An epidemic prevention optimization method for a logistics system based on big data

Abstract

The present application relates to a logistics system epidemic prevention optimization method based on big data, comprising data collection, data matching, data evaluation, risk level confirmation, determination of sampling proportion and disinfection level according to the confirmed risk level, and posting of quarantine certification terminal after evaluating the disinfection effect. The epidemic prevention optimization method provided by the present application adopts a combination model of "random forest + KNN + expert scoring method" for risk prediction, combines big data for rapid evaluation of the risk level of the logistics system epidemic, improves the speed of inspection and quarantine of the entry logistics system, designs the inspection and quarantine of the logistics system and the disinfection effect certification mark and mutual recognition standard, provides the logistics system and the five-element inspection and quarantine disinfection safety level certificate, so that the entry logistics system is exempted from repeated inspection and quarantine and disinfection or excessive disinfection, and the scientific and accurate improvement of the quarantine efficiency of the logistics system, the reduction of service cost, the improvement of entry facilitation and the optimization of the business environment of the double circulation of service are realized.

Description

Technical Field

[0001] This invention relates to the field of epidemic prevention technology, specifically to an optimization method for epidemic prevention of infectious diseases in logistics systems based on big data. Background Technology

[0002] Domestic quarantine work for inbound personnel, means of transport, and goods mainly consists of two stages: declaration and quarantine. Currently, information collection during the declaration stage follows standardized methods for small-data statistical analysis, while the inspection and quarantine process adheres to standardized personnel allocation rules and workflows. In recent years, policies for the prevention and control of infectious diseases related to logistics activities have been improved, standardizing the prevention and disinfection work related to infectious diseases in each function of logistics activities. Furthermore, the established procedures for the prevention and control of infectious diseases among personnel, means of transport, and goods have clarified the relevant workflows for the prevention and control of infectious diseases among inbound personnel and goods, defined the information collection content for inbound personnel and goods, and enabled the tracing of some movement trajectories within the logistics system.

[0003] However, current domestic methods for preventing infectious diseases in logistics systems still have the following shortcomings: 1. The risk assessment mechanism for personnel and goods awaiting inspection in inspection and quarantine work is not highly intelligent, and the ability to assess the risks of other elements in the logistics system is insufficient. Although customs has developed relatively comprehensive and standardized information collection documents in recent years, the extraction of information from these documents is still done manually; the assessment of risk levels is mainly based on qualitative analysis, making it difficult to achieve accurate assessment. 2. The sampling ratio of goods lacks a scientific and precise mechanism, and different sampling ratios are not adopted for goods from areas with different risk levels, resulting in a one-size-fits-all approach. After an infectious disease outbreak, due to the fear that sampling will lead to missed inspections of logistics systems from high-risk areas and cause transmission, most logistics nodes adopt full inspection, which consumes a lot of time, increases costs, and reduces efficiency for logistics systems from low-risk areas. 3. Regarding disinfection models and the evaluation of disinfection effects. On the one hand, the current approach fails to consider the possibility of cross-infection of infectious disease pathogens between logistics systems and their various components. There is no unified standard for disinfection of all elements within the logistics system. Failure to fully disinfect logistics system elements from high-risk areas poses a risk of virus transmission, while full disinfection of elements from low-risk areas would consume significant public resources. On the other hand, existing disinfection effectiveness evaluation methods are time-consuming. Currently, all disinfection effectiveness evaluations stipulated in domestic health regulations are conducted in laboratories, requiring precise detection of the number of pathogens in samples taken before and after disinfection. Without evaluation of disinfection effectiveness, its effectiveness cannot be guaranteed. Furthermore, the current evaluation methods, due to their lengthy process, lead to cargo backlogs within customs-controlled areas of the inbound logistics system, congestion in customs clearance zones, inefficient inbound logistics, low efficiency, and a poor business environment. To address the issue of excessively long testing times for disinfection effectiveness, some scholars have proposed using poliovirus as a standardized control sample to assess disinfection effectiveness. This involves simultaneously disinfecting with poliovirus and the target pathogen; if the poliovirus disinfection effect meets hygiene requirements, the target pathogen disinfection is deemed effective. Other scholars have suggested that GFP-labeled adenovirus can function similarly to poliovirus, providing easier and faster experimental results. However, current poliovirus testing still requires laboratory testing and is time-consuming. 4. The feedback mechanism in quarantine work is inadequate. Problems arising in subsequent circulation stages of quarantined personnel, goods, and vehicles are not considered as feedback information for continuous adjustments to quarantine policies to improve safety. Historical data on infectious disease pathogen testing results for personnel, goods, and vehicles entering through the inbound logistics system, as well as data on the number of public health crises caused by related logistics system elements after entry, are largely absent from the formulation of inbound quarantine reference standards. Consequently, there is almost no dynamic adjustment or iterative optimization of the quarantine mechanism. The scientific accuracy and safety of quarantine work are low. 5. Due to the lack of disinfection assessment and certification and mutual recognition standards, administrative jurisdiction responsibilities are divided and uncoordinated.Due to the lack of disinfection assessment and certification and mutual recognition standards, there is insufficient trust in quarantine work among different regions and lower-level administrative agencies within the country. This results in the need for quarantine and disinfection at all logistics nodes along the route of inbound logistics, leading to duplicate or excessive quarantine and disinfection. This not only wastes public resources but also reduces the efficiency of the logistics system. Summary of the Invention

[0004] The purpose of this invention is to provide an optimization method for the prevention and control of infectious diseases in logistics systems based on big data, which can solve the technical problems mentioned in the background section.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for optimizing the prevention and control of infectious diseases in a logistics system based on big data, comprising the following steps:

[0006] Data collection: Collect infectious disease risk data for elements of the logistics system;

[0007] Data matching: The infectious disease risk data is matched with the infectious disease risk assessment database to obtain the risk level of the logistics system elements;

[0008] Data evaluation: A combined model of expert judgment and KNN algorithm was used to determine the overall risk level of the logistics system;

[0009] Based on the overall risk level, the sampling ratio for detecting elements of the logistics system is matched;

[0010] Based on the overall risk level, match the disinfection level and carry out disinfection.

[0011] Assess the effectiveness of disinfection;

[0012] Save data for retrieval and post quarantine certificates at the terminal.

[0013] As a preferred embodiment, the logistics system elements include at least the following two: drivers, goods, goods packaging, turnover vehicles, and transportation vehicles.

[0014] As a preferred option, the infectious disease risk assessment database is derived from information on the risk of infectious diseases in various regions published by government departments.

