CN112786210A - Epidemic propagation tracking method and system - Google Patents
Epidemic propagation tracking method and system Download PDFInfo
- Publication number
- CN112786210A CN112786210A CN202110059223.5A CN202110059223A CN112786210A CN 112786210 A CN112786210 A CN 112786210A CN 202110059223 A CN202110059223 A CN 202110059223A CN 112786210 A CN112786210 A CN 112786210A
- Authority
- CN
- China
- Prior art keywords
- infection risk
- personal
- personnel
- probability
- relationship
- 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.)
- Granted
Links
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
Landscapes
- Health & Medical Sciences (AREA)
- Public Health (AREA)
- Engineering & Computer Science (AREA)
- Medical Informatics (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Pathology (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides an epidemic propagation tracking method and system, and the method is a research method for fusing the incidence risk factors of new individual coronary pneumonia and an individual close contact dynamic relationship network based on space-time big data and an artificial intelligence technology, and a more comprehensive individual infection risk assessment model is constructed. The invention combines the real-time data mining technology and the knowledge map technology, and constructs the personnel spatio-temporal relationship map to realize the dynamic management of close contact relationship of infectors and the knowledge fusion of cross-department. Meanwhile, the infectious disease model is applied to a dynamic personnel relation map with a local community structure, is combined with epidemiological investigation and research, comprehensively considers personnel individual morbidity risk factors and individual close contact relation, realizes scientific, rapid, accurate and dynamic individual infection risk assessment and modeling prediction, is convenient to discover potential infectors with high risk as soon as possible, and is beneficial to improving epidemic situation prevention and control management capacity.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to an epidemic propagation tracking method and system.
Background
The new coronary pneumonia has the obvious characteristic of 'people transmission', and after the epidemic situation appears, the new coronary pneumonia can be rapidly spread in a large area, so that the new coronary pneumonia becomes a great public health safety event threatening the world. The sudden outbreak of the new crown epidemic situation puts new requirements on the public health safety emergency management and disposal capability and the scientific and technological informatization capability. The evaluation of individual infection risks in the process of spreading the new coronary pneumonia is an important premise and basis for accurately positioning epidemic situations and developing the work of prevention and control and emergency management of the new coronary pneumonia, and how to scientifically, accurately, quickly and dynamically predict individual infection risks is an urgent problem to be solved in the work of epidemic prevention and control of the new coronary pneumonia.
The existing disease infection risk assessment research method mainly focuses on two aspects: the method comprises the following steps of firstly, infection trend prediction research based on epidemiological investigation; and infection risk assessment research based on the personal relationship network. 1. Epidemiological-based research has mainly focused on: taking individual morbidity risk factors as evaluation factors of regional infection risks, and establishing an evaluation model to predict the total number and the total trend of possible infection in the whole region from a macroscopic level; predicting the space-time distribution rule and the development trend of the new crown epidemic situation by applying a space epidemiology research method; the machine learning algorithm model is applied to predicting the development trend of the new crown epidemic situation, and the like, so that the research is mostly to carry out macroscopic trend prediction on the new crown epidemic situation; 2. the research is based on the infection risk evaluation research of the personal relationship network, the research is based on the mutual association relationship among individuals, such as family relationship, social relationship, close contact relationship and the like to construct the relationship network, the prediction evaluation is carried out on the infection risk of personal diseases, the research is mostly based on the relatively static relationship network, and the problem of dynamic reconstruction of the relationship network caused by the personal space-time behavior is not considered.
Research on prediction of personal infection risk based on epidemiological investigation methods:
the epidemiological investigation and research method refers to the prediction of the transmission trend of diseases and the infection risk through means of observation, experiment, statistical analysis and the like, and is one of the most common means for evaluating the individual infection risk at present. For example, in the SARS outbreak period of 2003, Li Jun Z and the like adopt a statistical method to analyze sample data collected by epidemic disease investigation, and successfully predict the trend of SARS epidemic situation; the Chinese disease prevention and control center Lizhongjie doctor collects local malaria case data in 2011-2015 of China by applying an epidemiological investigation method, and evaluates the risk of local transmission caused by input malaria cases by adopting a descriptive analysis method. With the increasingly close integration of geographic information science and biological information science research, the spatial-temporal distribution law for predicting disease incidence by applying a spatial epidemiological research method has become a common means for epidemiological investigation and research. The influence of various geographic environmental factors on the mortality of related diseases such as cancer is explored by Yankeen researchers at the Chinese academy of geography and the like through an epidemiological investigation method; the United states space navigation administration combines a geographic information system with epidemiology, and researches the spatial change rule of the incidence of traumatic diseases; amal et al scientifically review the influence of ecological factors such as environment and climate change on the spread of the epidemic disease of Egypt malaria by using geographic information; lysien I et al provides accurate epidemic disease data for a public health system by combining clinical disease characteristics and geographic information, and efficiently and accurately realizes risk prediction of travelers visiting a specific area; fahui Wang et al propose a geographical information system-based regional disease analysis method which is applied to analysis of rare diseases (including cancer) occurring in a minority of partial regions to search for causes of the diseases; in the period of preventing and controlling the new coronary pneumonia, the perihormone red and the like indicate that the space analysis of the epidemic situation and the application of the air information have important guiding significance on the epidemic situation prevention and control decision.
With the rapid development of data collection and storage technologies such as big data and cloud platforms, methods for predicting disease infection risks by applying machine learning algorithm models are receiving more and more attention. The method can train an intelligent algorithm model for solving related business problems based on sample data collected by epidemiological investigation, and further carry out prediction and evaluation on the prevalence trend of diseases and the infection risk. For example, Lim S and other people find potential rules of epidemic diseases and change rules of morbidity by using social media data through an unsupervised learning method; nan Y et al train an AIDS personal incidence prediction model by applying an Artificial Neural Network (ANN) based on AIDS sample data from 2011 to 2017; vervier, K vin and the like apply a structured machine learning method to fuse personal prior information and improve the diagnosis efficiency of diseases; volkova S and the like use a recurrent neural network method to predict the incidence trend of influenza, the model gives full play to the advantage that long-time information dependence can be realized by a long-short memory network (LSTM), and the real-time prediction of the incidence rate of influenza is realized. The application of machine learning algorithm models to predict the development trend of diseases has become a main means of epidemiological research in the present stage.
