CN115410720A - Clustered infectious disease epidemic situation early warning method based on confirmed patient trajectory characteristics - Google Patents
Clustered infectious disease epidemic situation early warning method based on confirmed patient trajectory characteristics Download PDFInfo
- Publication number
- CN115410720A CN115410720A CN202211341361.3A CN202211341361A CN115410720A CN 115410720 A CN115410720 A CN 115410720A CN 202211341361 A CN202211341361 A CN 202211341361A CN 115410720 A CN115410720 A CN 115410720A
- Authority
- CN
- China
- Prior art keywords
- data
- early warning
- risk
- influence
- points
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 33
- 208000015181 infectious disease Diseases 0.000 title claims abstract description 25
- 208000035473 Communicable disease Diseases 0.000 title claims abstract description 22
- 238000012545 processing Methods 0.000 claims abstract description 35
- 238000012549 training Methods 0.000 claims description 53
- 238000012216 screening Methods 0.000 claims description 33
- 230000014759 maintenance of location Effects 0.000 claims description 17
- 238000004422 calculation algorithm Methods 0.000 claims description 15
- 238000007405 data analysis Methods 0.000 claims description 9
- 201000010099 disease Diseases 0.000 claims description 4
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 claims description 4
- 238000005192 partition Methods 0.000 claims description 4
- 230000004044 response Effects 0.000 claims description 4
- 230000007613 environmental effect Effects 0.000 claims description 3
- 238000003745 diagnosis Methods 0.000 description 19
- 238000005070 sampling Methods 0.000 description 16
- 230000000694 effects Effects 0.000 description 8
- 238000004458 analytical method Methods 0.000 description 6
- 238000000605 extraction Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 230000002411 adverse Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 241001441571 Hiodontidae Species 0.000 description 1
- 230000033228 biological regulation Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004138 cluster model Methods 0.000 description 1
- 230000008094 contradictory effect Effects 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000009792 diffusion process Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 238000010030 laminating Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000001902 propagating effect Effects 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
- 208000024891 symptom Diseases 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/80—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for detecting, monitoring or modelling epidemics or pandemics, e.g. flu
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/29—Geographical information databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Public Health (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Medical Informatics (AREA)
- Tourism & Hospitality (AREA)
- Pathology (AREA)
- Epidemiology (AREA)
- Biomedical Technology (AREA)
- General Engineering & Computer Science (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Remote Sensing (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention provides an aggregated infectious disease epidemic early warning method based on confirmed patient trajectory characteristics, which at least comprises the following steps: collecting and processing personnel data; establishing a detention risk estimation model to obtain the propagation risk degree and the influence radius of a detention point; and dividing early warning areas of different grades according to the propagation risk degree and the influence radius of the stagnation point. The invention provides an epidemic situation early warning method for clustered infectious diseases based on confirmed patient track characteristics, which can establish a risk estimation model according to patient flow information and divide early warning areas of different levels, thereby solving the problem that the position and range of a risk area are difficult to judge.
Description
Technical Field
The invention relates to the technical field of infectious disease epidemic situation early warning, in particular to a clustered infectious disease epidemic situation early warning method based on confirmed patient trajectory characteristics.
Background
Establishing mathematical modeling to predict the epidemic situation development trend, solving related parameters and model results through an MATLAB calculation simulation program, evaluating the error of the results by using statistical indexes, and then using the model with better evaluation effect to perform short-term prediction and medium-term prediction on the epidemic situation development trend.
In the prior art, the position and the diffusion range of the gathering occurrence are extremely difficult to judge, so how to early warn the gathering epidemic occurrence area in advance efficiently and accurately is a big problem to be solved by governments and various epidemic prevention and control units.
Disclosure of Invention
The invention aims to overcome the defects of the existing epidemic situation risk area prediction method and provide a clustered infectious disease epidemic situation early warning method based on confirmed patient track characteristics.
A clustered infectious disease epidemic early warning method based on confirmed patient trajectory characteristics comprises the following steps:
collecting and processing personnel data;
establishing a detention risk estimation model to obtain the propagation risk degree and the influence radius of a detention point;
and classifying early warning areas of different grades according to the propagation risk degree and the influence radius of the stagnation point.
Collecting data and processing includes:
acquiring data;
screening and processing data;
and matching and integrating data.
Establishing a retention risk estimation model to obtain the propagation risk degree and the influence radius of the retention point comprises the following steps:
inputting model parameters according to the data;
overlapping the influence of the stagnation point of the confirmed patient on a block map;
and calculating to obtain the propagation risk degree and the influence radius of each stagnation point.
Superimposing confirmed patient stagnation point effects onto a block map comprises:
converting the input WGS84 longitude and latitude coordinates (lng and lat) into web mercator projection coordinates (x, y), and if the duration = -1, determining the patient ward stagnation point, and if the duration ≠ -1, determining the other stagnation points.