[0015] As a preferred approach, before data matching, the number of infectious disease cases in the precise country is used as a training sample, and a random forest regression algorithm is used to obtain a prediction model. The prediction model is then used to process the infectious disease risk database of the non-precise country to obtain a geographical location infectious disease risk value database.

[0016] As a preferred approach, in the data evaluation, scores for low-risk, medium-risk, and high-risk levels are obtained through expert judgment, and values ​​are assigned to the elements of the logistics system to obtain the risk matrix of the logistics system. The risk matrix is ​​then input into the KNN algorithm to obtain the overall risk level of the logistics system.

[0017] As a preferred option, based on the confirmed risk level, disinfection methods are matched for known and unknown pathogens, and then disinfection is carried out.

[0018] As a preferred method, the evaluation of the disinfection effect is carried out by spraying an aqueous solution containing GFP-labeled adenovirus onto the surface of the logistics system elements before disinfection, collecting surface samples, and calculating the number of adenoviruses. After disinfection, surface samples are collected again to calculate the number of adenoviruses.

[0019] As a preferred embodiment, the quarantine certificate terminal includes a radio frequency information database, which includes information such as the label manufacturer, manufacturing time, registration number, and risk information, risk level, disinfection method, and disinfection effect for each element of the logistics system. The registration number is encrypted using a key.

[0020] Compared with existing methods for preventing and controlling infectious diseases in logistics systems, the beneficial effects of this invention are as follows: This invention addresses the shortcomings of current methods for preventing and controlling infectious diseases in inbound logistics systems. It fully utilizes available data on the impact of infectious disease pathogens carried by logistics systems, considers the transmission probability of infectious pathogens among various elements in imported logistics vehicles, and employs a combined algorithm of random forest regression, expert judgment, and KNN to rapidly assess the risk level of infectious diseases in the logistics system as a whole and its individual elements. Furthermore, it optimizes the disinfection methods and effectiveness evaluation methods for logistics systems, improves the efficiency of quarantine work at entry ports, and designs a quarantine result certificate for the logistics system to provide safety assurance for inbound logistics systems, preventing them from being subject to repeated quarantine and disinfection.

[0021] This invention proposes a combined algorithm of random forest regression, expert judgment, and KNN. Random forest provides a more comprehensive data source for expert judgment and KNN. The combination of qualitative and quantitative methods of expert judgment and KNN effectively quantifies the risks of each element and the overall risk of the logistics system. Unlike single random forest regression, the results obtained from the quantitative analysis of random forest regression are qualitatively processed by expert judgment. The judgment interval of expert judgment reduces the error of random forest. Unlike single expert judgment, two expert judgments are used. The first expert judgment obtains the risk levels of the five elements of the logistics system. The second KNN algorithm combines the records of the overall risk of the logistics system judged by the expert judgment. KNN quantitatively processes the results obtained from the qualitative analysis of expert judgment. The case analysis shows that the overall risk judgment model of the logistics system after KNN training is different from the model set by expert judgment. Attached Figure Description

[0022] Figure 1 This is a flowchart illustrating an optimization method for epidemic infectious disease prevention in a logistics system based on big data, according to the present invention.

[0023] Figure 2 This is a schematic diagram illustrating the matching of various logistics system elements in this invention.

[0024] Figure 3 This is a schematic diagram of the quarantine certificate identification and processing in this invention.

[0025] Figure 4 This is a schematic diagram illustrating the relationship between the individual algorithms in the innovative combined algorithm of this invention.

[0026] Figure 5 This is a schematic diagram illustrating the prediction principle of the random forest algorithm in this invention.

[0027] Figure 6 This is a schematic diagram of the logical structure of the quarantine certificate in this invention.

[0028] Figure 7 This is a schematic diagram illustrating the concept and structure of the quarantine certificate in this invention.

[0029] Figure 8 This is the front view of the quarantine certificate in this invention.

[0030] Figure 9 This is a parameter tuning diagram for random forest regression in an embodiment of the present invention.

[0031] Figure 10 This is a graph showing the fit of the random forest regression test set in an embodiment of the present invention.

[0032] Figure 11 This is a parameter tuning diagram of the KNN model in an embodiment of the present invention.

[0033] Figure 12 This is a scatter plot for a visualization of an embodiment of the present invention. Detailed Implementation

[0034] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0035] See Figure 1 This application discloses an optimization method for the prevention and control of infectious diseases in logistics systems based on big data. The steps of this method are as follows:

[0036] 1. Analysis of Logistics System Elements

[0037] In the context of an infectious disease outbreak, direct or indirect contact, as well as the spread of viral mixtures through the air, can lead to the transmission of viruses. Therefore, there is a possibility of cross-infection among all elements within the logistics system. Considering the transmission characteristics of infectious diseases, all elements within the logistics system are included in the data collection scope for infectious diseases, primarily including five elements: drivers, goods, goods packaging, transit vehicles, and transportation vehicles.

[0038] 2. Collection of key data related to epidemic infectious diseases

[0039] Based on whether the logistics system has quarantine certificate, the logistics system is divided into the first entry logistics system and the already entered logistics system. The entry quarantine port collects all information of the first entry logistics system. After entry, each quarantine node obtains the geographical location information of the already entered logistics system through the RFID information of the quarantine certificate.

[0040] 2.1 Collection of Key Data on Epidemic Infectious Diseases in the First-Time Inbound Logistics System

[0041] ① Data collection and sources of infectious disease risk for drivers

[0042] Based on the application areas of the data, driver information is divided into three categories: basic information, health status, and infectious disease risk assessment data. Basic information includes name, gender, age, and nationality; health status includes body temperature, infectious disease test reports, and electronic health certificates; and infectious disease infection risk data includes place of residence, point of origin, transit points, origin and destination regions, and two types of container loading methods (logistics company loading and manufacturer loading).

[0043] The driver's basic information is derived from their ID card or driver's license; body temperature information is obtained from a thermometer; infectious disease test reports and electronic health certificates are obtained from the government's infectious disease tracing platform. If the government's infectious disease tracing platform does not have infectious disease risk information, then infectious disease testing information is provided through paper or electronic versions of infectious disease pathogen test reports; place of residence, starting point of driving, and transit points are derived from the big data travel card. If there is no big data travel card information, the place of residence is provided through the road transport operation permit; import customs declarations provide data on the origin and destination regions; and container loading methods require warehouse entry slips. To obtain this data, drivers must provide their ID card, driver's license, infectious disease test report, electronic health certificate, big data travel card, road transport operation permit, import customs declaration, warehouse entry slip, and other documents or QR codes, barcodes, or other radio frequency codes or codes certified by national standards bodies.

