A rapid determination method of real-time risk grade of new infectious disease based on improved moving average prediction limit method

Abstract

The application provides a new infectious disease real-time risk grade rapid determination method based on a modified moving average prediction limit method. The basic idea of the control chart theory is used, and through difference, standardization, moving average and other technical means, the epidemic "outbreak", the epidemic "peak passing", the growth trend and other indexes are introduced, and the thresholds of the indexes are reasonably set. The epidemic risk grade is determined on the basis of identifying the stage of the epidemic, so that the analysis efficiency and the accuracy of the results are improved.

Description

Technical Field

[0001] This invention relates to a rapid algorithm for determining the real-time risk level of emerging infectious diseases based on an improved moving average prediction limit method, belonging to the field of big data analysis. Technical Background

[0002] This invention overcomes the limitations of previous methods, such as the Optimized Moving Average Prediction Limit (Op-MAPL), which were only applicable to trend assessment and timely risk prediction after an epidemic had developed to a certain stage, and the rapid decline in risk level assessment in the later stages of an epidemic. The improved MAPL is then integrated into a real-time online monitoring system based on open-source internet data. By analyzing the cumulative number of infections of emerging infectious diseases in trackable countries and regions, the system can quickly determine the local epidemic risk level and provide a reference for prevention and control decisions. Summary of the Invention

[0003] The technical problem this invention aims to solve is to provide a rapid real-time risk level determination method for emerging infectious diseases based on a modified moving average prediction limit method. This method uses the cumulative number of trackable emerging infectious disease infections as a basis, draws on the basic ideas of control chart theory, and employs techniques such as differencing, standardization, and moving averages to introduce indicators such as epidemic "outbreak," epidemic "peak," and growth trend. By reasonably setting thresholds for each indicator, the risk level of the epidemic is determined based on the identified stage of the epidemic, thereby improving the efficiency of the analysis and the accuracy of the results.

[0004] The technical solution adopted in this invention is: a rapid algorithm for real-time risk level determination of emerging infectious diseases based on the improved moving average prediction limit method, comprising the following steps:

[0005] Step 1: Establish a time series of new cases for the day: Calculate the number of new cases on day j by comparing the cumulative number of infections on day j-1 with the cumulative number of infections on day j-1, using data released in real-time by an authoritative platform (e.g., GitHub). If the cumulative number of infections on day j is missing, the number of new cases on that day is assumed to be zero.

[0006] Step 2: Standardization of New Case Data: Using T as the time interval, standardize the data for day j using the number of new cases from day jT to day j-1, to obtain z. j Simultaneously, the data for day j is standardized with a 2-day buffer period using the number of new cases from day j-(T+2) to day j-3, resulting in...

[0007] Step 3: Calculate the growth trend value for day j based on the standardized data: To reduce the interference of outliers, first...i The data is truncated (i = j - (T-1), ..., j). Then, to mitigate the impact of sudden events or various random factors, a moving average is calculated on the processed data over time interval T to obtain the growth trend value MA for the j-th day. j ;

[0008] Step 4: Calculate the indicators for distinguishing between the "outbreak" and the "peak" of the epidemic: Calculate the high-sensitivity indicator C3 on day j. j This is used to determine in the early stages whether the epidemic is in an "outbreak" state; at the same time, from a "cautious" perspective, the low-sensitivity detection negative change index C1 on day j is calculated. j This is used to determine whether the daily new cases growth rate has declined "significantly", thereby clarifying whether the epidemic has "passed its peak", that is, whether the epidemic risk level has entered a relatively low state;

[0009] Step 5: Determining the COVID-19 risk level on day j: using the COVID-19 "outbreak" indicator C3 j And the "peak" indicator C1 j Clearly define the state of the epidemic on day j, and comprehensively consider the growth trend value MA. j Based on the number of new cases on that day, determine the risk level of the epidemic in that area on day j and give a risk score.

[0010] Specifically, when C3 on day j i When the threshold value is greater than or equal to the preset threshold I, the j-th day is marked as the "outbreak" node.