In the work of preventing and controlling new coronary pneumonia, the disease epidemic prevention control center generally adopts an epidemiological investigation method to investigate potential infectors, explore epidemic situation propagation characteristics, predict epidemic situation development trend and provide basis for government decision. However, many epidemiological investigation methods search and determine individual disease risk factors based on statistical theory, abstract the individual infection risk into the accumulation of individual disease probability in the population, thereby realizing the macro prediction of the population or regional infection risk. However, since new coronary pneumonia has a significant "people-borne" characteristic, the individual infection risk of new coronary pneumonia cannot be understood as a simple accumulation of the individual morbidity of the population in isolation without breaking the association between individuals in the population, and cannot simply see the exposure and morbidity of each individual in the population, but neglect the close contact relationship between individuals and their dynamic changes. Therefore, how to evaluate individual infection risks at a microscopic level based on a close contact relationship network between individuals is another important research direction for evaluating the infection risks of new coronary pneumonia.
Infection risk prediction research based on personal relationship network:
with the development of graph data mining technology and complex network theory, many scholars at home and abroad apply a classical infectious disease model to a complex network, and the problem of disease infection risk assessment considering the mutual association relationship among individuals is solved. Diekmann et al considered a model of SIR disease transmission with repeated contacts between the same individuals on a regular network, assuming that each individual had k contacts with other individuals, where k contacts are a random sampling of the entire population, the study derived the basic regeneration number of the model and the final scale expression of the disease; newman has studied how to model the epidemic trend of infectious diseases by direct contact of people with the network. Cauchemez et al quantify how influenza propagates through social networks. The above research considers the human relationship network as a whole, and studies and judges the transmission trend of the disease on the whole network based on the model of the infectious disease. However, the spread of the disease often varies in different regions, with significant differences between different populations. The development of the community division theory and algorithm lays a foundation for analyzing the propagation model of the local network. In recent years, research on a disease propagation model of a local network has been advanced after division of complex network communities is completed. Zhu G and the like research the infectious disease transmission rule in a local personnel relationship network; boussoid et al have studied the use of models of infectious diseases in local social networks to determine overall risk of disease; reppa et al have studied the individual epidemic disease model based on local network relationship, have utilized the topological network relationship of the individual node to construct the computational model, explore the better infectious disease control strategy.
Infection risk prediction research based on personal relationship networks is generally based on static network models, and the relatively static social relationship between people is considered. However, disease transmission is a dynamic process, closely related to spatiotemporal behavior of humans. Hai-Feng Z, Ren G and other scholars successively put forward an algorithm model for supporting the dynamic evolution of the network community structure, and based on the algorithm model, the close relationship between the individual space-time behavior and the individual infection risk is proved; brdar S and the like capture the behavior and activity rules of individuals by using mobile phone signaling data and research the spatial change rule of the morbidity of the HIV disease. In the prevention and control work of the new coronary pneumonia, the chaimenwei professor provides the space-time behaviors of individuals to be fully researched, and accurate epidemic prevention information is provided for the public by comparing the space-time behavior tracks of confirmed patients and non-confirmed residents. At present, a dynamic relationship network closely contacted with an infected person is mined based on personal space-time behavior data, a potential infected person space-time relationship map is reconstructed by combining a personal social relationship map, and the dynamic assessment of personal infection risk is one of the difficulties of research in the field of disease prevention and control.
At present, the individual infection risk assessment method commonly adopted by the new crown epidemic situation is as follows:
in order to find out the potential infected person of the new coronary pneumonia, epidemic prevention is carried out in advance. The research personnel of the disease control center is combined with the public security organization on the basis of massive epidemiological investigation, and the basic information and the dynamic information of the personnel mastered by the public security organization are combined, so that a great amount of manpower and material resources are spent to screen, analyze and reason the personal contact history of the epidemic situation through methods such as manually drawing a thinking guide graph, a relation graph and the like, and finally, potential infectors are locked, and the epidemic situation detection is reported by media. The success cases show that data resources, personnel relation maps and a time-space relation mining model accumulated by a public security organization in the aspect of accurate personal information analysis are crosswise fused into the personal infection risk assessment business of the new coronary pneumonia, so that the accuracy and the efficiency of personal infection risk assessment of the new coronary pneumonia can be improved.
At present, the prior art lacks a multidimensional individual infection risk assessment model which combines infection trend prediction research based on epidemiological investigation and infection risk assessment research based on personal relationship network and comprehensively considers the close contact relationship between individual incidence risk factors and individuals. As is well known, the risk of individual infection of new coronary pneumonia is closely related to factors such as individual morbidity risk factors, personal relationship networks, individual space-time behaviors, dynamic changes of the personal relationship networks and the like. Simply considering a factor, it is clear that the individual risk of infection for new coronary pneumonia cannot be fully assessed.
The current scheme for evaluating the infection risk of individual spreading diseases seems to achieve certain results in the actual effect of coping with the new coronary pneumonia epidemic situation, but has the following problems:
1. the information level of the disease prevention control center is relatively delayed, and the information sharing is insufficient:
in the new coronary pneumonia prevention and control work, because the information construction level of a disease prevention and control center is relatively lagged, data fusion and knowledge sharing among a plurality of departments are insufficient, many researchers only can manually collect case data and then seek the assistance of departments such as public security, traffic and the like to draw a thinking guide map, and then high-risk infected people can be tracked and positioned, so that the prevention and control work efficiency is low.
2. The mining application of the time-space behavior data is insufficient:
the spatiotemporal behavior data plays an important role in epidemic prevention and control and emergency management because the spatiotemporal behavior data can provide accurate individual information. However, its "epidemic" resistance value is not fully exploited. One of the main reasons is the lack of an effective spatio-temporal behavior data mining analysis method.
3. There is a lack of rapid, accurate, dynamic methods and models for assessing the risk of personal infection:
how to combine a statistical analysis method focusing on individual morbidity risk factors with a complex network model method focusing on a personnel relationship network to exert respective advantages and adapt to the risk assessment of the new crown epidemic situation, and no mature research method exists at present.
Disclosure of Invention
The present invention provides an epidemic propagation tracking method and system to solve the above technical problems.