[0044] ② Data collection and sources of infectious disease risk on transportation vehicles

[0045] Similar to ①, the transportation data is divided into two categories: basic information and data on the risk assessment of infectious diseases. The former includes vehicle type and entry permit information, while the latter includes the place of registration, license plate, country (region) of origin, port of export, port of loading, domestic source of goods, and transit points.

[0046] The vehicle type, entry permit information, license plate, port of export, country (region) of origin, port of loading, and domestic cargo origin are all derived from the import customs declaration; the vehicle registration location is provided by the road transport operation permit; and the route is based on the driver's big data travel card. To obtain this data, drivers need to provide documents such as the import customs declaration, road transport operation permit, and big data travel card.

[0047] ③ Data collection and sources of risk of infectious diseases in cargo

[0048] Similar to ①, cargo data is divided into cargo name, quantity information, and infectious disease risk assessment data. Quantity information includes commodity name, specifications, quantity, and unit. Infectious disease risk assessment data includes the cargo's place of origin and port of loading. All data originates from import customs declarations, therefore, the document required from the driver is the import customs declaration.

[0049] ④ Data collection and sources of epidemic infectious disease risk for turnover tools

[0050] Similar to ①, the data on transportation tools is divided into basic information and data on the risk assessment of infectious diseases. Basic information includes the model of the transportation tool and the container number. The data on the risk of infectious diseases related to transportation tools includes the port of loading and transit points of the container. The data on the model of the transportation tool, the container number (i.e., the port of loading), is derived from the import customs declaration, while the transit points are referenced from the driver's big data travel card.

[0051] ⑤ Methods for collecting data on the risk of infectious diseases in cargo packaging

[0052] Since cargo packaging in logistics activities generally takes place in production areas and logistics transit centers, and comes into contact with the environment outside the logistics system during loading, the factors affecting cargo packaging are categorized into three elements: packaging type, country of origin, and port of loading. However, when container loading is done by the logistics company, the driver has direct contact with the cargo packaging; therefore, the container loading type is also included as an evaluation factor. The data for packaging type, country of origin, and port of loading are all derived from import customs declarations. Drivers are required to provide import customs declarations.

[0053] 2.2 Collection of key data on infectious diseases already entering the logistics system

[0054] When an inbound logistics system passes through a domestic quarantine checkpoint, the RFID in the quarantine certificate emits a radio frequency signal. The RFID information reader at the quarantine checkpoint receives the quarantine and disinfection information of the logistics system and the geographical location information of the route. The geographical location information of the route is used as the main basis for re-evaluating the risk level of infectious diseases in the logistics system.

[0055] The certification standards for quarantine results in quarantine certificates are as follows:

[0056] If the logistics system has a quarantine certificate, the mutual recognition of quarantine results in the certificate is based on standards established through visualized data. These standards mainly include: ① After the logistics system is certified upon entry, quarantine certification checks are required at the borders of two regions it passes through within the country. If the logistics system has not passed through medium- or high-risk areas, disinfection is exempt. ② Based on data retrieval using big data visualization, if the entire transportation process does not involve medium- or high-risk areas within the administrative jurisdiction, disinfection is exempt when leaving that area. ③ If a new, unknown infectious disease outbreak occurs, the logistics system's risk information needs to be updated based on the risk locations of the new infectious disease when entering the next logistics node, and a new quarantine certification needs to be conducted. ④ All methods, data, equipment, and data transmission are completed under government-certified standards and secure conditions.

[0057] 2.3 Database for Risk Assessment of Epidemic Infectious Diseases

[0058] The collected data, particularly key data related to infectious diseases, are stored in a temporary sample database within the Infectious Disease Risk Assessment Database. This facilitates subsequent matching with other relevant databases within the Infectious Disease Risk Assessment Database. The following is an introduction to the Infectious Disease Risk Assessment Database:

[0059] The infectious disease risk assessment database D includes D1, D2, ..., D... j The database includes sub-databases such as a global geographic location-based infectious disease risk level database, a quarantine and disinfection database, an infectious disease pathogen environmental monitoring results database, a human health database, and a vehicle model database. The data type code for the geographic location-based infectious disease risk level is 1, i.e., D1.

[0060] The risk indicator criteria are based on a country's infectious disease risk zone classification policy, which defines a high-risk area as one with more than 50 cumulative cases and an outbreak of infectious disease within 14 days. The risk indicator is derived by applying the country's infectious disease risk zone classification policy based on its population density; that is, the risk level indicator value for a specific geographical location abroad is as follows:

[0061]

[0062] In formula (2), NC represents the number of newly infected individuals with a certain infectious disease in the region within 14 days, and Pd represents the number of newly infected individuals with a certain infectious disease in the region within 14 days.L This indicates the population density of the area. This indicates the population density of the aforementioned country.

[0063] 3. Evaluation Methods

[0064] This invention employs a combined prediction algorithm of random forest regression, expert judgment, and KNN, and uses this combined algorithm as the evaluation method for this invention. The algorithm principle is as follows: Figure 4 As shown.

[0065] A risk assessment database for infectious diseases is established using information on the risk of infectious diseases in various regions published by government departments as the data source. Random forest regression is used to improve the risk assessment database. A matching method is used to match the infectious disease information of the logistics system with the infectious disease risk assessment database (the infectious disease risk assessment of the first-time inbound logistics system uses the infectious disease data collected at the port of entry quarantine, and the infectious disease risk assessment of the logistics system that has already entered the country uses the geographical location information of the logistics nodes after entry). This yields the infectious disease risk value of each element of the logistics system. The risk level of each element of the logistics system is then used as input to the trained KNN model to finally obtain the overall infectious disease risk level of the logistics system.

[0066] ①. Random Forest Regression

[0067] This study uses several countries with the number of infectious disease infections pinpointed to the first-level and lower administrative levels (hereinafter referred to as "precision countries") as training samples. Data on the proportion of new infections in each of these countries' lowest-level administrative regions within 14 days relative to the national total, population gathering activity rates, population share, GDP share, passenger traffic volume as a percentage of the national total, freight traffic volume as a percentage of the national total, and the Human Development Index are compiled into a dataset. A certain proportion of data from this dataset is extracted as the training set, and the remainder as the test set. A prediction model is obtained using the random forest regression method. For non-precision countries, datasets containing the proportion of population gathering activity rates, population share, GDP per capita, passenger traffic volume as a percentage of the national total, freight traffic volume as a percentage of the national total, and the Human Development Index at the lowest administrative level are also compiled. These datasets are used as input to the prediction model, yielding the output term: the proportion of new infections relative to the national total. This allows for the determination of the distribution of infectious disease infections in each province of all countries where the number of infectious disease infections cannot be pinpointed to the provincial level. Figure 5 As shown.