[0011] When C1 on day j j When the threshold value is greater than or equal to the preset threshold II, the j-th day is marked as the node "the epidemic begins to decline".

[0012] Specifically, in Step 2, the data on day j is standardized (z). j The calculation formula is:

[0013]

[0014] Where x j This represents the number of new cases on day j. This represents the average number of new cases over the T observation days prior to the j-th day, i.e.

[0015]

[0016] s j This represents the sample standard deviation of the number of new cases on the T-th observation day prior to the j-th day, i.e.

[0017]

[0018] Standardization of Day j data with a 2-day buffer period The formula for calculating the average is:

[0019]

[0020] Specifically, in Step 3, the growth trend value MA on the j-th day... j The calculation formula is:

[0021]

[0022] Specifically, in Step 4, the day j's "outbreak" indicator C3 j The calculation formula is:

[0023]

[0024] Where K represents the minimum deviation of the observed value from the expected value. K is a variable value. Ideally, it should be derived by experts in the relevant field after multiple outbreaks of a certain infectious disease, by reviewing historical data and basing it on professional knowledge, subsequent actual situations, and certain statistical methods. However, for emerging infectious diseases, due to the lack of historical data and the lack of broad consensus on issues such as outbreak criteria, this invention considers using historical settings. For example, referring to the Salmonella Outbreak Detection Algorithm (SODA) of the US Centers for Disease Control and Prevention (CDC), K is set to 1, and the reference threshold is 2, that is, when C3 j When the value is ≥2, the epidemic on day j is considered to be in an "outbreak" state.

[0025] Day j, the COVID-19 "peak" indicator C1 j The calculation formula is:

[0026] C1 j =|min{0,z j +K}|

[0027] The meaning and definition of K here are the same as above. C1 j The reference threshold is 3, that is, when C1 j A threshold of ≥3 indicates a "significant" decline in the daily new case growth rate, clearly indicating that the epidemic has passed its peak. This threshold of 3, as set above, is also a reference value.

[0028] Specifically, the relationship between the applicable conditions and reference range of each indicator in Step 5 and the epidemic risk level and score is shown in Table 1 below:

[0029]

[0030]

[0031] Table 1: Correspondence between indicator range and epidemic risk level on day j, and conditions for indicator use

[0032] Where P20 P 40 P 60 and P 80 These represent the 20th, 40th, 60th, and 80th percentiles of the standard Logistic distribution, respectively. In this invention, their specific values ​​are: P 20 =-1.3863, P 40 = -0.4055, P 60 =0.4055, P 80 =1.3863.

[0033] Specifically, the process for determining the risk level is as follows:

[0034] S1. Start monitoring from a state of no epidemic or sporadic outbreaks, and calculate the C3 value, MA value and C1 value for the day;

[0035] S2. Determine whether the C3 value on or before the Mth day is greater than or equal to the preset threshold I;

[0036] When the result is "yes", proceed to S2-1; that is, it indicates that the epidemic has either experienced an outbreak or is currently in an outbreak phase.

[0037] When the result is "No", proceed to S2-2; that is, it indicates that the epidemic has just begun or that the epidemic is not very contagious.

[0038] S2-1. Determine whether day M-1 is "risk-free".

[0039] If the result is "yes", proceed to S2-1-1;

[0040] S2-1—1, determine whether the epidemic has broken out / re-emerged based on whether the C3 value is greater than or equal to the preset threshold I;

[0041] When the result is "No", determine whether the C1 signal is greater than or equal to the preset threshold II;

[0042] If the result is "yes", proceed to S2-1-2-1;

[0043] If the result is "No", proceed to S2-1-2-2;

[0044] S2-1-2-1, risk level is determined based on MA value and number of new cases;

[0045] S2-1-2-2, determine whether the number of new cases in the past M days is less than or equal to the set value;

[0046] When the result is "yes", it is judged as "no risk".

