periods.

[0059] Define the parameter set {S, O, π, A, B} of the hidden Markov model, specifically:

[0060] Initial state probability distribution:

[0061] π = {π1, π2, π3}

[0062] Where: π i = P(q i = s i ), q i represents the hidden state variable of the system at time i; s i represents the specific state at the i-th moment in the set of hidden states. For example, in the analysis of zoonotic disease transmission, s i can represent the contact intensity level of "low contact rate (L)", "medium contact rate (M)", or "high contact rate (H)"; P(q i = s i ) represents the probability that the system is in state s i at time i.

[0063] State transition probability matrix:

[0064] A = [a ij 3×3 , where, a ij = P(q t+1 = s j |q t = s i )

[0065] Observation probability distribution:

[0066] B = {B i (o t )}, where, B i (o t ) = P(o t |q t = s i )

[0067] Optionally, B i (o t ) selects the Gaussian distribution:

[0068]

[0069] Where, μ i , are the means of state s iExpectation and variance under

[0070] Step S40: Train the hidden Markov model with the change rate of the number of human-animal cross-infection as the observation sequence, and determine the state distribution, state transition probability matrix, and observation probability distribution parameters of the hidden Markov model.

[0071] Specifically, calculate the forward probability and backward probability given the observation sequence O under the current parameter estimation, respectively:

[0072] α t (i) = P(o1, o2,..., o t , q t = s i |λ)

[0073] β t (i) = P(o t+1 , o t+2 ,..., o T |q t = s i , λ)

[0074] Among them, α t (i) = P(o1, o2,..., o t , q t = s i |λ) represents the joint probability that, under the condition of the model given parameter λ, from time 1 to time t, the observation sequence o1, o2, L, o t is successively observed, and the system is in state s i at time t; λ represents the parameter set of the hidden Markov model λ = (π, A, B); β t (i) = P(o t+1 , o t+2 ,..., o T |q t = s i , λ) represents the backward probability, which represents the conditional probability that, under the condition of the given model parameter λ, from time t + 1 to the final time T, the observation value sequence o t+1 , o t+2 , L, o T is successively observed, and the system is in the hidden state s i at time t.

[0075] Use the above probabilities to update the parameters:

[0076] Initial state probability:

[0077]

[0078] State transition probability:

[0079]

[0080] Among them,

[0081]

[0082] Observation probability parameter (Gaussian distribution parameter):

[0083]

[0084] Iteratively train until the model parameters converge or reach the set number of iterations to obtain the optimally estimated π * , A * , B * .

[0085] Step S50, infer the current human-animal contact rate according to the current change rate of the number of human-animal cross-infections and the trained hidden Markov model.

[0086] Specifically, based on the current and historical observations o1, o2... o t , calculate the probability that the system is in each hidden state at time t. The formula is as follows:

[0087]

[0088] Where:

[0089] α t (i) represents the forward probability, indicating the probability that the system reaches state s i and observes the first t data.

[0090] β t (i) represents the backward probability, indicating the probability of observing the remaining observations starting from state s i .

[0091] Calculate the posterior probability γ t (i) of each state, and take the state s max corresponding to the maximum value as the most likely hidden state at the current moment, that is, the current human-animal contact rate level:

[0092]

[0093] According to the inferred hidden state s max , assign the corresponding contact rate value C ha (t), which is defined as follows:

[0094] If s max = L, then C ha (t) = C0 × 0.5;

[0095] If s max = M, then C ha (t) = C0 × 1.0;

[0096] If s max = H, then C ha (t) = C0 × 1.5.

[0097] Where C0 is the base contact rate constant, set according to historical statistics or literature.

[0098] Step S60: Establish a transmission model based on the human-animal contact rate, and predict the infection data according to the transmission model.

[0099] According to the transmission mechanism of zoonosis, construct a transmission dynamics model including human and animal dual-host populations, and use an extended SEIR (Susceptible-Exposed-Infected-Recovered) model to describe the within-population transmission and human-animal cross-transmission processes in the human and animal populations respectively.

[0100] Let the state variables in the model include:

[0101] Human population:

[0102] S h (t) represents the number of susceptible human individuals;

[0103] E h (t) represents the number of latent human individuals;;

[0104] I h (t) represents the number of infected human individuals;

[0105] R h (t) represents the number of recovered human individuals.