In order to solve the above technical problem, an embodiment of the present invention provides an epidemic propagation tracking method, including:
constructing according to the obtained public safety field big data resource to obtain a personnel relation map;
adopting a pre-constructed personal close contact relation dynamic mining algorithm model to mine personal close contact relations in real time, and updating the personal close contact relations into the personnel relation map in real time;
calculating the personal infection risk evaluation probability of each person node in the person relation graph based on a person infection risk evaluation regression analysis model obtained through pre-training;
carrying out community structure detection on preset infected persons according to the person relation maps, and then carrying out community structure division on the person relation maps to obtain local person relation maps;
and calculating the personal infection risk probability according to the local personnel relationship map, and calculating by combining the personal infection risk evaluation probability and the personal infection risk probability to obtain the overall personal infection risk probability.
Further, the constructing according to the obtained public safety field big data resource to obtain the personnel relationship map specifically includes:
acquiring public safety field big data resources; the public safety field big data resources comprise personnel basic data, address data and key personnel data;
performing domain professional knowledge fusion on the public safety domain big data resources through an entity extraction and relationship extraction method;
and storing the field professional knowledge by using a graph database Neo4j to construct the human relationship graph.
Further, the training mode of the human infection risk assessment regression analysis model comprises the following steps:
screening all pre-selected risk factors by adopting a Pearson correlation analysis algorithm according to the collected related case sample data to obtain target risk factors of which the absolute values of the correlation coefficients are greater than a preset threshold value;
and training the human infection risk assessment regression analysis model by adopting a Kmeans-Boosting algorithm according to the obtained real human sample data set.
Further, the community structure detection is performed on preset infected persons according to the person relationship graph, and then community structure division is performed on the person relationship graph to obtain a local person relationship graph, and the method specifically includes:
taking a preset node corresponding to an infected person as a community structure center;
performing path diffusion by adopting a preset community structure division algorithm based on the community structure center;
and merging the path nodes meeting the judgment indexes into the community structure by taking the modularity increment as the judgment indexes, and traversing the nodes in the path to obtain a local personnel relationship map corresponding to the preset infected personnel.
Further, the calculating the personal infection risk probability according to the local personnel relationship graph, and calculating by combining the personal infection risk evaluation probability and the personal infection risk probability to obtain the personal infection risk overall probability specifically includes:
calculating personal infection risk probability according to the local personnel relationship map;
and calculating a weighted summation value of the personal infection risk evaluation probability and the personal infection risk probability according to a preset weight to obtain the overall personal infection risk probability.
In order to solve the same technical problem, the invention also provides an epidemic propagation tracking system, which comprises:
the map construction module is used for constructing according to the acquired public safety field big data resources to obtain a personnel relation map;
the relationship updating module is used for mining personal close contact relationships in real time by adopting a pre-constructed personal close contact relationship dynamic mining algorithm model and updating the personal close contact relationships into the personnel relationship map in real time;
the first probability calculation module is used for calculating the personal infection risk evaluation probability of each person node in the person relation graph based on a person infection risk evaluation regression analysis model obtained through pre-training;
the relation graph dividing module is used for carrying out community structure detection on preset infected persons according to the person relation graph and then carrying out community structure division on the person relation graph to obtain a local person relation graph;
and the second probability calculation module is used for calculating the personal infection risk probability according to the local personnel relationship map and calculating by combining the personal infection risk evaluation probability and the personal infection risk probability to obtain the whole personal infection risk probability.
Further, the map building module is specifically configured to: acquiring public safety field big data resources; the public safety field big data resources comprise personnel basic data, address data and key personnel data; performing domain professional knowledge fusion on the public safety domain big data resources through an entity extraction and relationship extraction method; and storing the field professional knowledge by using a graph database Neo4j to construct the human relationship graph.
Further, the training mode of the human infection risk assessment regression analysis model comprises the following steps:
screening all pre-selected risk factors by adopting a Pearson correlation analysis algorithm according to the collected related case sample data to obtain target risk factors of which the absolute values of the correlation coefficients are greater than a preset threshold value;
and training the human infection risk assessment regression analysis model by adopting a Kmeans-Boosting algorithm according to the obtained real human sample data set.
Further, the relationship graph partitioning module is specifically configured to: taking a preset node corresponding to an infected person as a community structure center; performing path diffusion by adopting a preset community structure division algorithm based on the community structure center; and merging the path nodes meeting the judgment indexes into the community structure by taking the modularity increment as the judgment indexes, and traversing the nodes in the path to obtain a local personnel relationship map corresponding to the preset infected personnel.
Further, the second probability calculation module is specifically configured to: calculating personal infection risk probability according to the local personnel relationship map; and calculating a weighted summation value of the personal infection risk evaluation probability and the personal infection risk probability according to a preset weight to obtain the overall personal infection risk probability.
Compared with the prior art, the invention has the following beneficial effects:
the invention is based on space-time big data and artificial intelligence technology, a research method for fusing the individual morbidity risk factors of the new coronary pneumonia and the individual close contact dynamic relationship network is used for constructing a more comprehensive individual infection risk assessment model. The invention combines the real-time data mining technology and the knowledge map technology, and constructs the personnel spatio-temporal relationship map to realize the dynamic management of close contact relationship of infectors and the knowledge fusion of cross-department. Meanwhile, the infectious disease model is applied to a dynamic personnel relation map with a local community structure, is combined with epidemiological investigation and research, comprehensively considers personnel individual morbidity risk factors and individual close contact relation, realizes scientific, rapid, accurate and dynamic individual infection risk assessment and modeling prediction, is convenient to discover potential infectors with high risk as soon as possible, and is beneficial to improving epidemic situation prevention and control management capacity.
Drawings
Fig. 1 is a schematic flow chart of an epidemic propagation tracking method according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of an overall technical route provided by an embodiment of the present invention;
FIG. 3 is a flow chart of basic people relationship graph construction according to an embodiment of the present invention;
FIG. 4 is a technical roadmap for dynamic mining of personal affinity provided by an embodiment of the present invention;
FIG. 5 is a technical roadmap for epidemiological personal risk of infection assessment provided by an embodiment of the present invention;
FIG. 6 is a flowchart of a local community structure detection algorithm according to an embodiment of the present invention;
FIG. 7 is a flowchart of a personal infection risk assessment algorithm based on a local community structure according to an embodiment of the present invention;
fig. 8 is a schematic structural diagram of an epidemic propagation tracking system according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the accompanying 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.