[0047] If the result is "No", proceed to S2-1-2-2-2;

[0048] S2-1-2-2-2, determine the risk level based on the MA value;

[0049] S2-2. Determine if there are any new cases within M+2 days;

[0050] When the result is "No", it is judged as "No risk";

[0051] If the result is "yes", check if the C3 value exists;

[0052] When the result is "No", it is judged as "No risk".

[0053] When the result is "yes", the outbreak is determined based on whether the C3 value is greater than or equal to the preset threshold I.

[0054] The risk level assessment method for S2-1-2-2-2 is as follows:

[0055] When P 40 <MA j ≤P 60 At this time, the risk level is "medium risk".

[0056] When P 60 <MA j ≤P 80 At that time, the risk level was "medium to high risk";

[0057] When MA j >P 80 At this time, the risk level is "high risk".

[0058] P 40 P 60 and P 80 These are the 40th, 60th, and 80th percentiles of the standard Logistic distribution, respectively.

[0059] The risk level assessment method for S2-1-2-1 is as follows:

[0060] When P 20 <MA j ≤P 40 or MA j ≤P 40 x j When the value is greater than the rated value A, the risk level is "medium-low risk".

[0061] When MA j ≤P 20 At that time, and X j When the value is ≤ A, the risk level is "low risk". 20 P 40These are the 20th and 40th percentiles of the standard Logistic distribution.

[0062] The beneficial effects of this invention are:

[0063] In this invention, by selecting an appropriate time interval, it is proposed that the standardized number of new daily cases follows a standard Logistic distribution, thereby improving the accuracy of judging the relative growth rate of cases.

[0064] This invention constructs a growth trend indicator to more fully mine the potential information in the "daily new cases" sequence data. By drawing on the cumulative and control chart method based on short-term baseline data, it introduces epidemic "outbreak" indicators and epidemic "peak" indicators for auxiliary judgment, overcoming the problems of previous methods that were only applicable to trend assessment and timely risk prediction after tracking the development of the epidemic to a certain extent, and the problem of the risk level assessment declining too quickly in the later stage of the epidemic;

[0065] This invention, through the combination of various indicators, determines the risk level of an epidemic based on its current stage, and partially achieves self-correction, improving the efficiency and accuracy of the analysis. Compared with other algorithms, this invention requires less computation and consumes fewer system resources. When integrated into a real-time online monitoring system based on open-source internet data, it enables faster processing of big data, providing favorable conditions for better real-time monitoring and early warning of emerging infectious disease outbreaks.

[0066] It is worth mentioning that the judgment of epidemic risk level in this invention is mainly based on the "relative growth rate" of cases. The significance of this invention lies in providing a relatively simple and easy-to-use "ruler" for measuring this "relative growth rate". The prototype of this "ruler" comes from the Shewhart control chart in the field of quality management. The main work of this invention is to combine the subsequent development results of control chart theory, some practices of other scholars in introducing control chart methods into infectious disease outbreak detection, and the results of traditional infectious disease ward models, to modify the "prototype ruler" to make it more "suitable" for measuring the "relative growth rate" of infectious disease cases, and to provide it with a specific and comparable value. However, unlike "quality" in the field of management, there is no unified standard for the "fast" and "slow" growth rate of newly emerging infectious disease cases in different countries and regions. In this invention, based on the idea that "the scale of the ruler is generally evenly divided", from a probabilistic perspective, our "ruler" is divided into "low", "low-medium", "medium", "medium-high", and "high" into "five equal parts". This is an attempt, and the classification of levels can be diversified according to the differences in the actual development of the epidemic. Attached Figure Description

[0067] Figure 1 This is a flowchart illustrating the daily COVID-19 risk level determination process for this invention.