[0068] The calculation methods for each index value in the training sample set are as follows:

[0069] Table 1. Calculation methods for each indicator value in the training sample set

[0070]

[0071] Basic model of random forest algorithm:

[0072]

[0073] In equation (3), F * (x) represents the predicted result, and the model means the output dataset F(x,θ) based on the independent variable x. i The mean of A decision trees.

[0074] The random forest algorithm process is as follows:

[0075] The first step is to select the proportion of newly infected cases within 14 days to the total number of infected cases nationwide as the feature variable y, and the rest as non-feature variables Y. The hold-out method is then used to refine the original dataset {M}. m+n 80% of the data was randomly selected as the training sample set. The remaining 20% ​​is used as a test set.

[0076] The second step is to randomly select τ features from the six features in the training sample set. The selected τ features do not change during the random forest generation process, and the unselected features are split into nodes.

[0077] The third step involves repeating the first and second steps, adjusting the parameters, and finally obtaining the optimal prediction model for the random forest.

[0078] The fourth step is to process the test set data. Using the proportion of population gatherings, population size, GDP share, passenger turnover share, freight turnover share, and Human Development Index of each administrative region as inputs, and the proportion of each administrative region's 14-day new infections to the national total as output, a test prediction set was obtained. Compare the test prediction set with the test set The data was compared with real samples. The test set contains actual values ​​of the proportion of newly infected cases in each province within 14 days to the total number of infected cases nationwide. The predictive utility discriminant is then:

[0079]

[0080] In equation (4), S EVS This represents the regression score explaining the variance, where var{·} represents the variance.

[0081] The predictive model is considered effective when the variance regression score is not less than 0.9.

[0082] The fifth step is to combine the non-feature variable Y data from various regions of the non-precise country to form a dataset. Substituting into the prediction model ∈ yields a prediction dataset containing k data points. Prediction dataset Each predicted value and the non-precise number of new infections nationwide on the 14th Multiplying these yields a precise dataset of the number of newly infected individuals nationwide over the past 14 days for each administrative region. Will and Horizontal merging into datasets (Non-precise country), will {M m+n In the dataset, the feature variable y is changed to the number of newly infected people on the 14th. (Precision Nation), and Horizontal merging yields dataset {M m+n+k As a preferred option, extract the dataset {M}. m+n+k The feature variable y in the dataset {y} is obtained. m+n+k}, {y m+n+k Substituting the discriminant formula (2) for the geographical location risk level index, we obtain the dataset. Dataset With {M m+n+k Vertical merging yields the geographic location risk value dataset {M} for administrative regions of various countries. m+n+k This dataset serves as the geographic location-based infectious disease risk value database D1.

[0083] [Note]: Horizontal merging adds rows, and vertical merging adds columns. Subscripts in the dataset indicate the number of samples, i.e., the number of rows.

[0084] ②. Matching method

[0085] The collected key data related to infectious diseases in the logistics system are matched with the risk assessment database for infectious diseases in the logistics system to obtain the risk level of each element in the logistics system.

[0086] The matching model is shown below:

[0087] map(i jk D j ), (i=α, β, γ, λ, Ω) (5)

[0088] In equation (5), α, β, γ, λ, and Ω represent the driver, vehicle, transport vehicle, packaging, and goods, respectively, and map(·) represents the matching function. jk This represents the k-th data point in data type j of element i, where D represents the infectious disease risk database. j This represents the data type corresponding to data type j of element i in the infectious disease risk database. The matching diagram is shown below. Figure 2 As shown.

[0089] Based on the characteristic that infectious diseases are easily transmitted over short distances, vehicles are divided into driver's cab and cargo area, transit vehicles into inner and outer walls, and packaging into inner and outer walls. Analyzing the five elements of the logistics system according to contact type, the elements that generate close contact are: the outer wall of packaging and the inner wall of transit vehicles, the driver's cab and the driver, and the driver and the outer wall of packaging (when the container loading type is logistics company loading).

[0090]

[0091] In equation (6), max(·) is a function that takes the highest risk value. The risks arising from the geographical locations traversed by element i. The expression for the risk level of infectious diseases in geographical locations experienced by other logistics system elements that have close contact with element i. This indicates the risk level represented by the environmental detection results of the infectious disease pathogen of element i.

[0092] ③. A method for determining the risk of a logistics system based on expert judgment and the KNN algorithm.

[0093] The scoring rules are as follows:

[0094] The scoring rules include four indicators: survival rate of infectious diseases (1-fatality), non-infection rate (1-infection rate), impact on public safety, and impact on socioeconomic status. The weights of the four indicators are jointly determined by the expert group. The infection rate and fatality rate of infectious diseases are determined based on the actual situation of the infectious disease. The expert group scores the risk level based on the impact of infectious diseases on public safety and the impact on socioeconomic status, with a scoring range of 0 to 1.

[0095] The expert rating sheet is as follows:

[0096] Table 2. Expert Scoring Sheet

[0097]

[0098] In the expert scoring table, X1 to X4 represent the weights of four indicators: infection rate, mortality rate, impact on public safety, and impact on socioeconomic status of the infectious disease. i1 To H i4 These are the high-risk assessment values ​​for the four indicators given by expert i, where H i1 H i2 H is a fixed value that cannot be changed by experts. i3 H i4The scores assigned by experts are similar to those for high risk, with medium and low risk corresponding to similar indicator scores. The weighted scores from expert i for the three risk levels are as follows:

[0099]

[0100] The expert panel's scoring results are as follows:

[0101] Table 3. Expert Panel Scoring Results

[0102]

[0103] in:

[0104]

[0105] In the formula, L i M represents the recommended score for the low-risk level by the i-th expert. i H represents the recommended score for the medium-risk level by the i-th expert. i Let represent the recommended score for the high-risk level by the i-th expert, n represent the number of experts in the expert group, L represent the final low-risk level value adopted, M represent the final medium-risk level value adopted, and H represent the final high-risk level value adopted.

[0106] The risk assessment value δ for the logistics system carrying epidemic pathogens was obtained through the above steps. R According to the index δ R The risk of logistics systems carrying infectious pathogens is classified into the following levels: δ R ∈[L,M], the logistics system is of low risk; δ R ∈[M,H], the logistics system is of medium risk; δ R ∈[H,1], the logistics system is of high risk.