Referring to fig. 1, an embodiment of the present invention provides an epidemic propagation tracking method, including the steps of:
and S1, constructing according to the obtained public safety field big data resource to obtain a personnel relation map.
In the embodiment of the present invention, further, step S1 specifically includes:
acquiring public safety field big data resources; the public safety field big data resources comprise personnel basic data, address data and key personnel data;
performing domain professional knowledge fusion on the public safety domain big data resources through an entity extraction and relationship extraction method;
and storing the field professional knowledge by using a graph database Neo4j to construct the human relationship graph.
And S2, mining the personal close contact relationship in real time by adopting a pre-constructed personal close contact relationship dynamic mining algorithm model, and updating the personal close contact relationship into the personnel relationship map in real time.
And S3, calculating the personal infection risk assessment probability of each person node in the person relation graph based on the person infection risk assessment regression analysis model obtained through pre-training.
In an embodiment of the present invention, further, the training mode of the regression analysis model for human infection risk assessment includes:
screening all pre-selected risk factors by adopting a Pearson correlation analysis algorithm according to the collected related case sample data to obtain target risk factors of which the absolute values of the correlation coefficients are greater than a preset threshold value;
and training the human infection risk assessment regression analysis model by adopting a Kmeans-Boosting algorithm according to the obtained real human sample data set.
S4, carrying out community structure detection on preset infected persons according to the person relation graph, and then carrying out community structure division on the person relation graph to obtain a local person relation graph.
In the embodiment of the present invention, further, step S4 specifically includes:
taking a preset node corresponding to an infected person as a community structure center;
performing path diffusion by adopting a preset community structure division algorithm based on the community structure center;
and merging the path nodes meeting the judgment indexes into the community structure by taking the modularity increment as the judgment indexes, and traversing the nodes in the path to obtain a local personnel relationship map corresponding to the preset infected personnel.
S5, calculating the personal infection risk probability according to the local personnel relationship map, and calculating by combining the personal infection risk evaluation probability and the personal infection risk probability to obtain the overall personal infection risk probability.
In the embodiment of the present invention, further, step S5 specifically includes:
calculating personal infection risk probability according to the local personnel relationship map;
and calculating a weighted summation value of the personal infection risk evaluation probability and the personal infection risk probability according to a preset weight to obtain the overall personal infection risk probability.
Based on the above scheme, in order to better understand the epidemic propagation tracking method provided by the embodiment of the present invention, the following detailed description is provided (see fig. 2 for an overall technical route):
scheme for dynamically mining and managing close contact person time-space relation based on infector time-space behavior data
The research of the personal close contact relation dynamic mining and management scheme based on the personal space-time behavior data mainly solves the problem of mining personal close contact relations in real time and the management problem of dynamic update of association relations, and specifically comprises the following two key links:
1) integrating data resources, and building a relationship map of the basic personnel:
the personnel relationship map is a basic data resource developed by the research of the invention. Firstly, acquiring big data resources in the public safety field, wherein the big data resources comprise various structured data such as personnel basic data, address data, key personnel data and the like; then, a professional knowledge base of the public security interpersonal relationship field is formed through entity extraction and relationship extraction, and the field professional knowledge fusion is realized; and finally, realizing the storage of professional knowledge based on the graph database Neo4j, and completing the construction of the personnel relationship graph. The construction process of the basic personnel relationship map is shown in figure 3.
2) Personal intimate contact relationship mining based on personal spatiotemporal behavioral data:
in order to ensure timeliness of relation data resources in a personnel relation map, the invention needs to solve the problem of dynamic mining of personal close contact relation, especially the problem of dynamic mining of spatio-temporal accompanying relation of an infected person based on personal spatio-temporal behavior data.
The invention realizes real-time data mining algorithms such as stream-type space-time accompaniment, space-time co-occurrence, peer-to-peer co-existence and the like based on a spark stream flow calculation framework. The core thought for realizing the real-time data mining algorithm of the streaming computing framework is as follows: according to the time attribute of the data, the data is subjected to granularity division by adopting an inclined time window, the recent data is subjected to fine-granularity recording, and the more recent data is subjected to coarse-granularity recording, so that the pressure of data storage is reduced on the premise of keeping data information of a longer time period. And finally, realizing a real-time personal close contact relation dynamic mining algorithm model on the basis of realizing real-time data persistence. The mined personal close contact relations are updated to the personal relation map in real time for management. The personal intimate contact relationship dynamic mining technology route is shown in fig. 4.
Secondly, a new individual coronary pneumonia infection risk assessment algorithm model based on epidemiological survey data is as follows:
the method for evaluating the individual infection risk of the new coronary pneumonia based on epidemiological investigation data and the individual morbidity risk factors comprises the following key problems:
1) screening attribute factors combined with new coronary pneumonia case characteristics:
firstly, the invention combines the traditional epidemiology research and development method, and comprehensively selects a plurality of individual morbidity risk factors related to new coronary pneumonia cases, such as: whether people have a travel history of an epidemic area recently, the times of participating in intensive activities of people, the number of confirmed cases near a residence, the number of population of the residence, the times of recent outgoing, basic medical history, and relevant factors such as environmental factors and climate change factors; then, the invention collects sample data of related cases, and adopts Pearson correlation analysis algorithm to primarily screen all risk factors, so as to obtain factors with the absolute value of the correlation coefficient not less than 0.2.
2) Training a personal infection risk assessment model based on machine learning:
the invention takes clinical diagnosis samples provided by Guangdong province disease control center and real sample data of persons without infection after medical observation as a data set. Considering the unbalance of sample data resources, namely the number of confirmed cases is far less than that of non-confirmed cases, the invention trains a personal infection risk assessment regression analysis model by adopting a Kmeans-Boosting algorithm which is researched by the invention team and specially aims at the training of an unbalanced data regression model.
3) Managing evaluation results based on the knowledge graph:
and adding individual morbidity risk factor infection rate to each personnel node in the personnel relationship map, and storing and managing the individual infection risk evaluation probability quantitatively calculated in the second step.
The technical route of the research scheme for the individual infection risk assessment of the new coronary pneumonia based on epidemiological survey data is shown in fig. 5.