[0068] Figure 2In this embodiment of the invention, indicator C3 represents the period from February 17th to March 2nd. j value;

[0069] Figure 3 This refers to the daily growth trend values ​​from February 21 to May 10 in this embodiment of the invention;

[0070] Figure 4 In this embodiment of the invention, indicator C1 represents the period from May 11 to June 21. j Value and daily growth trend value;

[0071] Figure 5 In this embodiment of the invention, indicator C1 is the index from June 22 to October 22. j Value and daily growth trend value;

[0072] Figure 6 In this embodiment of the invention, indicator C1 represents the period from October 23 to December 3. j value. Detailed Implementation

[0073] The present invention will be further explained below with reference to the accompanying drawings and specific embodiments.

[0074] Example 1: As Figure 1-6 As shown, an algorithm for rapid real-time risk level determination of emerging infectious diseases based on an improved moving average prediction limit method is presented.

[0075] S1. Start monitoring from a state of no epidemic or sporadic outbreaks, and calculate the C3 value, MA value and C1 value for the day;

[0076] S2. Determine whether the C3 value within the Mth day (inclusive) is greater than or equal to the preset threshold I;

[0077] If the result is "yes", proceed to S2-1;

[0078] If the result is "No", proceed to S2-2;

[0079] S2-1. Determine whether day M-1 is "risk-free".

[0080] If the result is "yes", proceed to S2-1-1.

[0081] S2-1-1, determine whether the epidemic has broken out / re-emerged based on whether the C3 value is greater than or equal to the preset threshold I;

[0082] When the result is "No", determine whether the C1 signal is greater than or equal to the preset threshold II;

[0083] If the result is "yes", proceed to S2-1-2-1;

[0084] If the result is "No", proceed to S2-1-2-2;

[0085] S2-1-2-1, risk level is determined based on MA value and number of new cases;

[0086] S2-1-2-2, determine whether the number of new cases in the past M days is less than or equal to the set value;

[0087] When the result is "yes", it is judged as "no risk";

[0088] If the result is "No", proceed to S2-1-2-2-2;

[0089] S2-1-2-2-2, determine the risk level based on the MA value;

[0090] S2-2. Determine if there are any new cases within M+2 days;

[0091] When the result is "No", it is judged as "No risk".

[0092] When the result is "yes", it is determined whether a C3 value exists; in this embodiment, the C3 value can only be calculated when there are 9 days of data, therefore,

[0093] When the result is "No", it is judged as "No risk";

[0094] When the result is "yes", that is, when there is a sufficient amount to be calculated, the outbreak is determined based on whether the C3 value is greater than or equal to the preset threshold I. That is, when the C3 value is greater than or equal to the preset threshold I, it is determined to be of medium to low risk, and when the C3 value is less than the preset threshold I, it is determined to be "no risk".

[0095] In this embodiment, since the reference epidemic is a new virus with no historical record, the threshold is referenced to the Salmonella Outbreak Detection Algorithm (SODA) of the U.S. Centers for Disease Control and Prevention (CDC), and is set as follows: threshold I is 2 and threshold II is 3. The threshold can be adjusted according to the differences in the actual application environment.

[0096] The above determination process is carried out according to the rules listed in Table 1. These rules can be adjusted according to differences in actual application environments.

[0097] Taking a specific implementation method as an example, the specific implementation steps are as follows:

[0098] Step 1: Obtain the daily cumulative number of infections for a certain infectious disease in a certain country from February 5th to December 3rd of a certain year (a complete wave of the epidemic), and perform backward difference to obtain the daily new cases from February 6th to December 3rd. After sorting, the daily new cases information table in Table 2 is as follows:

[0099]

[0100] Step 2: Standardize new case data: Take time interval T = 7, x j This represents the number of new cases on day j, where j = 1, ..., 301, corresponding to February 6th to December 3rd. Substitute the corresponding data from Table 2 into the formula.

[0101]

[0102] Standardization of daily new case counts from February 13 to December 3 has been completed.

[0103]

[0104] Furthermore, substitute the corresponding data from Table 2 into the formula.

[0105]

[0106] Complete the standardization of the daily new case count from February 15 to December 3, including the 2-day buffer period.

[0107] The time interval system can be adjusted.

[0108] In this embodiment, the standardized daily new cases follow a standard logistic distribution, which is mainly based on the conclusions of the classic infectious disease model SI. Therefore, it needs to meet three assumptions: ① the total number of people remains unchanged; ② there are only susceptible individuals and infected individuals in the population; ③ effective contact is transmission.

[0109] In this embodiment, the time interval T is set to 7, meaning the period is one week. This setting is based on the following:

[0110] First, regardless of the country or region, the population change within a week is negligible, so model assumption ① is satisfied.

[0111] Secondly, in this embodiment, the number of new cases per day comes from the difference in the cumulative number of infections between two adjacent days, so it can be considered that there are only two categories: susceptible individuals and infected individuals.

[0112] Finally, human factors such as lockdown measures and vaccination have a delayed and stable effect.

[0113] Therefore, it can be assumed that the "infectious environment" faced by the population within a week is relatively unchanged, that is, the standard of "effective contact equals infection" is uniform, and thus the time interval T is optional as long as it can approximately meet the above three conditions.

[0114] Step 3: Calculate the growth trend value based on the standardized data: Calculate the growth trend value based on the z8, ..., z8 values ​​obtained in Step 2. 301 Substitute the following formulas as needed.

[0115]

[0116] The daily growth trend (MA) value from February 19 to December 3 was calculated.

[0117] Step 4: Calculate the indicators for distinguishing between the "outbreak" and the "peak" of the epidemic: The indicators calculated in Step 2... Substitute the following formulas as needed.

[0118]

[0119] The daily C3 values ​​from February 17th to December 3rd were calculated; simultaneously, the z8, ..., z values ​​calculated in Step 2 were... 301 Substitute the following formulas as needed.

[0120] C1 j =|min{0,z j +K}|,j=8,…,301,

[0121] The daily C1 values ​​from February 13 to December 3 were calculated.

[0122] Step 5: Plot a graph based on the daily new case counts from February 6 to March 2 obtained in Step 1 and the daily C3 values ​​from February 17 to March 2 obtained in Step 4, as shown below. Figure 2 As shown in Table 1, the C3 value exceeded the threshold of 2 for the first time on February 20, signaling an outbreak. The risk level at this time was assessed as "low to medium risk" (risk score 2).

[0123] Step 6: Based on the daily new cases obtained in Step 1 and the daily growth trend (MA) value obtained in Step 3, plot the data from February 21 to May 10, as shown below. Figure 3 As shown: The MA value broke through P on February 21st. 60 The horizontal line, as shown in Table 1, can be assessed as "medium to high risk" (risk score 4).

[0124] During the 18 days from February 21st to March 10th, except for March 4th when the MA value was lower than P... 60 Except for the "medium risk" (score 3) value, the MA value was within the range of P for all other time periods. 60 and P 80 The risk level is between the horizontal lines, therefore the "medium to high risk" rating is maintained.

[0125] from Figure 3 It can be observed that the number of new daily cases was clearly on an upward trend during this period. On March 11, the MA value exceeded the P-value. 80 The horizontal line indicates a "high risk" rating (risk score of 5).

[0126] On the other hand, from Figure 3 As can be seen, since March 15, the number of new daily cases has entered a phase of accelerated increase compared to the previous period, meaning that the algorithm issued a "high-risk" warning 4 days in advance.

[0127] By May 2nd, the MA value had dropped to P. 80 Below the horizontal line, the rating has been downgraded from "high risk" to "medium-high risk", which means that the growth rate has slowed down, the epidemic has been controlled to a certain extent, and it indicates that the turning point of this round of the epidemic is coming.

[0128] Due to the slowdown in growth coupled with high-level fluctuations, the MA value declined rapidly, falling to P on May 10th. 60 The rating is downgraded to "medium risk".

[0129] Step 7: Based on the daily new case count obtained in Step 1, the daily growth trend MA value obtained in Step 3, and the daily C1 value obtained in Step 4, plot the data from May 11 to June 21, as shown below. Figure 4 As shown: On May 15th, the MA value broke downwards through P. 40 This indicates that the growth rate has begun to decline.