[0107] The risk level of infectious diseases for each of the five elements of the logistics system was obtained through the matching method. Then, the low, medium, and high risk levels obtained through expert judgment were used to assign values ​​to the five elements of the logistics system, resulting in the risk matrix X∈R of the logistics system. 1,5 The risk matrix is ​​input into the KNN algorithm to obtain the overall risk level of the logistics system.

[0108] KNN, also known as K-Nearest Neighbors, works by analyzing the feature values ​​of the feature variable F to be studied among the K nearest points (K values) to a sample Y in a plane, space, or high-dimensional space. A tree is constructed based on the K values, with a maximum number of trees to be constructed. The most frequent feature value F among these K points is recorded. * and F *The value of F is used as the feature variable of sample Y. Here, K is the radius of KNN, which selects the K nearest points to the sample as the basis for judging the sample feature value. There are three distance calculation methods: Manhattan distance, Euclidean distance, and Chebyshev distance. The distance formula is:

[0109]

[0110] In equation (8), L p (x i ,y j ) represents x i and y j The distance between two points When p is 1, L p (x i ,y j ) represents the Manhattan distance, and L is the distance when p is 2. p (x i ,y j ) represents the Euclidean distance, and L is the distance when the region p is infinite. p (x i ,y j ) represents the Chebyshev distance; a tree is constructed based on the distance formula. There are three types of tree construction: brute force implementation, KD tree implementation, and ball number implementation. Brute force implementation is suitable for situations where the dataset is widely distributed, KD tree is suitable for spaces with dimensions less than 20, and ball number is suitable for spaces with dimensions greater than 20. The maximum number of trees to construct is the threshold of the number of leaf nodes at which subtree construction stops, which is the upper limit of tree construction.

[0111] The risk values ​​of infectious diseases for the five elements of the logistics system can be obtained through matching and expert scoring methods. These five risk values ​​can form a sample matrix [R]. α ,R β ,R γ ,R λ ,R Ω Using this as the dataset to be predicted, the following steps are performed:

[0112] The first step is to determine the overall risk level of n logistics systems using expert judgment, and record this as historical data. Then, a matrix X is created by listing the risk levels of the five elements of the logistics system and the n historical records of the overall risk assessment. t ∈R n,6 The overall risk level (low, medium, and high) of the logistics system was extracted from the training sample dataset as data labels. Then, 80% of the data was randomly selected from this dataset using the hold-out method to form the training sample dataset. i∈{1,2,…,m}, and the sample set consisting of data labels in the training sample set is {y i The sample set consisting of the five elements of a logistics system is {Y}. iThe remaining 20% ​​was used as a test set. The sample set consisting of data labels in the training sample set is The sample set consisting of the five elements of a logistics system is

[0113] The second step is to... i} and {Y i Input the KNN model and train it to obtain the KNN model ω.

[0114] The third step is to process the test set data. Risk levels of the five elements of the Chinese logistics system Using the overall risk level of the logistics system as input and the overall risk level of the logistics system as output, we obtain the prediction set of the overall risk level of the logistics system. Test prediction set With test set real samples Inputting data into the ACC model yields the classification accuracy score. The prediction utility discriminant is:

[0115] ACC≥0.9 (9)

[0116] In equation (9), ACC represents the classification accuracy score.

[0117] A prediction model is considered effective when the classification accuracy score is not less than 0.9.

[0118] The fourth step is to adjust the parameters in the KNN model using ACC to maximize ACC, thus achieving the optimal model and obtaining the prediction model ω. K stands for underfit and overfit.

[0119] The fifth step is to substitute the data to be predicted into the prediction model ω to obtain a prediction dataset containing one data point. This data will be used as the overall risk level of the logistics system.

[0120] 4. Quarantine methods

[0121] Different sampling inspection procedures are applied to logistics systems with different risk levels. When the logistics system is at a high risk level, all five elements within the system are inspected. When the system is at a medium risk level, all elements at or above the medium risk level are inspected. When the system is at a low risk level, elements at or above the medium risk level are selected for inspection, and any one of the other elements is randomly selected for inspection.

[0122] Drivers with a low-risk rating are exempt from inspection; those with a medium-risk rating or higher will be tested for infectious diseases. Vehicles and transport vehicles with a medium-risk rating or higher will undergo quarantine inspection of both the interior and exterior of the driver's cab. Goods and packaging with a low-risk rating will be subject to random inspection at a rate of 2%; those with a medium-risk rating will be subject to random inspection at a rate of 20%; and those with a high-risk rating will require full quarantine inspection of both goods and packaging.

[0123] When no infectious disease pathogens are detected in the samples, the risk level is considered correct, and normal disinfection procedures are carried out. When infectious disease pathogens are detected in the samples, the logistics system is immediately upgraded to a high-risk level, and the logistics system is transferred and isolated, the drivers are isolated for observation, and other elements undergo comprehensive sterilization.

[0124] 5. Disinfection methods

[0125] my country's "Technical Specifications for Disinfection and Pest Control" classifies disinfection processes into four levels based on the effectiveness of disinfectant agents: sterilization, high-level disinfection, medium-level disinfection, and low-level disinfection. Sterilization refers to disinfection methods that kill all microorganisms; high-level disinfection refers to disinfection methods that kill all vegetative bacteria, viruses, fungi and their spores, and most bacterial spores; medium-level disinfection refers to disinfection methods that kill microorganisms other than bacterial spores; and low-level disinfection refers to disinfection methods that kill vegetative bacteria and lipophilic viruses. Since infectious diseases are caused by viruses, low-level disinfection is abandoned, and sterilization, high-level, and medium-level disinfection methods are used instead.

[0126] The sterilization method involves spraying a 5-fold diluted compound hydrogen peroxide disinfectant solution onto the object surface for 5 minutes; the high-level disinfection method involves spraying a 0.1% sodium hypochlorite solution onto the object surface for 5 minutes; and the medium-level disinfection method involves spraying a 75% ethanol solution onto the object surface for 10 minutes.