Third, personal infection risk assessment based on local personnel relation map
The personal infection risk assessment based on the local personnel relationship maps is based on the constructed personnel relationship maps, a personal infection risk assessment model which comprehensively considers individual morbidity risk factors and personal close contact relations is researched, and the following key problems are researched in an important way:
1) the method for dividing the local community of the person contact relationship map comprises the following steps:
in order to mine a local personnel relationship map related to a certain specified person within a certain space-time range from a personnel relationship map for storing and managing massive personnel relationship data resources, a local community structure of the personnel relationship map needs to be detected first. According to the invention, by using the thought of a classic community discovery algorithm based on modularity, the weight of the relationship is added into the weighing factors of community division, and then local community division is carried out by taking a certain node as a core.
The community structure division algorithm researched by the invention firstly takes the node corresponding to the infected person input in the person relationship graph as a community structure, namely a seed node; and then, taking the node as the center, diffusing the node to the periphery along a path which is directly or indirectly related to the infected person, traversing the nodes of the path one by one, and determining whether the node is merged into the community structure by taking the modularity increment as a determination index. The algorithm comprises the following specific steps:
1. the personal relationship graph network is CN ═ V, E, V denotes a set of nodes, and E denotes a set of edges.
2. Defining a priori weights for all relations:
rijis a node viAnd node vjA relationship of (a), (b), i.e. eij＝(vi，vj，rij) And e is aij∈E
3. Will infect vpAs a community structure, i.e. containing only vpCommunity C ofp
Cp＝(vp)
4. In the infected person vpAcquiring nodes forming relationship with the nodes and node weights V from the knowledge graph for the starting nodesr
Vr＝(vp，E)
Vr＝{∑vi1, 2, 3, m, v (o, p) represents a node, o represents an entity in the network, and p represents an attribute of the entity;
E＝{∑epj|j＝1，2，3...，l}，e＝(vp，vj，rpj)
5. and sequencing the nodes according to the weight:
Vr＝(v1，v2，....vm)
wp1＞wp2＞…＞wpm
6. traversing the sorted nodes, and assuming that the nodes are merged into the community CpCalculating modularity increment, if the modularity increment is larger than 0, determining to add the node to form a new community, otherwise, keeping the community unchanged;
ifΔQi＞0，Cp＝Cp+vi
wherein v isi∈VrM represents the number of clusters of the network,
dc，direspectively representing the node degrees of the node c and the node i, wherein the calculation mode is as follows:
7. for newly incorporated community CpEach node v ofiRepeating the steps 4-6, wherein the starting node of the step 4 needs to be changed into vi；
8. If no new nodes are incorporated into community C in the roundpIf the algorithm is finished, CpI.e. the final output including the infected person vpA community structure of (1).
The flow of the personal relationship map local community detection algorithm of the infected person is shown in fig. 6.
Meanwhile, due to the fact that the personnel time and space are along with the dynamic relation mining, the whole personnel relation map is dynamically changed, and therefore the community structure detected by the method is dynamically evolved according to the dynamic change of the personnel contact relation.
2) Individual infection risk assessment model based on local community structure
Based on the divided local community network of the personal contact relationship, the invention researches the application method of the classic static parameter infectious disease model SIR in the local personal relationship map with individual morbidity risk factor evaluation result, thereby realizing the refined evaluation of the personal infection risk, and the algorithm comprises the following specific steps:
1. and (V, E) setting the divided local community C as (V, E), wherein V is a person set contained in the community, E is a person relationship set, and the relationship weight definition is the same as the community division algorithm. Setting a traversal node stack Q in the community and infecting the number 0 with a node v0Pushing into stack Q.
2. Push node v out of Stack Q0Searching the neighboring node set { v) of the nodej，eij∈E}；
3. Traversing a set of neighboring nodes { vj，eijE, judging vjWhether it is an infected node. If the node belongs to the infected node, the node v is connectedjPushing into stack Q. If not, calculating the node vjThe risk of infection is calculated as follows:
pj＝pi*W*α*δ+qj*θs.t.δ+θ＝1
δ＞0；θ＞0
W＝∑wij，j∈{eij∈E}
in the formula, piIs a node viRisk of local network infection if viWhen it is infected, pi＝1；wijIs the relation eijAlpha is the transition probability between state S and state I in the Moxing country SIR of the infectious disease, qjIs a node vjThe personal infection risk probability evaluated by an epidemiological investigation researcher, wherein delta and theta are weights of the personal local personnel relationship map propagation risk and the individual morbidity risk factor risk, and can be flexibly configured according to actual conditions. After the calculation of the probability p is completed, the node v is calculatedjPushing into stack Q.
And jumping to the second step, and continuing the traversal process until the stack Q is empty, namely, all the nodes in the community are traversed.
The algorithm flow of the personal infection risk assessment model based on the local community structure is shown in fig. 7.
The individual infection risk assessment algorithm model based on the local personnel relationship graph realizes the research goal of combining individual morbidity risk factors and individual close contact relationship to finely assess individual infection risk by weighting and summing the individual infection risk probability assessed based on epidemiological investigation research and the individual infection risk probability assessed based on the local personnel relationship graph.
Fourth, algorithm model verification scheme
The embodiment of the invention verifies the effectiveness of the algorithm model by adopting a cross verification mode based on the new coronary pneumonia sample data set, and the specific verification steps of the algorithm model are as follows:
1. randomly splitting sample data according to the proportion of 8:2, wherein 80% of the sample data is a training set, and 20% of the sample data is a verification set;
2. training a KMeans-Boosting regression analysis model of the incidence rate of the new coronary pneumonia by adopting a training set, calculating the individual incidence risk factor infection risk probability of all people in a personnel relation map based on the model, and updating the relative joint point attribute of the personnel relation map;
3. dynamically mining the personal close contact relation of the verification set samples, and updating the related relation of the personnel relationship map;
4. and calculating the personal infection risk probability of the verification set sample by adopting methods of epidemiological investigation and research evaluation, personal close contact relation evaluation and personal infection risk evaluation based on the local community structure. And the performance of the three algorithm models is measured through the whole mean square error.
By completing the research of the four key technical links, the invention provides a quantitative calculation model for evaluating personal infection risk by comprehensively considering personal static attributes and dynamic contact relation on a microscopic level on the basis of the existing disease infection risk evaluation model, and the performance of the algorithm is measured through new coronary pneumonia sample data.