[0130] However, the daily new cases remained high at this time, and considerable risk remained. Afterwards, the MA value remained at P... 40 Below, the downward trend is established. By June 21, the C1 value had exceeded the threshold of 3, and as shown in Table 1, the risk level had dropped to "low to medium risk". At this point, the daily new cases had decreased by one order of magnitude compared to May 15.

[0131] Step 8: Based on the daily new case count obtained in Step 1, the daily growth trend (MA) value obtained in Step 3, and the daily C1 value obtained in Step 4, plot two data charts for June 22 to October 22 and October 23 to December 3, as follows: Figure 5 , Figure 6 As shown: Starting June 22, the rate of decline in new cases slowed down, such as Figure 5 As shown. By July 10th, the MA value broke through the P level upwards. 40 The risk level was initially at a horizontal line, and the rating returned to "medium risk." Afterward, the daily new cases fluctuated. On September 27th, the C1 value again exceeded the threshold of 3, and the risk rating returned to "low to medium risk," after which the daily new cases began to fluctuate downwards again. By October 22nd, the MA value returned to P... 40 Above the horizontal line, the algorithm again issues a "medium risk" rating. Subsequently, if... Figure 6As shown, while the daily new cases continued to fluctuate downwards, the rate of decline slowed significantly compared to before October 22nd. Since the C1 value did not exceed the threshold of 3 until December 2nd, the country maintained a "medium risk" rating since October 22nd. After December 2nd, a new wave of the epidemic emerged due to the invasion of a new strain.

[0132] Therefore, the proposed method for rapid real-time risk assessment of emerging infectious diseases based on the improved moving average prediction limit method aims to use the cumulative number of trackable emerging infectious diseases as a basis. Based on the inference that the standardized number of new daily cases follows a standard logistic distribution, and drawing on the basic ideas of control chart theory, this method introduces indicators such as epidemic "outbreak", epidemic "peak", and growth trend. By referring to the calculation results of classic infectious disease models and the research results of the US Centers for Disease Control and Prevention, the thresholds of each indicator are reasonably set, providing a simple, efficient and relatively accurate analytical method for rapidly determining the epidemic risk level.

[0133] This embodiment embeds the method into an online real-time monitoring system to obtain the results shown in the embodiment. Then, several senior experts specializing in infectious disease epidemiology and disease prevention and control were randomly invited to conduct peer review of the risk grading results, with a compliance rate of 88.89%.

Claims

1. A method for rapid real-time risk assessment of emerging infectious diseases based on a modified moving average prediction limit method, characterized in that: For the first The steps for determining the COVID-19 risk level in a certain area of ​​Japan are as follows: Step 1: Establish a time series of newly added cases for the day: This is based on real-time data released by authoritative platforms regarding the local number of cases. The daily cumulative number of infections, compared with the first The cumulative daily infection count is then differiated backwards to obtain the [number of infections]. Daily new cases; If the first If the daily cumulative infection count is missing, the default value is zero for the day's new cases. Step 2: Standardization of New Case Data: using... For time interval, using the first From the date The number of new cases on the day Standardize the daily data to obtain At the same time, utilizing the first From the date The number of new cases on the day Daily data is available. The standardization of the buffer period was obtained. ; Step 3: Calculate the first step based on the standardized data. Daily growth trend value: First, for Perform truncation processing ( Then, the processed data is analyzed regarding time intervals. Calculate the moving average to obtain the first... Daily growth trend value ; No. Daily growth trend value The calculation formula is: ； Step 4: Calculate the indicators for distinguishing between the "outbreak" and the "peak" of the epidemic: Calculate the... Hidaka Sensitivity Index This is used to determine in the early stages whether the epidemic is in an "outbreak" state; at the same time, from a "cautious" perspective, the calculation of the first... Low-sensitivity detection of negative change indicators This is used to determine whether the daily new cases growth rate has declined "significantly", thereby clarifying whether the epidemic has "passed its peak", that is, whether the epidemic risk level has entered a relatively low state; No. Japan's COVID-19 "outbreak" indicator The calculation formula is: No. Japan's COVID-19 "peak" indicator The calculation formula is: in This represents the minimum offset of the observed value relative to the expected value; Step5: No. Daily COVID-19 risk level assessment: using the COVID-19 "outbreak" indicator and "peak" indicators , clearly The current state of the daily epidemic and the overall growth trend value. Based on the number of new cases that day, determine the first The local COVID-19 risk level and risk score were announced.