[0106] Animal population:

[0107] S a (t) represents the number of susceptible animal individuals;

[0108] E a (t) represents the number of latent animal individuals;

[0109] I a (t) the number of infected animals;

[0110] R a (t) represents the number of recovered animal individuals.

[0111] Let the key transmission parameters in the transmission model be:

[0112] Human-to-human transmission rate: β hh , a constant;

[0113] Animal - animal transmission rate: β aa , which is a constant;

[0114] Human - animal transmission rate: β ha (t), which is a dynamic parameter and is calculated based on the current human - animal contact rate C ha (t): β ha (t) = k × C ha (t)

[0115] Where:

[0116] k represents the transmission efficiency constant;

[0117] C ha (t) represents the human - animal contact rate inferred in step S50;

[0118] β ha (t) represents the human - animal transmission rate at time t and is used for dynamic updating of the model.

[0119] Construct the differential equations of transmission for the human population and the animal population, specifically:

[0120] Equation for the human population transmission:

[0121]

[0122] Equation for the animal population transmission:

[0123]

[0124] Where:

[0125] N h , N a represents the total number of the human and animal populations;

[0126] σ h , σ a represents the reciprocal of the incubation period, i.e., the incubation rate;

[0127] γ h , γ a represents the recovery rate, i.e., the proportion of infected individuals turning into recovered individuals per unit time.

[0128] Set the initial conditions of the transmission model at time t = t0:

[0129] S h (t0), E h (t0), I h (t0), R h (t0), S a (t0), E a (t0), I a (t0), Ra (t0)

[0130] The initial value can be set according to the previous sampling and detection data, or set using a standardized ratio.

[0131] Discretize and solve the aforementioned differential equations to predict the infection change trends of the human and animal populations in the next T time periods, and obtain a predicted infection data sequence:

[0132] Human infection prediction sequence:

[0133]

[0134] Animal infection prediction sequence:

[0135]

[0136] See Figure 2 , in another embodiment, the present invention further provides a zoonotic disease epidemiological intelligent analysis system, including:

[0137] A data acquisition module, configured to collect samples and perform pathogen detection within a preset time period to obtain an infection data sequence;

[0138] A calculation module, configured to determine the change rate of the number of human-animal cross-infections according to the infection data sequence;

[0139] A modeling module, configured to determine the human-animal contact rate as a hidden state and construct a hidden Markov model;

[0140] A training module, configured to train the hidden Markov model with the change rate of the number of human-animal cross-infections as the observation sequence to determine the state distribution, state transition probability matrix, and observation probability distribution parameters of the hidden Markov model;

[0141] An inference module, configured to infer the current human-animal contact rate according to the current change rate of the number of human-animal cross-infections and the trained hidden Markov model;

[0142] A prediction module, configured to establish a transmission model according to the human-animal contact rate and predict infection data according to the transmission model.

[0143] It should be noted that the explanations of the foregoing embodiments of the zoonotic disease epidemiological intelligent analysis method also apply to the device of the embodiments of the present application, and will not be repeated here.

[0144] Those of ordinary skill in the art can realize that the units and algorithm steps described in the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of both. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0145] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0146] In several embodiments provided in this application, if any function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The foregoing storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (hereinafter referred to as ROM), random access memories (hereinafter referred to as RAM), magnetic disks, or optical discs.

[0147] The above is only the specific implementation manner of this application. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all of them should be covered by the protection scope of this application. The protection scope of this application shall be subject to the protection scope of the claims. For the part of the module structure that is not specifically defined in this invention, it shall be based on the content recorded in the prior art. The prior art mentioned in the foregoing background art part and specific embodiment part of this invention can be used as a part of this invention to understand the meaning of some technical features or parameters.

Claims

1. An intelligent analysis method for zoonotic disease epidemiology, characterized in that, It includes the following steps: Collect samples within a preset time period, conduct pathogen detection, and obtain an infection data sequence; Determine the change rate of the number of human-animal cross-infections according to the infection data sequence; Determine the human-animal contact rate as a hidden state and construct a hidden Markov model; Train the hidden Markov model with the change rate of the number of human-animal cross-infections as the observation sequence, and determine the state distribution, state transition probability matrix, and observation probability distribution parameters of the hidden Markov model; Infer the current human-animal contact rate according to the current change rate of the number of human-animal cross-infections and the trained hidden Markov model; Establish a transmission model according to the human-animal contact rate, and predict the infection data according to the transmission model.

2. The intelligent analysis method for zoonotic disease epidemiology according to claim 1, characterized in that: The definition of the change rate of the number of human-animal cross-infections is related to the number of human infections at time t, the number of animal infections at time t, the total number of sampled people, the total number of sampled animals, and the change in the human-animal infection intensity per unit time.

3. The intelligent analysis method for zoonotic disease epidemiology according to claim 2, characterized in that: The construction of the hidden Markov model includes: Set a set of hidden states; Use the change rate of the number of human-animal cross-infections calculated according to the infection data sequence as the observation variable of the hidden Markov model to form an observation sequence; Define the parameter set of the hidden Markov model.

4. The zoonosis epidemiological intelligent analysis method according to claim 3, characterized in that Training the hidden Markov model with the change rate of the number of human-animal cross-infections as the observation sequence to determine the state distribution, state transition probability matrix, and observation probability distribution parameters of the hidden Markov model specifically includes: Calculate the forward probability and backward probability of the given observation sequence O under the current parameter estimation; Update the parameters using the above probabilities: Initial state probability; Iteratively train until the model parameters converge or reach the set number of iterations to obtain the optimal estimate.

5. The zoonosis epidemiological intelligent analysis method according to claim 4, wherein Inferring the current human-animal contact rate according to the current change rate of the number of human-animal cross-infections and the trained hidden Markov model includes: Based on the current and historical observation values, calculate the probability that the system is in each hidden state at time t; Calculate the posterior probability γ for each state t (i), take the state s corresponding to the maximum value max as the most likely hidden state at the current moment, that is, determine the current human-animal contact rate level.

6. An intelligent analysis system for zoonotic disease epidemiology, characterized in that, It includes the following modules: A data acquisition module for collecting samples within a preset time period, conducting pathogen detection, and obtaining an infection data sequence; A calculation module for determining the change rate of the number of human-animal cross-infections according to the infection data sequence; A modeling module for determining the human-animal contact rate as a hidden state and constructing a hidden Markov model; A training module for training the hidden Markov model with the change rate of the number of human-animal cross-infections as the observation sequence, and determining the state distribution, state transition probability matrix, and observation probability distribution parameters of the hidden Markov model; An inference module for inferring the current human-animal contact rate according to the current change rate of the number of human-animal cross-infections and the trained hidden Markov model; A prediction module for establishing a transmission model according to the human-animal contact rate and predicting the infection data according to the transmission model.

7. The zoonosis epidemiological intelligent analysis system according to claim 6, wherein, The definition of the change rate of the number of human-animal cross-infections is related to the number of human infections at time t, the number of animal infections at time t, the total number of sampled people, the total number of sampled animals, and the change in the human-animal infection intensity per unit time.

8. The zoonosis epidemiology intelligent analysis system according to claim 7, characterized in that, The construction of the hidden Markov model includes: Set a set of hidden states; The change rate of the number of human-animal cross-infection cases calculated based on the infection data sequence is used as the observation variable of the hidden Markov model to form an observation sequence; Define the parameter set of the hidden Markov model.

9. The zoonosis epidemiology intelligent analysis system according to claim 8, characterized in that Training the hidden Markov model with the change rate of the number of human-animal cross-infection cases as the observation sequence to determine the state distribution, state transition probability matrix and observation probability distribution parameters of the hidden Markov model specifically includes: Calculate the forward probability and backward probability of the given observation sequence O under the current parameter estimation; Update the parameters using the above probabilities: Initial state probability; Iteratively train until the model parameters converge or reach the set number of iterations to obtain the optimal estimate.

10. The zoonosis epidemiology intelligent analysis system according to claim 9, characterized in that, Inferring the current human-animal contact rate based on the current change rate of the number of human-animal cross-infection cases and the trained hidden Markov model includes: Based on the current and historical observation values, calculate the probability that the system is in each hidden state at time t; Calculate the posterior probability of each state, and take the state corresponding to the maximum value as the most likely hidden state at the current moment, that is, determine the current human-animal contact rate level.

Images