[0127] Table 4. Disinfection Level Methods

[0128] Disinfection level Disinfectant used Duration of action (minutes) sterilization method 5-fold dilution of compound hydrogen peroxide disinfectant 5 High-level disinfection 0.1% sodium hypochlorite solution 5 Medium-level disinfection method 75% ethanol 10

[0129] Based on the confirmed risk level, disinfection methods are matched and carried out for known and unknown pathogens. When the pathogen of an epidemic infectious disease is unknown or non-spore-forming, sterilization is used for all logistics system elements at medium to high risk levels, excluding drivers. For non-spore-forming epidemic infectious disease pathogens, the following disinfection methods are used:

[0130] To reduce the risk level assessment deviations of various elements caused by unforeseen cross-infections within the logistics system, different disinfection methods were adopted for the five elements according to the risk level of the logistics system. When the system was high-risk, sterilization was used for high-risk elements within the system, while high-level disinfection was used for medium- and low-risk elements. When the system was medium-risk, high-level disinfection was used for high- and medium-risk elements within the system, while low-level disinfection was used for low-risk elements. When the system was low-risk, high-level disinfection was used for high-risk elements within the system, while medium- and low-risk elements were disinfected at a medium level.

[0131] Based on the risk assessment rules of the five elements, when the goods are of a high-risk level, the packaging must also be of a high-risk level. However, the packaging may also be contaminated by the driver due to the container loading method used by the logistics company. Therefore, the risk level of the packaging may be equal to or higher than the risk level of the goods. Due to the characteristics of the packaging enclosing the goods, the packaging and goods are disinfected as a whole. When the risk level of the goods is low, only the packaging is disinfected; when the risk level of the goods is medium to high, the goods are detained for virus testing, and both the packaging and the goods are disinfected. The quantity of goods can be determined through the cargo infectious disease risk database to select the appropriate disinfection method and obtain the expected disinfection time.

[0132] Because disassembling the transport vehicle and the vehicle itself is relatively difficult, and the driver's cab comes into direct contact with the driver, the vehicle is divided into two parts: the cab and the exterior. Treating the transport vehicle and the vehicle as a single unit, different disinfectants are sprayed according to their respective risk levels. The driver is disinfected separately, and the cab is also disinfected separately.

[0133] 6. Evaluation of disinfection effectiveness

[0134] Before disinfection of the logistics system, an aqueous solution containing a certain concentration of GFP-labeled adenovirus was sprayed onto the surface of the system. Since the adenovirus is environmentally compatible with the elements in the system, the kill rate of the virus carried by the system was assumed to be equal to the kill rate of the adenovirus. After spraying the adenovirus solution, surface samples were collected to calculate the adenovirus count. After disinfection, surface samples were collected again to calculate the adenovirus count.

[0135] In my country's "Technical Specifications for Disinfection and Pest Control," the standard for the kill rate of microorganisms in the air is 99.91%. Therefore, this kill rate is used as the indicator for qualified disinfection and pest control. When the disinfect rate reaches greater than 99.91%, the disinfection and pest control is considered qualified. The formula for the disinfect rate is:

[0136]

[0137] In equation (10), C is the virus disinfection rate, T0 is the number of labeled adenoviruses detected in the sample before disinfection, and T1 represents the number of labeled adenoviruses detected in the sample after disinfection.

[0138] 7. Quarantine Certificate

[0139] To demonstrate the safety of the processed logistics system, the precise geographical location of the logistics system's route after entry is recorded. This allows domestic epidemic prevention and quarantine checkpoints to conduct accurate risk assessments of re-emergence infectious diseases after the logistics system enters the country, reducing the increased quarantine time and costs associated with re-quarantining and disinfecting the logistics system upon entry. Certificates of quarantine and disinfection are issued at the port of entry, and the data is stored for future retrieval. Regarding the design of the quarantine result certification mark, this invention mainly involves three aspects: a quarantine certificate terminal, an RFID database, and anti-counterfeiting verification.

[0140] ① Quarantine Certificate Terminal

[0141] The appearance of the quarantine certificate terminal is as follows Figure 7 As shown in the image: 1 on the front of the quarantine certificate terminal, see details. Figure 8 The information includes the certificate number and the information of the supervising authority. Charging port 2 is the battery charging terminal, and electromagnet 3 is used to attach to the surface of the logistics vehicle. The quarantine certificate terminal includes a battery, charging port, electromagnet, navigation and positioning chip, memory, and active RFID. The connection principle between the components is as follows: Figure 6 As shown, the battery powers the electromagnet, navigation and positioning chip, memory, and active RFID. The positioning and navigation chip stores the location information in the memory. At the quarantine node after entry, the active RFID transmits and exchanges information on the risk of infectious diseases in the logistics system with the RFID reader. Finally, the information is transmitted to the infectious disease risk assessment database for evaluation by the quarantine node.

[0142] The battery is a lithium battery, the charging port is a USB Type-C interface, the electromagnet is JNP-15 / 15 from Shenzhen Jiupin Magnetic Technology Co., Ltd., the navigation and positioning chip is a Beidou positioning module, the memory is a TF card, and the active RFID is an ultra-high frequency RFID.

[0143] ② RFID Information Database

[0144] The RFID database contains three categories of information. Part 1 includes the tag manufacturer, production date, and registration number. Part 2 includes risk information, risk level, disinfection methods, and disinfection effectiveness for each element of the logistics system. Part 3 includes the geographical location information of the logistics system after entering the country. The database is shown below:

[0145] Table 5. RFID Storage Information Table

[0146]

[0147] See Figure 3 In this database, the disinfection effects, methods, and results of Parts 1 and 2 are directly displayed after the RFID tags are read. Part 2 risk information is not displayed but imported into the quarantine and disinfection database. Part 3 post-entry geographical location information is not displayed but imported into the quarantine and disinfection database. The port of entry and subsequent logistics nodes at all levels determine whether the logistics system has undergone quarantine and disinfection based on whether it possesses RFID certification. If no RFID certification is available, the logistics system undergoes quarantine and disinfection, and an RFID certificate is issued upon completion. If the logistics system has RFID certification, the risk of the logistics system is reassessed based on the updated risk information (i.e., the risk level of the areas it has traversed) not present in the RFID certificate information. If the updated risk information indicates low risk, the vehicle is released; otherwise, quarantine and disinfection are conducted again, and the RFID information is updated.

[0148] ③. Anti-counterfeiting verification of quarantine certificates

[0149] In Part 1 of the RFID information database, the registration number serves as a credential for verifying the validity of RFID tags. Logistics quarantine nodes determine the validity of RFID certificates by checking whether the registration number exists in the quarantine and disinfection database within the logistics system. The registration number is encrypted with a key; when the RFID tag is read at the logistics node, the encrypted registration number is displayed. It is decrypted and compared with already registered records when imported into the quarantine and disinfection database.

[0150] The following will further illustrate this point using an example of a certain infectious disease outbreak:

[0151] Example:

[0152] Background: On October 24, 2020, a regional infectious disease outbreak occurred in a certain region of country A. After tracing the source of the virus, the CDC found a virus with high homology to the virus in the region when sampling transit containers from three neighboring countries. It was determined that the infectious disease outbreak in the region originated from the overseas containers of the neighboring countries.