Compared with the prior art, the embodiment of the invention has the following beneficial effects:
the technical scheme of the invention innovatively combines the individual morbidity risk factors and the individual close contact relationship, and provides a new idea for constructing a refined quantitative evaluation model of the individual disease infection risk. The method discloses key influence factors for determining individual morbidity, integrates an individual dynamic contact relationship network, excavates potential high-risk carrying population, and assists in improving work efficiency related to individual infection risk assessment in new coronary pneumonia epidemic situation prevention and control work.
It should be noted that the above method or flow embodiment is described as a series of acts or combinations for simplicity, but those skilled in the art should understand that the present invention is not limited by the described acts or sequences, as some steps may be performed in other sequences or simultaneously according to the present invention. Further, those skilled in the art will appreciate that the embodiments described in the specification are exemplary embodiments and that no single embodiment is necessarily required by the inventive embodiments.
Referring to fig. 8, in order to solve the same technical problem, the present invention further provides an epidemic propagation tracking system, including:
the map construction module is used for constructing according to the acquired public safety field big data resources to obtain a personnel relation map;
the relationship updating module is used for mining personal close contact relationships in real time by adopting a pre-constructed personal close contact relationship dynamic mining algorithm model and updating the personal close contact relationships into the personnel relationship map in real time;
the first probability calculation module is used for calculating the personal infection risk evaluation probability of each person node in the person relation graph based on a person infection risk evaluation regression analysis model obtained through pre-training;
the relation graph dividing module is used for carrying out community structure detection on preset infected persons according to the person relation graph and then carrying out community structure division on the person relation graph to obtain a local person relation graph;
and the second probability calculation module is used for calculating the personal infection risk probability according to the local personnel relationship map and calculating by combining the personal infection risk evaluation probability and the personal infection risk probability to obtain the whole personal infection risk probability.
Further, the map building module is specifically configured to: acquiring public safety field big data resources; the public safety field big data resources comprise personnel basic data, address data and key personnel data; performing domain professional knowledge fusion on the public safety domain big data resources through an entity extraction and relationship extraction method; and storing the field professional knowledge by using a graph database Neo4j to construct the human relationship graph.
Further, the training mode of the human infection risk assessment regression analysis model comprises the following steps:
screening all pre-selected risk factors by adopting a Pearson correlation analysis algorithm according to the collected related case sample data to obtain target risk factors of which the absolute values of the correlation coefficients are greater than a preset threshold value;
and training the human infection risk assessment regression analysis model by adopting a Kmeans-Boosting algorithm according to the obtained real human sample data set.
Further, the relationship graph partitioning module is specifically configured to: taking a preset node corresponding to an infected person as a community structure center; performing path diffusion by adopting a preset community structure division algorithm based on the community structure center; and merging the path nodes meeting the judgment indexes into the community structure by taking the modularity increment as the judgment indexes, and traversing the nodes in the path to obtain a local personnel relationship map corresponding to the preset infected personnel.
Further, the second probability calculation module is specifically configured to: calculating personal infection risk probability according to the local personnel relationship map; and calculating a weighted summation value of the personal infection risk evaluation probability and the personal infection risk probability according to a preset weight to obtain the overall personal infection risk probability.
It can be understood that the above system embodiments correspond to the method embodiments of the present invention, and the epidemic propagation tracking system provided by the embodiments of the present invention can implement the epidemic propagation tracking method provided by any one of the method embodiments of the present invention.
While the foregoing is directed to the preferred embodiment of the present invention, it will be understood by those skilled in the art that various changes and modifications may be made without departing from the spirit and scope of the invention.
Claims (10)
1. An epidemic propagation tracking method is characterized by comprising the following steps:
constructing according to the obtained public safety field big data resource to obtain a personnel relation map;
adopting a pre-constructed personal close contact relation dynamic mining algorithm model to mine personal close contact relations in real time, and updating the personal close contact relations into the personnel relation map in real time;
calculating the personal infection risk evaluation probability of each person node in the person relation graph based on a person infection risk evaluation regression analysis model obtained through pre-training;
carrying out community structure detection on preset infected persons according to the person relation maps, and then carrying out community structure division on the person relation maps to obtain local person relation maps;
and calculating the personal infection risk probability according to the local personnel relationship map, and calculating by combining the personal infection risk evaluation probability and the personal infection risk probability to obtain the overall personal infection risk probability.
2. The epidemic propagation tracking method according to claim 1, wherein the constructing according to the obtained public safety field big data resource to obtain the personnel relationship map specifically comprises:
acquiring public safety field big data resources; the public safety field big data resources comprise personnel basic data, address data and key personnel data;
performing domain professional knowledge fusion on the public safety domain big data resources through an entity extraction and relationship extraction method;
and storing the field professional knowledge by using a graph database Neo4j to construct the human relationship graph.
3. The epidemic propagation tracking method of claim 1, wherein the training mode of the regression analysis model for human infection risk assessment comprises:
screening all pre-selected risk factors by adopting a Pearson correlation analysis algorithm according to the collected related case sample data to obtain target risk factors of which the absolute values of the correlation coefficients are greater than a preset threshold value;
and training the human infection risk assessment regression analysis model by adopting a Kmeans-Boosting algorithm according to the obtained real human sample data set.
4. The epidemic propagation tracking method according to claim 1, wherein the community structure detection is performed on preset infected persons according to the person relationship graph, and then community structure division is performed on the person relationship graph to obtain a local person relationship graph, specifically comprising:
taking a preset node corresponding to an infected person as a community structure center;
performing path diffusion by adopting a preset community structure division algorithm based on the community structure center;
and merging the path nodes meeting the judgment indexes into the community structure by taking the modularity increment as the judgment indexes, and traversing the nodes in the path to obtain a local personnel relationship map corresponding to the preset infected personnel.
5. The epidemic propagation tracking method according to claim 1, wherein the calculating of the individual infection risk probability according to the local personnel relationship map and the calculation of the individual infection risk evaluation probability and the individual infection risk probability are combined to obtain the overall probability of the individual infection risk, specifically comprises:
calculating personal infection risk probability according to the local personnel relationship map;
and calculating a weighted summation value of the personal infection risk evaluation probability and the personal infection risk probability according to a preset weight to obtain the overall personal infection risk probability.