The calculation of the propagation risk degree and the influence radius of each stagnation point specifically includes:
the infection rate parameter of the patient is a parameter which is more than 1, and the infection rate parameter is common to all stagnation points;
the key position identification parameters take environmental and population factors into consideration; if the population is dense and the urban area is occupied, the obtained parameter is larger, and conversely, the parameter is correspondingly smaller in the area with sparser population in the suburb area;
residence time and number of residence points are mainly considered by the residence response coefficient, and the influence range of the regions with long residence time and large number is larger;
the corresponding coefficient of the influence radius and the partition size parameter are used for calculating an influence range on the propagation influence degree.
Before establishing a retention risk estimation model to obtain the propagation risk degree and the influence radius of the retention point, training of the risk estimation model can be further included, specifically:
based on a big data analysis platform, carrying out data processing to obtain a training data set of a risk estimation model;
the training data set is trained.
Based on a big data analysis platform, data processing is carried out to obtain a training data set of a risk estimation model, and the method specifically comprises the following steps:
processing the former aggregated epidemic situation occurrence case data to obtain complete data of the confirmed patient;
defining a series of labeled training data sets
Wherein D is (1) Showing the data set of the previous cases of the occurrence of the aggregated epidemic, the data which was first used to train the network, D (2) Indicating each incremental learningTraining set of (1), x j Representing all stagnation Point data under each class, y j Representing corresponding category labels, each category corresponding to a set of some confirmed patient stagnation points with similar impact radii;
with L ( t ) Representing a set of categories in the t-th dataset and defining that the categories in the series of datasets do not cross, namely:
at this time, the requirement of few samples is added into the increment learning, namely, the requirement D ( t > 1 ) Is from C classes, each class having only K samples, and this is set to C-way K-shot.
Training the training data set specifically includes:
performing basic training;
a plurality of incremental trainings are performed.
The early warning areas with different grades are divided according to the propagation risk degree and the influence radius of the stagnation point, and the early warning areas comprise:
screening out the risk range of stagnation points of newly increased cases in x days;
combining areas with similar positions, risk degrees and influence radiuses through a clustering algorithm;
and enveloping the risk area through an algorithm and sending out early warning information.
Enveloping the risk region through an algorithm, and sending out early warning information specifically comprises the following steps:
setting a judgment radius R;
assuming that the stagnation point data set has n unordered points, drawing a circle with the radius of R through any two points P1 and P2, if no other data point exists in any circle, considering the points P1 and P2 as boundary points, and using a connecting line P1P2 as a boundary line segment;
n data points are connected pairwise to form (n x (n-1))/2 line segments, and judgment and solution are carried out one by one;
and connecting boundary stagnation points of the three risk regions meeting the conditions to form a polygonal region, and enveloping points in the polygonal region to form a polygonal surface so as to form three risk level early warning region data.
The method comprises the steps of collecting names of personnel, primary screening of data such as positive sampling time, current addresses, primary screening of positive report time, retention addresses, birth dates and ages, and screening of the collected data, wherein the flow modulation data only comprises the names, the primary screening of the positive sampling time and the longitude and latitude data, and the track data only comprises the names, the addresses, the arrival time, the retention time and the longitude and latitude data. The data is screened, so that the time of subsequent processing can be reduced, and the adverse effect of redundant data on the final result is avoided. And integrating the screened data to obtain form data comprising names, addresses, arrival time, detention time, longitude and latitude of a detention point, longitude and latitude of a patient, diagnosis time, arrival timestamp and diagnosis timestamp. The system has the advantages that the system is convenient for workers to check, the situation is easier to know, and a large amount of data are supported, so that the output result is more comprehensive and accurate. The form data is input into a stagnation point risk estimation model, the propagation risk degree and the influence radius of the stagnation point are judged according to the model, the model estimation is faster and more accurate than manual estimation, the influence of the data can be comprehensively considered by the model, the analysis time and labor of workers are saved, and the risk area division precision and efficiency are improved.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description, serve to explain the principles of the invention.
In order to more clearly illustrate the embodiments or technical solutions in the prior art of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious for those skilled in the art to obtain other drawings without inventive labor.
Fig. 1 is a flowchart of an early warning method for epidemic situation of infectious disease caused by gathering based on the track characteristics of a patient to be diagnosed according to the present invention;
FIG. 2 is a test result diagram of the early warning method for epidemic situation of infectious disease caused by gathering based on the track characteristics of confirmed patients according to the present invention;
FIG. 3 is a flow chart of data collection and processing of a collective infectious disease epidemic situation early warning method based on confirmed patient trajectory characteristics according to the present invention;
fig. 4 is a flow chart of establishing a stagnation risk estimation model to obtain the propagation risk degree and the influence radius of a stagnation point according to the early warning method for the aggregated infectious disease epidemic situation based on the confirmed patient trajectory characteristics provided by the invention;
fig. 5 is a flow chart of the method for establishing an early warning region cluster model to send out early warning information according to the established patient trajectory characteristics-based clustered infectious disease epidemic early warning method provided by the invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It should be noted that all directional indicators (such as up, down, left, right, front, back \8230;) in the embodiments of the present invention are only used to explain the relative positional relationship between the components, the motion situation, etc. in a specific posture (as shown in the attached drawings), and if the specific posture is changed, the directional indicator is changed accordingly.