[0153] Method: Scenario Reenactment

[0154] We select port a, which is relatively close to a certain region of country A, as the port of implementation, and choose country B, the trading country of port a, as the research object of the logistics system. The following uses the above method to process the logistics system from country B at port a:

[0155] 1. Processing of the Epidemic Infectious Disease Risk Database

[0156] The following is the data on newly added infections of a certain infectious disease outbreak in Country B within the 14 days prior to October 24, 2020:

[0157] Table 6. Data on Infected Individuals in a Certain Infectious Disease Outbreak in Country B

[0158]

[0159]

[0160] The data from the seven first-level administrative regions of countries M, N, P, Q, R, A, and T are used as the training dataset for the model.

[0161] Using the hold-out method, samples are drawn from the data with a sampling ratio of 8:2 between the training and test sets and 15 degrees of freedom to obtain the training set. With test set The training set is input into the random forest model, and the possible values ​​of the random forest model parameters are iterated through to obtain the optimal parameter combination when the explained variance S-regression score of the random forest model reaches its maximum value. The optimal values ​​of each parameter are shown in Table 7.

[0162] Table 7. Random Forest Model Parameters

[0163]

[0164] Input the training set into the training model, and after the model is trained, input the test set. The relevant indicators are input into the model to obtain predictions, and the prediction set is tested. Compare the test prediction set with the real dataset of feature variables in the test set. Compare the results. Finally, we obtain S. EVS =0.9595175953737675>0.9, the random forest model is effective, and the goodness of fit of the random forest to the test set is as follows: Figure 10 As shown in the figure, the fit is quite good.

[0165] The dataset to be predicted, consisting of data related to country B, is as follows:

[0166] Table 8. Dataset to be predicted for Country B

[0167]

[0168]

[0169] The prediction dataset consists of the prediction results As shown in Table 9, y_test_pred represents the proportion of newly infected individuals in the first-level administrative region of country B within 14 days, as predicted by the random forest model, relative to the total number of infected individuals nationwide. By combining the data on infected individuals with a certain infectious disease outbreak in country B within the previous 14 days in Table 6 with the predicted dataset for country B in Table 15, the number of newly infected individuals in each first-level administrative region of country B within 14 days can be obtained. The smallest integer greater than the predicted value is selected when calculating the number of newly infected individuals within 14 days. According to formula (2), the risk value of infectious diseases in each first-level administrative region of country B can be obtained:

[0170] Table 9. National B Prediction Dataset

[0171]

[0172] 2. Logistics system data collection and matching

[0173] Based on the main types of goods imported from country B by country A, and the domestic industrial and market structure of country B, information on possible logistics systems from country B is matched with the global geographic location infectious disease risk level database, the infectious disease pathogen environmental monitoring results database, and the human health database from the infectious disease risk assessment database. This yields the risk values ​​for each of the five elements of the logistics system, as shown below, and is further supported by Table 10-14.

[0174] Table 10. Information Collection Form for Personnel Accompanying Inbound Goods

[0175]

[0176] Table 11. Information Collection Form for Sample Inbound Transportation Vehicles

[0177]

[0178]

[0179] Table 12. Sample Inbound Goods Collection Information Form

[0180]

[0181] Table 13. Information on Sample Entry Transit Vehicles

[0182]

[0183] Table 14. Sample Packaging Information Form for Imported Goods

[0184]

[0185]

[0186] 2. Logistics system data matching

[0187] According to Table 10-14, the risk values ​​of the five elements of the logistics system can be obtained:

[0188] R α =max(0,0,0,0.99,0.99,0.99,0.99,0.99,0)=0.99,

[0189] R β =max(0.99,0.99,0.99,0.99,0.99,0.99)=0.99,

[0190] R γ =max(0.99,0.99,0.99,0.99)=0.99,

[0191] R λ =max(0.99,0.99,0.99)=0.99,

[0192] R Ω =max(0.99,0.99,0.99)=0.99.

[0193] Based on the contact situation of the five elements of the logistics system, according to formula (6), we can obtain:

[0194] R β =max(R) β ,R α ) = 0.99,

[0195] R γ =max(R) γ ,R Ω =0.99.

[0196] Prediction set data composed of the risk values ​​of the five elements of the logistics system [0.99 0.99 0.99 0.99 0.99].

[0197] 3. Determination of the five key elements and overall risk level of the logistics system.

[0198] Let δ be the risk assessment value for the logistics system carrying infectious pathogens obtained by the expert judgment method. R For: δ R ∈[0,0.3], the logistics system is of low risk; δ R ∈[0.3,0.9], the logistics system is of medium risk; δ R ∈[0.9,1], the logistics system is of high risk.

[0199] Set up 5 columns and 1000 rows of random numbers. The 5 columns represent the risk values ​​of personnel, vehicles, transportation tools, cargo packaging, and cargo, respectively. 80% of the values ​​are in the range [0, 0.2], 16% are in the range [0.2, 0.8], and 4% are in the range [0.8, 1]. Randomize the 5000 values ​​and arrange them into matrix X. t ∈R 1000,5 Each row in the matrix X represents a sample data point. t ∈ R 1000,5 The following steps are taken: ① Randomly select 200 data points from the matrix and set the loading method of the transport vehicle to "logistics company packing". Based on the contact situation of the five elements of the logistics system, assuming contact between personnel and cargo packaging, assign the highest risk value between the driver and cargo packaging to personnel and cargo packaging. ② Based on the contact situation of the five elements of the logistics system, assign the highest risk value between personnel and transport vehicle to transport vehicle, while keeping the personnel risk value unchanged. ③ Based on the contact situation of the five elements of the logistics system, assign the highest risk value between the transport vehicle and cargo packaging to the transport vehicle, while keeping the cargo packaging risk value unchanged.

[0200] Assuming expert judgment rules, and using the expert judgment method to obtain the risk levels of 5000 values ​​in the historical dataset, the risk level of infectious diseases for the overall logistics system of each sample is defined as follows: ① When all five elements of the logistics system have a low risk level, the overall infectious disease risk level of the logistics system is considered low. ② When one or two of the five elements of the logistics system have a medium risk level, the overall infectious disease risk level of the logistics system is considered medium. ③ When three of the five elements of the logistics system have a medium risk level, the overall infectious disease risk level of the logistics system is considered high. ④ When one or more of the five elements of the logistics system have a high risk level, the overall infectious disease risk level of the logistics system is considered high. Based on these rules, the risk levels of all sample logistics systems can be obtained. Low risk is designated as label 2, medium risk as label 1, and high risk as label 0, resulting in matrix X. t ∈ R 1000,6 This dataset was used as a historical dataset for the KNN algorithm.