6. An epidemic propagation tracking system, comprising:
the map construction module is used for constructing according to the acquired public safety field big data resources to obtain a personnel relation map;
the relationship updating module is used for mining personal close contact relationships in real time by adopting a pre-constructed personal close contact relationship dynamic mining algorithm model and updating the personal close contact relationships into the personnel relationship map in real time;
the first probability calculation module is used for calculating the personal infection risk evaluation probability of each person node in the person relation graph based on a person infection risk evaluation regression analysis model obtained through pre-training;
the relation graph dividing module is used for carrying out community structure detection on preset infected persons according to the person relation graph and then carrying out community structure division on the person relation graph to obtain a local person relation graph;
and the second probability calculation module is used for calculating the personal infection risk probability according to the local personnel relationship map and calculating by combining the personal infection risk evaluation probability and the personal infection risk probability to obtain the whole personal infection risk probability.
7. The epidemic propagation tracking system of claim 6, wherein the map building module is specifically configured to: acquiring public safety field big data resources; the public safety field big data resources comprise personnel basic data, address data and key personnel data; performing domain professional knowledge fusion on the public safety domain big data resources through an entity extraction and relationship extraction method; and storing the field professional knowledge by using a graph database Neo4j to construct the human relationship graph.
8. The epidemic propagation tracking system of claim 6, wherein the human infection risk assessment regression analysis model is trained by:
screening all pre-selected risk factors by adopting a Pearson correlation analysis algorithm according to the collected related case sample data to obtain target risk factors of which the absolute values of the correlation coefficients are greater than a preset threshold value;
and training the human infection risk assessment regression analysis model by adopting a Kmeans-Boosting algorithm according to the obtained real human sample data set.
9. The epidemic propagation tracking system of claim 6, wherein the relationship map partitioning module is specifically configured to: taking a preset node corresponding to an infected person as a community structure center; performing path diffusion by adopting a preset community structure division algorithm based on the community structure center; and merging the path nodes meeting the judgment indexes into the community structure by taking the modularity increment as the judgment indexes, and traversing the nodes in the path to obtain a local personnel relationship map corresponding to the preset infected personnel.
10. The epidemic propagation tracking system of claim 6, wherein the second probability calculation module is specifically configured to: calculating personal infection risk probability according to the local personnel relationship map; and calculating a weighted summation value of the personal infection risk evaluation probability and the personal infection risk probability according to a preset weight to obtain the overall personal infection risk probability.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110059223.5A CN112786210B (en) | 2021-01-15 | 2021-01-15 | Epidemic propagation tracking method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110059223.5A CN112786210B (en) | 2021-01-15 | 2021-01-15 | Epidemic propagation tracking method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112786210A true CN112786210A (en) | 2021-05-11 |
| CN112786210B CN112786210B (en) | 2021-10-08 |
Family
ID=75756887
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110059223.5A Active CN112786210B (en) | 2021-01-15 | 2021-01-15 | Epidemic propagation tracking method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112786210B (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113312915A (en) * | 2021-05-28 | 2021-08-27 | 北京航空航天大学 | Intelligent epidemiology investigation system |
| CN113314231A (en) * | 2021-05-28 | 2021-08-27 | 北京航空航天大学 | Infectious disease propagation prediction system and device integrating spatio-temporal information |
| CN113362959A (en) * | 2021-06-03 | 2021-09-07 | 重庆南鹏人工智能科技研究院有限公司 | Sudden respiratory infectious disease risk prediction model for regional epidemic prevention and control |
| CN113658716A (en) * | 2021-07-27 | 2021-11-16 | 之江实验室 | New coronary pneumonia infection crowd analysis method and system based on constrained subgraph calculation |
| CN113743711A (en) * | 2021-06-28 | 2021-12-03 | 航天科工智能运筹与信息安全研究院(武汉)有限公司 | Epidemic situation response benefit evaluation method and device |
| CN113889284A (en) * | 2021-09-16 | 2022-01-04 | 同济大学 | Infectious disease contact target tracking method based on public transport knowledge graph |
| CN114068036A (en) * | 2021-11-18 | 2022-02-18 | 江苏商贸职业学院 | Infection propagation prediction method and system based on Internet of things perception |
| CN114781952A (en) * | 2022-06-23 | 2022-07-22 | 济宁市任城区畜牧兽医事业发展中心(济宁市任城区动物疫病预防控制中心、济宁市任城区动物卫生检疫中心) | Risk early warning method for epidemic disease prevention and control in animal husbandry |
| CN114842980A (en) * | 2022-04-14 | 2022-08-02 | 浙江大学 | Contact tracking pre-screening method for infectious disease susceptible population based on WiFi matching |
| CN115274133A (en) * | 2022-07-15 | 2022-11-01 | 宝鸡市交通信息工程研究所 | Track identification method based on stream modulation big data |
| CN115587593A (en) * | 2022-06-16 | 2023-01-10 | 中关村科学城城市大脑股份有限公司 | Information extraction method and device, electronic equipment and computer readable medium |
| WO2023078082A1 (en) * | 2021-11-02 | 2023-05-11 | International Business Machines Corporation | Determining infection risk levels |
| CN116682574A (en) * | 2023-08-03 | 2023-09-01 | 深圳市震有智联科技有限公司 | Health management method and system for associated crowd |
| WO2023246705A1 (en) * | 2022-06-22 | 2023-12-28 | 清华大学 | Epidemiological investigation data processing method and apparatus, and computer device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180181714A1 (en) * | 2016-12-27 | 2018-06-28 | Cerner Innovation, Inc. | Healthcare System Based on Devices and Wearables |