2. The method for rapid real-time risk assessment of emerging infectious diseases based on the improved moving average prediction limit method as described in claim 1, characterized in that: When C3 on day j j When the threshold value is greater than or equal to the preset threshold I, the j-th day is marked as the "outbreak" node.

3. The method for rapid real-time risk assessment of emerging infectious diseases based on the improved moving average prediction limit method as described in claim 1, characterized in that: When C1 on day j j When the threshold value is greater than or equal to the preset threshold II, the j-th day is marked as the "epidemic begins to decline" node.

4. The method for rapid real-time risk assessment of emerging infectious diseases based on the improved moving average prediction limit method as described in claim 1, characterized in that: In step 2, the first Daily Data Standardization The calculation formula is: in, Indicates the first Daily new cases Indicates the first Recently The average number of new cases per observation day, i.e. Indicates the first Recently The sample standard deviation of the number of new cases observed each day, i.e. have The first day of the buffer period Standardization of daily new case data The calculation formula is: 。 5. The method for rapid real-time risk assessment of emerging infectious diseases based on the improved moving average prediction limit method as described in claim 1, characterized in that: The specific process for determining the risk level is as follows: S1. Start monitoring from a state of no epidemic or sporadic outbreaks, and calculate the C3 value, MA value and C1 value for the day; S2. Determine whether the C3 value before the Mth day (including the Mth day) is greater than or equal to the preset threshold I; If the result is "yes", proceed to S2-1; If the result is "No", proceed to S2-2; S2-1. Determine whether day M-1 is "risk-free". When the result is "yes", proceed to S2-1-1. S2-1-1 determines whether the epidemic has broken out / re-broken out based on whether the C3 value is greater than or equal to a preset threshold I. When the result is "No", determine whether the C1 signal is greater than or equal to the preset threshold II; If the result is "yes", proceed to S2-1-2-1; If the result is "No", proceed to S2-1-2-2; S2-1-2-1, risk level is determined based on MA value and number of new cases; S2-1-2-2, determine whether the number of new cases in the past M days is less than or equal to the set value; When the result is "yes", it is judged as "no risk"; When the result is "No", proceed to S2-1-2-2-2, where the risk level is determined based on the MA value. S2-2. Determine if there are any new cases within M+2 days; When the result is "No", it is judged as "No risk"; If the result is "yes", then check if the C3 value exists. When the result is "No", it is judged as "No risk". When the result is "yes", the outbreak is determined based on whether the C3 value is greater than or equal to the preset threshold I.

6. The method for rapid real-time risk assessment of emerging infectious diseases based on the improved moving average prediction limit method as described in claim 5, characterized in that: The risk level assessment method for S2-1-2-2-2 is as follows: when At this time, the risk level is "medium risk". when At that time, the risk level was "medium to high risk"; when At this time, the risk level is "high risk"; The , and These are the first and second parts of the standard Logistic distribution. , , percentile.

7. The method for rapid real-time risk assessment of emerging infectious diseases based on the improved moving average prediction limit method as described in claim 5, characterized in that: The risk level assessment method for S2-1-2-1 is as follows: when or , When the rated value is A, the risk level is "low to medium risk"; when At that time, and X j When the value is ≤ A, the risk level is "low risk"; The , The first is different from the standard Logistic distribution. , percentile.

8. The method for rapid real-time risk assessment of emerging infectious diseases based on the improved moving average prediction limit method as described in any one of claims 5-7, characterized in that: ， ， ， 。

Images