[0201] Using the hold-out method, 80% of the historical dataset is selected as the training sample dataset. The rest are used as the test set. The overall risk level of the logistics system was extracted as the data label from the training sample dataset. The KNN model was used to train the training sample dataset. The possible values ​​of the parameters of the hold-out method and the KNN model were iterated to obtain the optimal parameter combination when the ACC reaches its maximum value. The optimal values ​​of each parameter are shown in Table 15.

[0202] Table 15. Optimal values ​​of parameters in the KNN model

[0203]

[0204] The final classification accuracy score is: ACC = 0.975. According to formula (9), ACC > 0.9, the KNN model ω is effective.

[0205] Data to be predicted Substituting ω into the prediction model yields the prediction dataset. The prediction result is [0], meaning that the risk level of infectious diseases among the research subjects in the logistics system of country B is high. The scatter plot is as follows. Figure 12 As shown in the scatter plot, the dots represent historical datasets, the stars represent the research objects of the logistics system, white represents low-risk levels, gray represents medium-risk levels, and black represents high-risk levels. The score of 0.975 on the test set indicates that KNN quantitatively processes the results obtained from the qualitative analysis of expert judgment, blurring the definition of expert judgment and making the model more flexible.

[0206] 4. Determination of Quarantine Methods for the Five Elements of the Logistics System

[0207] Based on the rules for determining quarantine methods, the quarantine methods for the five key elements of the logistics system research object are shown in Table 16:

[0208] Table 16. Quarantine Methods for Various Elements of the Research Object in the Logistics System

[0209]

[0210] 5. Determination of disinfection methods for the five key elements of the logistics system.

[0211] Based on the rules for determining the disinfection methods, the disinfection methods for the five elements of the logistics system research object are shown in Table 17:

[0212] Table 17. Disinfection methods for various elements of the research object in the logistics system

[0213]

[0214] Before disinfection, a certain concentration of GFP-labeled adenovirus aqueous solution is sprayed onto four elements: vehicles, transit vehicles, cargo packaging, and cargo. The surfaces of each part are then wiped with quarantine swabs, and the number of GFP-labeled viruses is recorded. After disinfection, the surfaces of each part are wiped again with quarantine swabs, and the number of GFP-labeled viruses is recorded. The disinfection effect is judged using formula (10). If the disinfection effect is not up to standard, further disinfection is performed.

[0215] 6. Quarantine certificate

[0216] If the disinfection and sterilization of each part of the logistics system meets the standards, an RFID quarantine result certificate will be issued for the logistics system. The RFID quarantine result certificate information is as follows:

[0217] Table 18. RFID Quarantine Result Certificate Information Form

[0218]

[0219]

[0220] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for optimizing the prevention and control of infectious diseases in a logistics system based on big data, characterized in that, Includes the following steps: Data collection: Collect infectious disease risk data for elements of the logistics system; Data matching: Using the number of infectious disease infections in the precise country as a training sample, a random forest regression algorithm is used to obtain a prediction model. The prediction model is then used to process the infectious disease risk database of non-precise countries to obtain a geographical location infectious disease risk value database. The infectious disease risk data is then matched with the infectious disease risk assessment database to obtain the risk value of logistics system elements. The term "precise country" refers to several countries that can accurately count the number of people infected with infectious diseases down to the first-level or lower administrative level; the term "inaccurate country" refers to several countries that cannot accurately count the number of people infected with infectious diseases down to the first-level or lower administrative level. Data Evaluation: A combined model of expert judgment and KNN algorithm was used to confirm the overall risk level of the logistics system. The risk values ​​of the logistics system elements were assigned using expert judgment to obtain a risk matrix. This risk matrix was then input into the KNN algorithm to obtain the overall risk level of the logistics system. The training process of the KNN algorithm included: determining the overall risk level of n logistics systems based on expert judgment as historical records; creating a matrix of the risk levels of the five elements of the logistics system and the n historical records of the overall risk judgment as a training sample dataset; randomly selecting 80% of the data using the hold-out method as the training sample set and the remaining 20% ​​as the test set; and adjusting the KNN model parameters to maximize the ACC, with a classification accuracy ACC ≥ 0.9 as the criterion for effective prediction. Based on the overall risk level, the sampling ratio for detecting elements of the logistics system is matched; Based on the overall risk level, match the disinfection level and carry out disinfection. Assess the effectiveness of disinfection; Save data for retrieval and post quarantine certificates at the terminal.

2. The method for optimizing the prevention and control of infectious diseases in logistics systems based on big data according to claim 1, characterized in that, The logistics system elements include at least the following two: drivers, goods, goods packaging, turnover vehicles, and transportation vehicles.

3. The method for optimizing the prevention and control of infectious diseases in logistics systems based on big data as described in claim 1, characterized in that, The infectious disease risk assessment database is derived from information on the risk of infectious diseases in various regions published by government departments.

4. The method for optimizing the prevention and control of infectious diseases in logistics systems based on big data according to claim 1, characterized in that, In the data evaluation, scores for low-risk, medium-risk, and high-risk levels are obtained through expert judgment, and values ​​are assigned to the elements of the logistics system to obtain the risk matrix of the logistics system. The risk matrix is ​​then input into the KNN algorithm to obtain the overall risk level of the logistics system.

5. The method for optimizing the prevention and control of infectious diseases in logistics systems based on big data according to claim 1, characterized in that, Based on the confirmed risk level, disinfection methods are matched for known and unknown pathogens, and disinfection is carried out.

6. The method for optimizing the prevention and control of infectious diseases in logistics systems based on big data according to claim 1, characterized in that, The evaluation of the disinfection effect is carried out by spraying an aqueous solution containing GFP-labeled adenovirus onto the surface of the logistics system elements before disinfection, collecting surface samples, and calculating the number of adenoviruses. After disinfection, surface samples are collected again to calculate the number of adenoviruses.

7. The method for optimizing the prevention and control of infectious diseases in logistics systems based on big data according to claim 1, characterized in that, The quarantine certificate terminal includes a radio frequency information database, which includes information such as the label manufacturer, production time, registration number, and risk information, risk level, disinfection method, and disinfection effect of each logistics system element.

8. The method for optimizing the prevention and control of infectious diseases in logistics systems based on big data according to claim 7, characterized in that, The registration number is encrypted with a key.

Images