In addition, the descriptions relating to "first", "second", etc. in the present invention are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In addition, technical solutions between the embodiments may be combined with each other, but must be based on the realization of the technical solutions by a person skilled in the art, and when the technical solutions are contradictory to each other or cannot be realized, such a combination should not be considered to exist, and is not within the protection scope of the present invention.
The invention discloses a collective infectious disease epidemic situation early warning method based on confirmed patient track characteristics, which comprises the following steps:
step 100, collecting and processing personnel data;
step 200, establishing a detention risk estimation model to obtain the propagation risk degree and the influence radius of a detention point;
and step 300, dividing early warning areas of different grades according to the propagation risk degree and the influence radius of the stagnation point.
The method comprises the steps of collecting names of personnel, primary screening of data such as positive sampling time, current addresses, primary screening of positive report time, retention addresses, birth dates and ages, and screening of the collected data, wherein the flow modulation data only comprises the names, the primary screening of the positive sampling time and the longitude and latitude data, and the track data only comprises the names, the addresses, the arrival time, the retention time and the longitude and latitude data. The data is screened, so that the time of subsequent processing can be reduced, and the adverse effect of redundant data on the final result is avoided. And integrating the screened data to obtain form data comprising names, addresses, arrival time, detention time, latitude and longitude of a detention point, latitude and longitude of a patient, diagnosis time, arrival timestamp and diagnosis timestamp. The system has the advantages that the system is convenient for workers to check, the situation is easier to know, and a large amount of data are supported, so that the output result is more comprehensive and accurate.
The form data is input into a stagnation point risk estimation model, the propagation risk degree and the influence radius of the stagnation point are judged according to the model, the model estimation is faster and more accurate than manual estimation, the influence of the data can be comprehensively considered by the model, and the analysis time and labor of workers are saved.
Preferably, the step 100 of collecting and processing data comprises:
step 101, data acquisition;
firstly, the circulation person records the information of the confirmed patient by telephone or other means, and records the information in a form, wherein the information at least comprises the name, the preliminary screening positive sampling time, the current address, the preliminary screening positive reporting time, the retention address and other data.
The case multi-source data comprises initial screening positive sampling time, stagnation point longitude and latitude, addresses, arrival time, stagnation duration and the like, and relevant information such as the diagnosis time of a patient, the position longitude and latitude coordinates of each stagnation point, the stagnation time, the location type and the like is obtained through a semantic recognition technology.
The multi-source data provides powerful data support for subsequent model analysis, the more detailed data enables early warning time to be more accurate, the early warning range to be more accurate, the prediction mode is more diversified, a large amount of processing time is saved, and the processing efficiency is improved.
Step 102, screening and processing data;
after the record, screening the information data of the flow modulation of the confirmed patient, because the flow modulation information data of the confirmed patient is complex and comprises data such as name, gender, primary screening positive sampling time, birth date, age, current address, primary screening positive report time and the like, the algorithm only needs to screen the data fields such as name, primary screening positive sampling time and longitude and latitude for extraction, as the flow modulation data does not have longitude and latitude, the longitude and latitude can only be extracted through a hundred-degree map through an interface provided by a map information provider such as hundred degrees or high moral and the like, after extraction, the BD-09 longitude and latitude of the hundred-degree map are firstly converted into WGS84 longitude and latitude, then the field of the 'primary screening positive sampling time' is processed, and the data format is as follows: unified processing into standard format% Y/% m/% d.
Secondly, screening and processing the track data of the confirmed patient, also screening required name, address, arrival time, residence time and longitude and latitude data, and processing the field format of the arrival time in the same way to uniformly process the data into the data format% Y/% m/% d of the initial screening positive sampling time.
Because the diagnosis patient circulation information data is complex and comprises data such as name, gender, preliminary screening positive sampling time, birth date, age, current address, preliminary screening positive report time and the like, the algorithm only needs to screen out the fields such as name, preliminary screening positive sampling time and longitude and latitude data for extraction.
The WGS84 coordinate system is a geodetic coordinate system and a coordinate system which is widely used in the GPS global positioning system, the GCJ02 is a mars coordinate system, which is a coordinate system established by the chinese national survey and drawing bureau, and is obtained by encrypting the WGS84, and the BD09 is a hundred-degree coordinate system, which is encrypted again on the basis of the GCJ02 coordinate system. The Baidu coordinate system is converted into a universal WGS84 coordinate system to be subjected to data processing through software, and the classification conversion format is% Y/% m/% d, so that the machine classification identification is facilitated.
In addition, the longitude and latitude coordinates of a Gaode map, a Google map China area and an Tencent map can be selected for processing, the longitude and latitude coordinates of the map software are GCJ02, and the same effect can be achieved only by converting the GCG02 into the WGS84 coordinates.
And 103, matching and integrating data.
Firstly traversing the screened and processed confirmed patient trajectory data, then performing a table-linking operation on the two processed data of 1.1, if the confirmed patient in the screened and processed patient trajectory data can be found in the confirmed patient flow modulation information data, determining the confirmed patient diagnosis time as the processed 'initial screening positive sampling time', converting the time into a time stamp format, and if the matched time is not obtained, temporarily setting the confirmed patient diagnosis time and the time stamp as null values.
Secondly, after the screened and processed diagnosed patient flow regulation information data is processed, the data and the screened and processed diagnosed patient track data are integrated to form table data containing names, addresses, arrival time, detention time, latitude and longitude of a detention point, longitude and latitude of a patient, confirmed diagnosis time, an arrival timestamp and a confirmed diagnosis timestamp. As shown in table 1 below.
For example: if the confirmed patient in the confirmed patient trajectory data can be found in the confirmed patient flow modulation information data, the confirmed patient's time of diagnosis is determined as the processed "initial screening positive sampling time", and the time is converted into a time stamp format, if no match is found, the patient's time of diagnosis and the time stamp are temporarily set to null values.
The form of the table can convert complex characters into simple and effective information, so that the time for workers to extract effective information is saved, the workers can quickly know various information of cases, and the information missing in the table by supplementing the cases can be quickly found.
The invention can also directly put the information of the patient on the map and display the information to the staff in a red dot mode, thus judging the severity of the epidemic situation in the region according to the number of the red dots and then containing other effective information in the red dots.
TABLE 1
Field(s) | Data format | Examples of the invention |
Name (I) | Str | Zeng book |
Address | Str | 38 # of one city in the white cloud area of Guangzhou City |
Time of arrival | Datetime | 2022/04/08 07:50:00 |
Residence time | Number | 10 |
Stagnation Point longitude | Float | 113.26767943722965 |
Latitude of stagnation point | Float | 23.215174492203467 |
Longitude of the point of the patient | Float | 113.31074270982894 |
Latitude of the patient | Float | 23.065356612330746 |
Time of diagnosis | Date | 2022/05/08 |
Time of arrival stamp | Timestamp | 1649498160 |
Time stamp for determining diagnosis | Timestamp | 1649375400 |
The data acquisition method based on the telephone inquiry comprises the steps of name acquisition, preliminary screening of positive sampling time, current address acquisition, preliminary screening of positive report time, address retention and the like, different information acquisition modes can be adopted for different personnel, the method is more humanized than a mechanical mode, the personnel to be checked can accept and cooperate more easily, information can be preliminarily screened artificially, the simplicity of data is improved, the redundancy of the data is reduced, a large amount of workload is reduced for subsequent data processing, and time and energy are saved.
Because the diagnosis patient flow information data is complex, and the algorithm only needs name, preliminary screening of positive sampling time and longitude and latitude data, the data needs to be further screened, redundant data are eliminated, time for distinguishing disordered data is reduced, and processing is convenient.
And then, the data are subjected to table connection operation and integrated into table data containing names, addresses, arrival time, detention time, longitude and latitude of a stagnation point, longitude and latitude of a patient, diagnosis time, arrival timestamp and diagnosis timestamp, so that personnel information is clear at a glance, and the personnel can understand and check conveniently.
Preferably, the step 200 of establishing a stagnation risk estimation model to obtain the propagation risk degree and the influence radius of the stagnation point includes:
step 201, inputting model parameters according to the data;
taking the processed data, the model input parameters [ lng, lat, duration, radius, block ] are respectively corresponding to longitude, latitude, whether the data is a sick house point, radius and an influence block for modeling.
Step 202, overlapping the influence of the stagnation point of the confirmed patient on a block map;
converting the input longitude and latitude coordinates (lng and lat) of the WGS84 into web mercator projection coordinates (x, y), and if the duration = -1, determining the patient's doctor stagnation point, and if the duration ≠ -1, determining the other stagnation points. These two modeling flows differ. Generally, the residence time of the residence point of the patient is long, the mask is not arranged on the home surface, the environment is a closed space, the propagation risk degree is high, and the propagation risk degree of other residence points is relatively low.
The block map is divided into one block with the map, and each block all has corresponding name, has included information such as corresponding street, building, overlaps the influence of stagnation point on the block map, can directly mark the map, makes things convenient for the staff to look over, and audio-visual tentatively knows which block region epidemic situation is more serious, and is inseparable with actual conditions laminating.
Constraint conditions shared by the patient stagnation point and other stagnation points:
when the disease is a stagnation point of a patient, the specific constraint conditions are as follows:
the objective function of the affected block of the ward stagnation point according with the conditions is obtained as follows:
formula one
the other stagnation point influence block objective functions meeting the conditions are obtained as follows:
formula two
And step 203, calculating the propagation risk degree and the influence radius of each stagnation point.
The input model parameters comprise longitude and latitude, whether the model parameters are a patient point, a radius and an influence block, whether the model parameters are transmitted or not is judged according to set constraint conditions, concepts are quantized, the result is more specific and clear, the relation with the data is tighter, the data can be more fully utilized, and the judged result is more persuasive.
The influence of the stagnation points is superposed on the block map and is tightly attached to the actual map, the longitude and latitude lines can be vertically intersected at any position by using the mercator projection coordinates, the map can be drawn on a rectangle, the correct direction between any two points can be displayed, and the linear scale in the coordinate system is kept unchanged around any point in the map.
Wherein divide into the sick house stagnation point and other stagnation points with the stagnation point, because the stagnation point detention time of generally sick house is longer, and the house lining does not have the gauze mask and the environment is airtight space, then the risk degree of propagating is higher, and other stagnation points propagate the risk degree relatively speaking lower. The propagation risk degree is relatively low, the propagation risk degree and the propagation risk degree have special constraint conditions, different conditions are respectively processed, the range division of the affected block is more accurate, and the condition that the range does not accord with the actual range during unified standard processing is avoided.
Preferably, the step 202 of superimposing the confirmed patient stagnation point effect onto the block map comprises:
converting the input WGS84 longitude and latitude coordinates (lng and lat) into web mercator projection coordinates (x, y), and if the duration = -1, determining the patient ward stagnation point, and if the duration ≠ -1, determining the other stagnation points.
The use of the Mooney coordinate system enables the longitude and latitude lines to be vertically intersected at any position, so that a map can be drawn on a rectangle, the correct direction between any two points can be displayed, and the linear scale in the coordinate system keeps unchanged around any point in the map.
The stagnation point is marked according to the situation, the efficiency of data processing can be improved, the data processing is pointed, different methods are used for different situations, and a scientific and reasonable guiding idea is provided for processing the range of the stagnation point. Meanwhile, workers can sort the results when checking the results, and the results are clear at a glance and are not disorderly.
The step 203 of calculating the propagation risk degree and the influence radius of each stagnation point specifically includes:
the infection ratio parameter of the patient is a parameter which is more than 1, and the infection ratio parameter is common to all stagnation points;
the key position identification parameters take environmental and population factors into consideration; if the population is dense and the urban area is occupied, the obtained parameter is larger, and conversely, the parameter is correspondingly smaller in the area with sparser population in the suburb area;
residence time and number of residence points are mainly considered by the residence response coefficient, and the influence range of the regions with long residence time and large number is larger;
the corresponding coefficient of the influence radius and the partition size parameter are used for calculating an influence range on the propagation influence degree.
The parameters influencing the risk area range comprise the infection proportion of a patient, the infection proportion, the key position identification, the resident response coefficient, the corresponding coefficient of the influence radius, the partition size and the like, the influence of each parameter on the risk area range is independently analyzed, and finally all influences are combined to divide the risk area.
Preferably, before the step 300 of establishing the retention risk estimation model to obtain the propagation risk degree and the influence radius of the retention point, training of the risk estimation model may further include:
step 201A, based on a big data analysis platform, performing data processing to obtain a training data set of a risk estimation model;
step 202A, a training data set is trained.
The training model of the invention uses a big data analysis platform, has more data support, can enable the system to rapidly master the characteristics of experimental data, can enable the established model to be closely related to the actual situation through a large amount of actual data training, can solve the proposed problem by combining the actual situation, enables the model to be closer to the reality, and improves the universality and the popularization. Reference experience is provided for establishing a more reasonable model.
Preferably, step 104 is based on a big data analysis platform, and performs data processing to obtain a training data set of the risk estimation model, which specifically includes:
processing the previous aggregated epidemic situation occurrence case data to obtain complete data of the confirmed patients;
defining a series of labeled training data sets
Wherein D is (1) Representing a data set of previous cases of occurrence of an aggregate epidemic, the data that was first used to train the network, D (2) Represents the training set in each incremental learning, x j Represents all stagnation point data under each category, y j Representing corresponding category labels, each category corresponding to a set of some confirmed patient stagnation points with similar impact radii;
by L ( t ) Representing a set of categories in the t-th dataset and defining that the categories in the series of datasets do not intersect with each other, namely:
at this time, the requirement of less samples is added into the increment learning, namely the requirement D ( t > 1 ) Is from C classes, each class having only K samples, and this is set to C-way K-shot.
The model is established on the basis of fully mining sample data, the relationship among all influence factors can be extracted by a system through a large amount of data obtained through big data analysis, unquantized indexes are scientifically and reasonably quantized, sufficient preparation is made for subsequent model training, the training effect is better, and the model is favorable for later-stage model establishment.
The training model of the invention uses a big data analysis platform, has more data support, can enable the system to rapidly master the characteristics of experimental data, can enable the established model to be closely related to the actual situation through a large amount of actual data training, can solve the proposed problem by combining the actual situation, enables the model to be closer to the reality, and improves the universality and the popularization. Reference experience is provided for establishing a more reasonable model.
The invention can also adopt the data of the characteristics of a certain block for analysis, only the name, address, arrival time, detention time, longitude and latitude of a detention point, longitude and latitude of a patient, diagnosis time, arrival timestamp, diagnosis timestamp and other data of the case of the block are input, and the model after analysis is more targeted and has better early warning effect on a certain area alone.
Preferably, the step 105 of training the training data set specifically includes:
step 1051, performing basic training;
basic training is performed by first using a base class dataset (dataset of cases of occurrence of aggregated epidemics) D (1) Training a classification network, wherein the classification network consists of a feature extractor theta F and a classifier theta C, and training the networks theta F and theta C by adopting basic cross entropy loss, so that the network can obtain basic knowledge. After training is complete, only Θ F remains. And finishing the classification training work of the base class data. The cross entropy loss is expressed as follows:
the basic training is carried out, a classification network is trained by using a base class data set (a data set of the gathering epidemic situation occurrence cases), the classification network is composed of a feature extractor theta F and a classifier theta C, and the networks theta F and theta C are trained by adopting basic cross entropy loss, so that the network can obtain basic knowledge.
Step 1052, perform multiple incremental training.
And performing incremental training, wherein during the incremental training, in order to adapt to a training set with few samples and meet the requirement of keeping old knowledge, a small part of network parameters in the specified theta F are selected to participate in the training, and the rest parameters are kept unchanged. When selecting the trainable parameters, firstly setting a threshold value, selecting the network parameters with the absolute values smaller than the threshold value as the trainable parameters, and setting the threshold value to ensure that the ratio of the trainable parameters to the total parameters is 1:10. the loss-in-training function is composed of three parts, and since the classifier Θ C is not available at this time, the cross-entropy loss is not used. The loss at this time is concentrated on the extracted x j Characteristically, three aspects are involved:
(1) triple loss: this loss is used to guarantee x belonging to the same class j Short characteristic distance, belonging to different classes of x j The features are far apart.
(2) Loss of regularization: the loss is used for ensuring that the updated network parameters are very close to the original values, so that the forgetting of old knowledge by the network is reduced.
(3) Cosine similarity loss: this loss keeps the class characteristics of the new class away from the class characteristics of the old class to ensure the accuracy of the network region classification.
Finally, the three losses are weighted and summed to obtain the total loss, which is used to train the trainable parameters in Θ F. Thus, one incremental training is completed.
The basic training is carried out, a classification network is trained by using a base class data set (a data set of the gathering epidemic situation occurrence cases), the classification network is composed of a feature extractor theta F and a classifier theta C, and the networks theta F and theta C are trained by adopting basic cross entropy loss, so that the network can obtain basic knowledge. During incremental training, in order to adapt to a training set with few samples and meet the requirement of keeping old knowledge, a small part of network parameters in specified Θ F are selected to participate in training, and the rest parameters are kept unchanged. A sequence of optimal results and parameters can be obtained after basic training and incremental training, so that the model estimation capability is greatly improved, the error probability is lower, and the output time is shorter.
Preferably, the step 300 of dividing the early warning areas with different levels according to the propagation risk degree and the influence radius of the stagnation point comprises the following steps:
step 301, screening out a stagnation point risk range of a newly increased case in the near x days;
step 302, merging the areas with similar positions, risk degrees and influence radiuses through a clustering algorithm;
and 303, enveloping the risk area through an algorithm and sending out early warning information.
The clustering algorithm has the advantages of easy implementation and high convergence speed, is particularly suitable for a grid division scene, combines areas with similar positions, risk degrees and influence radiuses, draws the same risk areas into the same color, makes the risk areas clear at a glance, facilitates workers to quickly know the risk area range of the same grade, synchronously updates the prediction range and real-time data, enables the workers to take first-hand epidemic situation risk information, and saves precious time.
Enveloping the risk area through an algorithm, and sending out early warning information specifically comprises the following steps:
setting a judgment radius R;
assuming that the stagnation point data set has n unordered points, drawing a circle with the radius of R through any two points P1 and P2, if no other data point exists in any circle, considering the points P1 and P2 as boundary points, and using a connecting line P1P2 as a boundary line segment;
n data points are connected pairwise to form (n x (n-1))/2 line segments, and judgment and solution are carried out one by one; the data points are judged one by one in the judgment radius, any data point cannot be missed, the possibility of error is reduced due to careless mistakes caused by too many data in manual analysis is avoided, and the accuracy of output results is improved. And outputting risk areas with different grades, facilitating follow-up workers to take different measures for different areas, and distributing more workers in serious areas, thereby scientifically and reasonably distributing human resources.
And connecting boundary stagnation points of the three risk regions meeting the conditions to form a polygonal region and enveloping points in the polygonal region to form a polygonal surface, so as to form early warning region data of three risk levels, as shown in table 2.
TABLE 2
Field(s) | Data format | Examples of the invention |
id | Number | 1 |
date | Datetime | 2022-06-07 17:10:11 |
Degree_of_risk | Str | low |
coordinates | List | [[113.31, 23.18], [113.31, 23.15] , [113.31, 23.15]] |
center_lng | float | 113.3329359530381 |
center_lat | float | 23.17585373075093 |
According to the propagation risk degree and the influence radius of the stagnation point, regions with similar positions, risk degrees and influence radii are combined through a clustering algorithm, and finally the risk regions are enveloped through an Alpha Shapes algorithm to send out early warning information. Different early warning information is sent according to the risk level, the situation that a plurality of invalid measures are taken by workers due to the fact that the risk level is not clear is avoided, the workers can conveniently take the medicine according to the symptoms, the risk data can be updated in real time, the workers can check the data on the mobile phone in real time, more flexible measures can be taken in time, and a scientific guiding thought and theory are provided for model prediction of risk areas.
The foregoing are merely exemplary embodiments of the present invention, which enable those skilled in the art to understand or practice the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. An aggregated infectious disease epidemic situation early warning method based on confirmed patient trajectory characteristics is characterized by comprising the following steps:
collecting and processing personnel data;
establishing a detention risk estimation model to obtain the propagation risk degree and the influence radius of a detention point;
and classifying early warning areas of different grades according to the propagation risk degree and the influence radius of the stagnation point.
2. The early warning method for epidemic disease of aggregated infectious diseases based on the confirmed patient trajectory characteristics of claim 1, wherein the collecting and processing personnel data comprises:
acquiring data;
screening and processing data;
and matching and integrating data.
3. The early warning method for the epidemic situation of the infectious disease with the aggregated characteristic based on the track characteristic of the confirmed patient according to claim 1, wherein the establishing of the detention risk estimation model to obtain the propagation risk degree and the influence radius of the detention point comprises the following steps:
inputting model parameters according to the data;
overlapping the influence of the stagnation point of the confirmed patient on a block map;
and calculating to obtain the propagation risk degree and the influence radius of each stagnation point.
4. The method as claimed in claim 3, wherein the step of superimposing the influence of the stagnation point of the confirmed patient on the block map comprises:
converting the input WGS84 longitude and latitude coordinates (lng and lat) into web mercator projection coordinates (x, y), and if the duration = -1, determining the patient ward stagnation point, and if the duration ≠ -1, determining the other stagnation points.
5. The method as claimed in claim 3, wherein the step of calculating the propagation risk degree and the influence radius of each stagnation point comprises:
the infection rate parameter of the patient is a parameter which is more than 1, and the infection rate parameter is common to all stagnation points;
the key position identification parameters take environmental and population factors into consideration; if the population is dense and the urban area is occupied, the obtained parameter is larger, and conversely, the parameter is correspondingly smaller in the area with sparser population in the suburb area;
residence time and number of residence points are mainly considered by the residence response coefficient, and the influence range of areas with long residence time and large number is larger;
the corresponding coefficient of the influence radius and the partition size parameter are used for calculating an influence range on the propagation influence degree.
6. The early warning method for the epidemic situation of the infectious disease with the aggregated structure based on the track characteristics of the confirmed patients according to claim 1, wherein before the establishment of the retention risk estimation model to obtain the propagation risk degree and the influence radius of the retention point, the method further comprises the training of the retention risk estimation model, specifically:
based on a big data analysis platform, carrying out data processing to obtain a training data set of a risk estimation model;
the training data set is trained.
7. The method for early warning of epidemic situation of infectious diseases in group based on confirmed patient trajectory feature of claim 6, wherein the training data set of the retention risk estimation model is obtained by data processing based on the big data analysis platform, and specifically comprises:
processing the previous aggregated epidemic situation occurrence case data to obtain complete data of the confirmed patients;
defining a series of labeled training data sets
Wherein D is (1) Showing the data set of the previous cases of the occurrence of the aggregated epidemic, the data which was first used to train the network, D (2) Represents the training set in each incremental learning, x j All stagnation points under each category are representedData, y j Representing corresponding category labels, wherein each category corresponds to a set of confirmed patient retention points with similar influence radiuses;
by L ( t ) Representing a set of categories in the t-th dataset and defining that the categories in the series of datasets do not cross, namely:
at this time, the requirement of few samples is added into the increment learning, namely, the requirement D ( t > 1 ) The data in (1) is from C categories, each category has only K samples, C is the number of categories in the category set, and K is the number of samples in the category, so that the C-way K-shot is set.
8. The early warning method for epidemic situation of infectious disease in cluster based on confirmed patient trajectory feature of claim 6, wherein the training of the training data set specifically comprises:
performing basic training;
a plurality of incremental trainings are performed.
9. The method for early warning of epidemic disease of aggregated infectious diseases based on confirmed patient trajectory characteristics according to claim 1, wherein the step of classifying early warning areas with different grades according to propagation risk degree and influence radius of stagnation point comprises:
screening out the risk range of stagnation points of newly increased cases in x days;
merging the areas with similar positions, risk degrees and influence radiuses through a clustering algorithm;
and enveloping the risk area through an algorithm and sending out early warning information.
10. The method for early warning of epidemic disease of aggregated infectious diseases based on the track characteristics of confirmed patients according to claim 9, wherein the enveloping the risk area by the algorithm and sending out early warning information specifically comprises:
setting a judgment radius R;
assuming that the stagnation point data set has n unordered points, drawing a circle with the radius of R through any two points P1 and P2, if no other data point exists in any one circle, considering the points P1 and P2 as boundary points, and using a connecting line P1P2 as a boundary line segment;
n data points are connected pairwise to form (n x (n-1))/2 line segments, and judgment and solution are carried out one by one;
and connecting boundary stagnation points of the three risk regions meeting the conditions to form a polygonal region, and enveloping points in the polygonal region to form a polygonal surface so as to form three risk level early warning region data.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211341361.3A CN115410720A (en) | 2022-10-31 | 2022-10-31 | Clustered infectious disease epidemic situation early warning method based on confirmed patient trajectory characteristics |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211341361.3A CN115410720A (en) | 2022-10-31 | 2022-10-31 | Clustered infectious disease epidemic situation early warning method based on confirmed patient trajectory characteristics |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115410720A true CN115410720A (en) | 2022-11-29 |
Family
ID=84168856
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211341361.3A Pending CN115410720A (en) | 2022-10-31 | 2022-10-31 | Clustered infectious disease epidemic situation early warning method based on confirmed patient trajectory characteristics |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115410720A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116564551A (en) * | 2023-05-17 | 2023-08-08 | 北京交通大学 | Data-knowledge driven urban rail transit risk identification method |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103310083A (en) * | 2012-03-09 | 2013-09-18 | 李晓松 | Infectious disease cluster detection and early warning system |
CN111383771A (en) * | 2020-02-26 | 2020-07-07 | 汤一平 | Epidemic disease virus field-based prevention and control system |
CN111885502A (en) * | 2020-06-28 | 2020-11-03 | 华东师范大学 | Epidemic situation prevention and control early warning and tracing system and method for protecting privacy |
CN115132368A (en) * | 2022-05-31 | 2022-09-30 | 山东省计算中心(国家超级计算济南中心) | Infectious disease prevention and control monitoring method and system based on big data platform |
-
2022
- 2022-10-31 CN CN202211341361.3A patent/CN115410720A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103310083A (en) * | 2012-03-09 | 2013-09-18 | 李晓松 | Infectious disease cluster detection and early warning system |
CN111383771A (en) * | 2020-02-26 | 2020-07-07 | 汤一平 | Epidemic disease virus field-based prevention and control system |
CN111885502A (en) * | 2020-06-28 | 2020-11-03 | 华东师范大学 | Epidemic situation prevention and control early warning and tracing system and method for protecting privacy |
CN115132368A (en) * | 2022-05-31 | 2022-09-30 | 山东省计算中心(国家超级计算济南中心) | Infectious disease prevention and control monitoring method and system based on big data platform |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116564551A (en) * | 2023-05-17 | 2023-08-08 | 北京交通大学 | Data-knowledge driven urban rail transit risk identification method |
CN116564551B (en) * | 2023-05-17 | 2024-01-26 | 北京交通大学 | Data-knowledge driven urban rail transit risk identification method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111403048A (en) | Unknown infectious disease early warning and tracing method | |
CN107967487A (en) | A kind of colliding data fusion method based on evidence distance and uncertainty | |
US20230041943A1 (en) | Method for automatically producing map data, and related apparatus | |
Chung et al. | Health and GIS: toward spatial statistical analyses | |
CN115410720A (en) | Clustered infectious disease epidemic situation early warning method based on confirmed patient trajectory characteristics | |
CN110990661A (en) | Interest area entrance and exit extraction method based on road network constraint and density clustering | |
CN107067727B (en) | A kind of road traffic service level evaluation method based on fuzzy KNN characteristic matching | |
CN111899888A (en) | Gynecological tumor disease risk prediction visualization system | |
CN109192312B (en) | Intelligent management system and method for adverse events of heart failure patients | |
CN113420059A (en) | Method and device for actively treating citizen hot line problem | |
CN109543893A (en) | Heterogeneous Information cyberrelationship prediction technique, readable storage medium storing program for executing and terminal | |
CN109523415A (en) | Heterogeneous Information cyberrelationship prediction meanss | |
CN116468043A (en) | Nested entity identification method, device, equipment and storage medium | |
Wheadon | Classification accuracy and consistency under item response theory models using the package classify | |
CN113689959B (en) | Epidemic situation prevention and control decision method, device, equipment and medium based on artificial intelligence | |
Kufa et al. | Syphilis screening coverage and positivity by HIV treatment status among South African pregnant women enrolled in the 2019 antenatal HIV sentinel survey | |
CN115712749A (en) | Image processing method and device, computer equipment and storage medium | |
CN112562864A (en) | New crown epidemic situation detection method, system and device based on block chain technology | |
Zheng et al. | Predicting the distribution of COVID-19 through CGAN—Taking Macau as an example | |
CN111476033B (en) | Bus stop name generation method and device | |
CN109559463A (en) | One kind is attended a banquet instantaneous operating conditions methods of exhibiting | |
CN117933954B (en) | Multi-hospital area manpower resource allocation evaluation method and system based on clinical specialty capability | |
CN117235089B (en) | Map checking method, map checking device, electronic equipment and readable storage medium | |
CN115662600B (en) | Emergency medical resource calling method and system based on geographic position | |
Mahmoud et al. | Development of Health Care Systems in Subareas of KSA Using GIS Concept |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20221129 |
|
RJ01 | Rejection of invention patent application after publication |