| CN111178614A (en) * | 2019-12-24 | 2020-05-19 | 成都数联铭品科技有限公司 | Enterprise risk prediction method and system |
| CN111863280A (en) * | 2020-07-30 | 2020-10-30 | 深圳前海微众银行股份有限公司 | Health detection method, system, terminal device and storage medium |
-
2021
- 2021-01-15 CN CN202110059223.5A patent/CN112786210B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180181714A1 (en) * | 2016-12-27 | 2018-06-28 | Cerner Innovation, Inc. | Healthcare System Based on Devices and Wearables |
| CN111178614A (en) * | 2019-12-24 | 2020-05-19 | 成都数联铭品科技有限公司 | Enterprise risk prediction method and system |
| CN111863280A (en) * | 2020-07-30 | 2020-10-30 | 深圳前海微众银行股份有限公司 | Health detection method, system, terminal device and storage medium |
Non-Patent Citations (2)
| Title |
|---|
| 李卫红、童昊昕: "针对非平衡警情数据改进的K-Means-Boosting-BP模型", 《中国图象图形学报》 * |
| 王晶: "具有社团结构的复杂网络上的SIR模型", 《中国优秀硕士学位论文全文数据库（电子期刊）基础科学辑》 * |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113312915A (en) * | 2021-05-28 | 2021-08-27 | 北京航空航天大学 | Intelligent epidemiology investigation system |
| CN113314231A (en) * | 2021-05-28 | 2021-08-27 | 北京航空航天大学 | Infectious disease propagation prediction system and device integrating spatio-temporal information |
| CN113314231B (en) * | 2021-05-28 | 2022-04-22 | 北京航空航天大学 | Infectious disease propagation prediction system and device integrating spatio-temporal information |
| CN113362959A (en) * | 2021-06-03 | 2021-09-07 | 重庆南鹏人工智能科技研究院有限公司 | Sudden respiratory infectious disease risk prediction model for regional epidemic prevention and control |
| CN113743711A (en) * | 2021-06-28 | 2021-12-03 | 航天科工智能运筹与信息安全研究院(武汉)有限公司 | Epidemic situation response benefit evaluation method and device |
| CN113658716A (en) * | 2021-07-27 | 2021-11-16 | 之江实验室 | New coronary pneumonia infection crowd analysis method and system based on constrained subgraph calculation |
| CN113658716B (en) * | 2021-07-27 | 2024-02-23 | 之江实验室 | New coronatine pneumonia infection crowd analysis method and system based on constraint subgraph calculation |
| CN113889284A (en) * | 2021-09-16 | 2022-01-04 | 同济大学 | Infectious disease contact target tracking method based on public transport knowledge graph |
| WO2023078082A1 (en) * | 2021-11-02 | 2023-05-11 | International Business Machines Corporation | Determining infection risk levels |
| CN114068036A (en) * | 2021-11-18 | 2022-02-18 | 江苏商贸职业学院 | Infection propagation prediction method and system based on Internet of things perception |
| CN114842980A (en) * | 2022-04-14 | 2022-08-02 | 浙江大学 | Contact tracking pre-screening method for infectious disease susceptible population based on WiFi matching |
| CN114842980B (en) * | 2022-04-14 | 2023-07-25 | 浙江大学 | WiFi (wireless fidelity) matching-based contact tracking pre-screening method for infectious disease susceptible people |
| CN115587593A (en) * | 2022-06-16 | 2023-01-10 | 中关村科学城城市大脑股份有限公司 | Information extraction method and device, electronic equipment and computer readable medium |
| WO2023246705A1 (en) * | 2022-06-22 | 2023-12-28 | 清华大学 | Epidemiological investigation data processing method and apparatus, and computer device |
| CN114781952A (en) * | 2022-06-23 | 2022-07-22 | 济宁市任城区畜牧兽医事业发展中心(济宁市任城区动物疫病预防控制中心、济宁市任城区动物卫生检疫中心) | Risk early warning method for epidemic disease prevention and control in animal husbandry |
| CN115274133A (en) * | 2022-07-15 | 2022-11-01 | 宝鸡市交通信息工程研究所 | Track identification method based on stream modulation big data |
| CN116682574A (en) * | 2023-08-03 | 2023-09-01 | 深圳市震有智联科技有限公司 | Health management method and system for associated crowd |
| CN116682574B (en) * | 2023-08-03 | 2023-11-24 | 深圳市震有智联科技有限公司 | Health management method and system for associated crowd |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112786210B (en) | 2021-10-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112786210B (en) | Epidemic propagation tracking method and system | |
| Zhang et al. | Big data and artificial intelligence based early risk warning system of fire hazard for smart cities | |
| CN108446470B (en) | Medical facility accessibility analysis method based on vehicle trajectory data and population distribution | |
| Colizza et al. | Epidemic modeling in complex realities | |
| Wiil et al. | Measuring link importance in terrorist networks | |
| CN112232389B (en) | Dynamic adjustment method and system for traffic emergency plan of large-scale activity emergency | |
| Farber et al. | Social interaction potential and the spatial distribution of face-to-face social interactions | |
| Liang et al. | Mapping the knowledge domain of soft computing applications for emergency evacuation studies: A scientometric analysis and critical review | |
| Chen et al. | Examining the importance of built and natural environment factors in predicting self-rated health in older adults: an extreme gradient boosting (XGBoost) approach | |
| Yu et al. | Team recognition in big scholarly data: Exploring collaboration intensity | |
| Arashi et al. | Spatial analysis and prediction of COVID-19 spread in South Africa after lockdown | |
| Hu et al. | Prioritizing disaster mapping tasks for online volunteers based on information value theory | |
| Dong et al. | Non-stationary spatio-temporal point process modeling for high-resolution COVID-19 data | |
| CN114819589A (en) | Urban space high-quality utilization determination method, system, computer equipment and terminal | |
| CN113362959A (en) | Sudden respiratory infectious disease risk prediction model for regional epidemic prevention and control | |
| Yao et al. | Assessing myocardial infarction severity from the urban environment perspective in Wuhan, China | |
| Shao et al. | Incorporating lstm auto-encoders in optimizations to solve parking officer patrolling problem | |
| CN118394873A (en) | AI-based geographical information crowdsourcing data analysis method and system | |
| Yang et al. | Unraveling nonlinear and interaction effects of multilevel built environment features on outdoor jogging with explainable machine learning | |
| Xuan et al. | Collaborative mechanism and simulation of low-carbon travel for residents in community-built environment based on evolutionary game | |
| CN118094304A (en) | City updating method and system based on city physical examination evaluation | |
| Cheng et al. | Optimizing the spatial scale for neighborhood environment characteristics using fine-grained data | |
| Wang et al. | A new flow-based centrality method for identifying statistically significant centers | |
| Lawlor et al. | Stratification structure of urban habitats | |
| Efiong | GIS-based network analysis for optimisation of public facilities closure: A study on libraries in Leicestershire, United Kingdom |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |