A method and system for event discovery and presentation based on knowledge graph

Abstract

The present invention discloses an event discovery and display method based on knowledge graph, including: S100, pre-processing each government service hotline data; S200, dividing the government service hotline data in the time period selected by the user into data frames according to the time interval specified by the user, performing event detection on the government service hotline data in each frame through the community discovery algorithm, and obtaining each city event; S300, matching the events of the two frames before and after based on the maximum weight matching algorithm of the bipartite graph, establishing the event chain of the evolution of each city event, analyzing the z-score value of each city event in the event chain, determining the suddenness and urgency of each city event, and performing early warning and multi-dimensional visualization of the city sudden events. The present invention can clearly understand various events occurring in the city, and can detect and warn the sudden events in the city.

Description

Technical Field

[0001] The present invention relates to the field of knowledge graph application and visualization display, and specifically to an event discovery and display method and system based on knowledge graph. Background Art

[0002] Citizens usually report various urban problems in the form of reports or complaints. Therefore, government service hotline data is a concentrated reflection of various urban problems, including various aspects of multi-granular urban problems, and also reflects citizens' subjective feelings about urban problems. Government service hotline data generally consists of complaint time, complaint content, receiving department, etc. Every government working day, a large number of citizen hotlines are received. In the past, government service hotline data was mostly processed manually by government personnel, which was inefficient and could not dig out the true value of the data set. Summary of the invention

[0003] In view of the defects existing in the prior art, the purpose of the present invention is to provide an event discovery and presentation method and system based on knowledge graph. Combining the domain knowledge model for government affairs to study the government service hotline data helps to better parse and enrich the semantic information of each 12345 hotline data, so as to timely and clearly understand the various events occurring in the city, and detect and warn of emergencies in the city.

[0004] To achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0005] A method for event discovery and presentation based on knowledge graph, comprising:

[0006] S100, preprocessing each piece of government service hotline data, including extracting time information and spatial information of each piece of government service hotline data, vectorizing each piece of government service hotline data, and adding time information and keyword information of the government service hotline data into the corresponding government service hotline data;

[0007] S200, dividing the government service hotline data within the time period selected by the user into data frames according to the time interval specified by the user, performing event detection on the government service hotline data in each frame through a community discovery algorithm, and obtaining events of each city, wherein each city event includes non-emergency events and emergency events of the city;

[0008] S300, based on the maximum weight matching algorithm of the bipartite graph, matches the events of the previous and next frames, establishes the event chain of the evolution of each city event, analyzes the z-score value of each city event in the event chain, determines the suddenness and urgency of each city event, and performs early warning and multi-dimensional visualization of the city sudden events.

[0009] Furthermore, in the method described above, in S100, the time information of each piece of government service hotline data is obtained through a regular expression or a call time.

[0010] Further, in the above method, in S100, extracting the spatial information of each piece of government service hotline data includes: extracting the location information of each piece of government service hotline data based on the Lattice LSTM model,

[0011] If one or more continuous location information is extracted, these location information are spliced ​​into complete location information and then converted into longitude and latitude information as the spatial information of the current government service hotline data;

[0012] If multiple locations with close distances are extracted, one of the locations with easy-to-retrieve longitude and latitude information is selected and converted into longitude and latitude information as the spatial information of the current government service hotline data;

[0013] If two locations with a long distance between them are extracted, one of the locations is selected and converted into longitude and latitude information as the spatial information of the current government service hotline data.

[0014] Further, in the above method, in S100, each government service hotline data is represented by vectorization, including:

[0015] Based on the Word2Vec model, a vector is used to represent each word in a piece of government service hotline data. Based on the TF-IDF value of each word, the word vectors are weighted and summed to obtain the text semantic representation of the government service hotline data.

[0016] Further, in the above method, in S200, event detection is performed on the government service hotline data in each frame by using a community discovery algorithm to obtain events in each city, including:

[0017] Based on the community discovery algorithm, text semantic clustering is performed on the government service hotline data in each frame, and the government service hotline data describing the same event are annotated with the same label to obtain different types of events in each city.

[0018] Furthermore, in the above method, in S300, based on the bipartite graph maximum weight matching algorithm, the events of the two frames are matched to establish the event chain of the event evolution of each city, including:

[0019] The event found in the previous frame is placed on the left side of the bipartite graph, and the event found in the next frame is placed on the right side of the bipartite graph. The weights between the nodes of the bipartite graph are determined by the semantic similarity of the events on the left and right sides.

[0020] If a node is not matched, it is defined as the start of a new event or the end of an old event.

[0021] Further, in the method as described above, S300 includes:

[0022] Through parameter configuration, multi-dimensional visualization of urban emergencies can be achieved, including locating the spatial distribution of different types of events on the map, showing the changes in the number of different events over time on the line graph, representing the degree of emergency of events on the event chain river graph, showing the importance of related elements of a certain type of event through word cloud graph, and observing urban events from global or selected local multi-perspectives, helping urban decision makers to determine the basic information and early warning information of urban emergencies in a variety of visualization forms. The basic information includes event spatiotemporal information and event content information.

[0023] An event discovery and presentation system based on knowledge graph, including:

[0024] A preprocessing module is used to preprocess each piece of government service hotline data, including extracting time information and spatial information of each piece of government service hotline data, vectorizing each piece of government service hotline data, and adding time information and keyword information of government service hotline data into corresponding government service hotline data;

[0025] An acquisition module is used to divide the government service hotline data within the time period selected by the user into data frames according to the time interval specified by the user, perform event detection on the government service hotline data in each frame through a community discovery algorithm, and acquire events of each city, wherein each city event includes non-emergency events and emergency events of the city;

[0026] The visualization module is used to match the events of the previous and next frames based on the maximum weight matching algorithm of the bipartite graph, establish the event chain of the evolution of each city event, analyze the z-score value of each city event in the event chain, determine the suddenness and urgency of each city event, and provide early warning and multi-dimensional visualization of the city emergencies.

[0027] A storage medium stores a computer program, wherein the computer program is configured to execute the method when running.

[0028] An electronic device comprises a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to execute the method.

[0029] The beneficial effects of the present invention are that the present invention can better parse and enrich the semantic information of each 12345 hotline data, so as to timely and clearly understand various events occurring in the city, and can detect and warn of emergencies in the city, and finally trace and analyze the causes and solutions of urban problems in the domain knowledge model oriented to government affairs, which helps the relevant government departments to take measures to reduce the losses caused by these events to citizens before these events become more serious. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 A schematic diagram of a process of an event discovery and presentation method based on a knowledge graph provided in an embodiment of the present invention;

[0031] Figure 2 A schematic diagram of the structure of an event discovery and presentation system based on a knowledge graph provided in an embodiment of the present invention;

[0032] Figure 3 A framework flow chart of an event discovery and presentation method based on a knowledge graph provided in an embodiment of the present invention;

[0033] Figure 4 A global display diagram of events in city A provided in an embodiment of the present invention;

[0034] Figure 5 A global display diagram of events in city B provided in an embodiment of the present invention;

[0035] Figure 6 This is a single display diagram of events in city A provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0036] In order to make the technical problems solved by the present invention, the technical solutions adopted and the technical effects achieved more clearly, the technical solutions of the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0037] The present invention provides a method for event discovery and presentation based on knowledge graph. Figure 1 As shown, the method includes:

[0038] S100. Preprocess each piece of government service hotline data, including extracting the time information and spatial information of each piece of government service hotline data, vectorizing each piece of government service hotline data, and adding the time information and keyword information of the government service hotline data into the corresponding government service hotline data.

[0039] In the embodiment of the present invention, in order to discover urban emergencies, it is necessary to obtain spatiotemporal information. The government service hotline data is recorded manually, and the time information contained in the data is recorded in a relatively standardized manner. The time information of each government service hotline data can be obtained by regular expressions or by approximating the call time. The main difficulty is that for government service hotline record data of different location modes, correct and appropriate location information needs to be extracted. In order to solve the problem of location extraction, after comprehensively comparing the effect indicators of the rule-based method and the three location extraction methods of Bi-LSTM and Lattice LSTM based on deep learning, it is found that the Lattice LSTM model is different from other NER methods. It makes full use of the context information of the document to find more suitable words, and makes full use of the explicit word information instead of only using the information of the character annotation sequence. Therefore, it can effectively reduce the impact of the word segmentation error in the Chinese document, so it has achieved the best effect. Finally, this method is selected to extract the spatial information of each government service hotline data. At the same time, since the complaint content may contain multiple addresses, the extracted addresses are also multiple. In order to obtain a unique and suitable address, we process it according to the rules in three cases: if the extracted location is one or more continuous locations, just splice these locations into a complete location; if the multiple extracted locations are close, select one of the locations that is easy to retrieve the longitude and latitude as the location information of the information; if there are two locations that are far apart, such as the pick-up and drop-off locations of taxis, in order to handle them consistently with other city issues, a unified regulation is adopted to select one of them as the location information. Finally, the location information is converted into longitude and latitude information through the API provided by Baidu, which is convenient for marking the location of the incident on the map.

[0040] In an embodiment of the present invention, the time information of each piece of government service hotline data is obtained through a regular expression or a call time. In S100, the spatial information of each piece of government service hotline data is extracted, including: based on the Lattice LSTM model, the location information of each piece of government service hotline data is extracted. If one or more continuous location information is extracted, these location information are spliced ​​into complete location information and then converted into longitude and latitude information as the spatial information of the current government service hotline data; if multiple location information with a close distance is extracted, one of the location information with easy longitude and latitude retrieval is selected and converted into longitude and latitude information as the spatial information of the current government service hotline data; if two location information with a long distance is extracted, one of the location information is selected and converted into longitude and latitude information as the spatial information of the current government service hotline data. In S100, each piece of government service hotline data is vectorized, including: based on the Word2Vec model, a vector is used to represent each word in a piece of government service hotline data, and based on the TF-IDF value of each word, the word vector is weighted and summed to obtain the text semantic representation of the government service hotline data.

[0041] S200. Divide the government service hotline data within the time period selected by the user into data frames according to the time interval specified by the user, perform event detection on the government service hotline data in each frame through the community discovery algorithm, and obtain each city event, which includes urban non-emergency events and urban emergency events.

[0042] In the embodiment of the present invention, citizens often reflect not trivial matters in life through the government service hotline, but more important things happening in the city. Government service hotlines with similar descriptions often reveal the occurrence of a city event. How to classify the problems reported by citizens in the city into multiple types according to their semantics, and each type corresponds to a city event type and represents them is a great difficulty. The present invention represents a city event with a set of government service hotlines with similar descriptions. First, for each hotline data, the Word2Vec model is used to vectorize each word in the hotline data, and then the word vectors are weighted and summed based on the TF-IDF value of each word to obtain the document semantic representation of the hotline data. Then, the text semantic clustering is performed based on the Louvain community discovery algorithm. The Louvain community discovery algorithm relies on the following greedy process. At the beginning, each node is a community. Then, the nodes in the network are traversed in a certain order. For each neighbor node j of node i in the network, try to transfer node i to the community where j is located, and calculate the modularity gain (the index of clustering defined by the algorithm) generated by the process. Node i eventually moves to the neighbor node with the highest modularity gain, and the modularity gain must be positive. Then, after the network is reconstructed in the previous step, the nodes in the same community are reconstructed into a super node, and the above two steps are repeatedly performed until the algorithm is stable. Based on the above algorithm, the government service hotline data describing the same event will eventually be labeled with the same label. The present invention adopts a community discovery algorithm to perform cluster analysis on the government service hotline data to complete the detection of events occurring within the selected time period.

[0043] In an embodiment of the present invention, in S200, event detection is performed on the government service hotline data in each frame through a community discovery algorithm to obtain events in each city, including: based on the community discovery algorithm, text semantic clustering is performed on the government service hotline data in each frame, and the government service hotline data describing the same event are labeled with the same label to obtain different types of city events.

[0044] S300, based on the maximum weight matching algorithm of the bipartite graph, the events of the previous and next frames are matched, and the event chain of the evolution of each city event is established. Based on the z-score value of each city event in the event chain, the suddenness and urgency of each city event are determined, and the sudden city events are warned and multi-dimensionally visualized.

[0045] In the embodiment of the present invention, urban events include non-emergency events and emergencies. When the frequency of a certain type of event within a certain time and space range is significantly higher than the historical statistical average, we call it an urban emergency. If a large number of citizens report the same event through the government service hotline in a short period of time, then the event may be an urban emergency. In order to effectively warn of urban emergencies, the present invention divides the data into data frames at a certain time interval and performs event detection in each frame. After that, we need to mark the government service hotline data describing the same event in the previous and next frames with the same label, which is a matching problem. The matching problem is usually solved by a bipartite graph algorithm. The event found in the previous frame is located on the left side of the bipartite graph, and the event found in the next frame is located on the right side of the bipartite graph. The weights between the nodes of the bipartite graph are determined by the semantic similarity of the events on the left and right sides. Finally, the maximum weight matching algorithm of the bipartite graph is used to match the events of the previous and next frames. If a node is not matched, it is defined as the beginning of a new event or the end of an old event. By analyzing the changes in the number of government service hotlines in each frame of the event chain matched in the previous step, urban events in the entire city or local areas of the city are sorted according to the z-score value to characterize the suddenness and urgency of urban events. The government service hotline data describing the same event are identified with the same ID, and early warnings are issued for urban emergencies to remind city managers to pay attention to such problems and resolve them as soon as possible.

[0046] In the embodiment of the present invention, by analyzing the government service hotline data to realize the city event discovery and event warning functions, it is possible to detect the sudden events in the city and issue a timely warning. Through parameter configuration, multi-dimensional visualization of events is achieved, including locating the spatial distribution of different types of events on Baidu Maps, representing the changes in the number of different events over time on a line graph, representing the suddenness of events on an event chain river graph, displaying the importance of related elements of a certain type of event through a word cloud graph, and observing urban events from a global or self-selected local multi-perspective, and using a variety of visualization forms to help urban decision makers determine the main content of urban emergencies, the main locations of occurrence and other information. The multi-dimensional visualization display technology for early warning of urban emergencies is of great significance to urban problem governance, urban construction and development.

[0047] The method of the embodiment of the present invention helps to better parse and enrich the semantic information of each 12345 hotline data, so as to timely and clearly understand various events occurring in the city, detect and warn of emergencies in the city, and finally trace and analyze the causes and solutions of urban problems in the domain knowledge model for government affairs, which helps the relevant government departments to take measures to reduce the losses caused by these events to citizens before these events become more serious. The design and implementation of the city event warning system based on government service hotline data is of great significance to both the government and citizens.

[0048] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the present invention is not limited by the described action sequence, because according to the present invention, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0049] According to another aspect of the embodiment of the present invention, a knowledge graph-based event discovery and presentation system is also provided. Figure 2 As shown, including:

[0050] A preprocessing module 100 is used to preprocess each piece of government service hotline data, including extracting time information and spatial information of each piece of government service hotline data, vectorizing each piece of government service hotline data, and adding time information and keyword information of the government service hotline data into the corresponding government service hotline data;

[0051] The acquisition module 200 is used to divide the government service hotline data in the time period selected by the user into data frames according to the time interval specified by the user, perform event detection on the government service hotline data in each frame through the community discovery algorithm, and obtain each city event, wherein each city event includes a city non-emergency event and a city emergency event;

[0052] The visualization module 300 is used to match the events of the previous and next frames based on the maximum weight matching algorithm of the bipartite graph, establish the event chain of the evolution of each city event, analyze the z-score value of each city event in the event chain, determine the suddenness and urgency of each city event, and provide early warning and multi-dimensional visualization of the city sudden events.

[0053] It should be noted that the system of the present invention and the aforementioned method belong to the same technical concept, and the specific implementation method will not be repeated here.

[0054] The present invention can be used for large-scale and efficient data import in a distributed knowledge graph system, and supports the import of super node data. The following application example of sending batch import requests from multiple application clients further illustrates the implementation mode of the present invention.

[0055] As long as the city's government service hotline data contains specific complaint content and complaint time information, it can be applied to the system of the present invention. The overall process is as follows: Figure 3As shown, it includes: S1. Preprocessing is performed on each piece of government service hotline data: the location of the event in the complaint content is extracted through the Lattice LSTM model, the complaint content is vectorized through the Word2Vec technology to facilitate subsequent machine processing, and then the complaint time information and keyword information representing the complaint content are added to complete the preprocessing of each 12345 hotline data; S2. The system will then accept parameter input from the user, divide the data in the time period selected by the user into data frames according to the time interval specified by the user, and monitor the data in each data frame through the Louvain-based community discovery clustering algorithm to divide the data into different event categories; S3. Then, the bipartite graph matching algorithm is used to establish the connection between the previous and next frame events to obtain the event chain of the evolution of each city event. Finally, the system uses the z-score value of the event, that is, the value that measures the suddenness of the event, to distinguish whether the city event is an emergency event, and present it through rich visualization technology.

[0056] In the system parameter list, you can select the display area of ​​the system, which is divided into global and local; you can select the time period and time interval you want to view; you can also choose to view the details of a certain type of event. The main display interface consists of three modules, namely the map module, the line graph module and the event chain river graph module. In the map module, the complaints within the selected time period are marked with points on the map according to their occurrence locations. The color of the points represents the event category to which they belong. Select a specific point to see its detailed complaint content; the line graph module shows the change curve of the number of 12345 hotlines over the time interval within the selected time period; the event chain river graph module shows the suddenness of each category of events within the selected time period in the form of a river graph. The higher the level, the higher the frequency of this type of event within this time range, and the corresponding problems need to be focused on and solved in a timely manner.

[0057] Take the urban emergency warning system based on the government service hotline data of City A and the urban emergency warning system based on the government service hotline data of City B as examples: the data used by the event system of City A is the government service hotline data of City A from September 2016 to April 2020. Figure 4 The global display interface of the city emergency warning system is shown from January 1, 2020 to March 1, 2020. It can be seen from the figure that during these two months, the number of events related to the epidemic increased sharply, and the emergency level was the highest, which required the government to pay attention and provide corresponding solutions, which was in line with the actual situation at the time. The data used by the event system of City B is the government service hotline data of City B from January 2019 to March 2021. Figure 5What is shown is the global display interface of the city emergency warning system from November 1, 2020 to January 1, 2021. It can be found that the demand for heating by citizens has increased sharply in these two months, which requires the focus of relevant government departments, which is also in line with the actual situation at the end of the year. The local display is similar to the global display, except that the scope of attention is narrowed from the city as a whole to the area selected by the user, which will not be repeated here. This system also provides a single display of a certain type of event in the city. On the basis of the global display or local display, select a certain type of event that you want to understand in depth, enter the event sequence in the parameter list, and you can enter the single display interface of the city event. Compared with the global / local display, the map module and the line chart module are retained; the difference is that first, only the selected category of events are displayed, and secondly, the event chain river map is removed, and replaced with a word cloud display map in different time intervals and a specific reporting content table, so as to better display a single event class. Figure 6 It is shown in Figure 2 Based on the global event display of city A, the interface for displaying epidemic events separately mainly shows that the layout of the single display interface is different from the global one, rather than the text content.

[0058] The system using the embodiment of the present invention helps to better parse and enrich the semantic information of each 12345 hotline data, so as to timely and clearly understand various events occurring in the city, detect and warn of emergencies in the city, and finally trace and analyze the causes and solutions of urban problems in the domain knowledge model for government affairs, which helps the relevant government departments to take measures to reduce the losses caused by these events to citizens before these events become more serious. The design and implementation of the city event warning system based on government service hotline data is of great significance to both the government and citizens.

[0059] The present invention also provides a storage medium, in which a computer program is stored, wherein the method of the present invention can be executed when the computer program is run. The storage medium includes at least one of the following: a floppy disk, an optical disk, a DVD, a hard disk, a flash memory, a USB flash disk, a CF card, an SD card, an MMC card, a SM card, a memory stick, an xD card, etc. The method of the present invention is converted into data (computer program) and recorded on the above storage medium, for example, the hard disk with the computer program of the method of the present invention is placed in a computer and run, so that the method of the present invention can be implemented.

[0060] The present invention also provides an electronic device, including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to execute the method of the present invention. The memory belongs to the aforementioned storage medium, and can store the computer program of the method of the present invention. The processor can process the data in the memory, and the electronic device can be a computer, a mobile phone, or any other device including a memory and a processor. After the computer is started, the processor is started to run the computer program of the method of the present invention in the memory, and the method of the present invention can be implemented.

[0061] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for event discovery and presentation based on knowledge graph, characterized in that: include: S100, preprocessing each piece of government service hotline data, including extracting time information and spatial information of each piece of government service hotline data, vectorizing each piece of government service hotline data, and adding time information and keyword information of the government service hotline data into the corresponding government service hotline data; S200, dividing the government service hotline data within the time period selected by the user into data frames according to the time interval specified by the user, performing event detection on the government service hotline data in each frame through a community discovery algorithm, and obtaining various city events, specifically including: based on the community discovery algorithm, performing text semantic clustering on the government service hotline data in each frame, annotating the government service hotline data describing the same event with the same label, and obtaining different types of various city events, wherein the various city events include urban non-emergency events and urban emergency events; S300. Based on the maximum weight matching algorithm of the bipartite graph, the events of the previous and next frames are matched, an event chain of the evolution of each city event is established, the z-score value of each city event in the event chain is analyzed, the suddenness and urgency of each city event is determined, and early warning and multi-dimensional visualization of the city sudden events are performed; wherein the event chain of the evolution of each city event is established specifically includes: the event found in the previous frame is located on the left side of the bipartite graph, and the event found in the next frame is located on the right side of the bipartite graph, and the weight between the nodes of the bipartite graph is determined by the semantic similarity of the events on the left and right sides; if a node is not matched, it is defined as the beginning of a new event or the end of an old event.

2. The method according to claim 1, characterized in that In S100, the time information of each government service hotline data is obtained through regular expressions or call time.

3. The method according to claim 1, characterized in that In S100, the spatial information of each piece of government service hotline data is extracted, including: based on the Lattice LSTM model, the location information of each piece of government service hotline data is extracted, If one or more continuous location information is extracted, these location information are spliced ​​into complete location information and then converted into longitude and latitude information as the spatial information of the current government service hotline data; If multiple locations with close distances are extracted, one of the locations with easy-to-retrieve longitude and latitude information is selected and converted into longitude and latitude information as the spatial information of the current government service hotline data; If two locations with a long distance between them are extracted, one of the locations is selected and converted into longitude and latitude information as the spatial information of the current government service hotline data.

4. The method according to claim 1, characterized in that: In S100, each government service hotline data is represented by vectorization, including: Based on the Word2Vec model, a vector is used to represent each word in a piece of government service hotline data. Based on the TF-IDF value of each word, the word vectors are weighted and summed to obtain the text semantic representation of the government service hotline data.

5. The method according to any one of claims 1 to 4, characterized in that: S300 includes: Through parameter configuration, multi-dimensional visualization of urban emergencies can be achieved, including locating the spatial distribution of different types of events on the map, showing the changes in the number of different events over time on the line graph, representing the degree of emergency of events on the event chain river graph, showing the importance of related elements of a certain type of event through word cloud graph, and observing urban events from global or selected local multi-perspectives, helping urban decision makers to determine the basic information and early warning information of urban emergencies in a variety of visualization forms. The basic information includes event spatiotemporal information and event content information.

6. An event discovery and presentation system based on knowledge graph, characterized in that: include: A preprocessing module is used to preprocess each piece of government service hotline data, including extracting time information and spatial information of each piece of government service hotline data, vectorizing each piece of government service hotline data, and adding time information and keyword information of government service hotline data into corresponding government service hotline data; An acquisition module is used to divide the government service hotline data within the time period selected by the user into data frames according to the time interval specified by the user, perform event detection on the government service hotline data in each frame through a community discovery algorithm, and acquire events of each city, wherein each city event includes non-emergency events and emergency events of the city; The visualization module is used to match the events of the previous and next frames based on the maximum weight matching algorithm of the bipartite graph, establish the event chain of the evolution of each city event, analyze the z-score value of each city event in the event chain, determine the suddenness and urgency of each city event, and provide early warning and multi-dimensional visualization of the city emergencies.

7. A storage medium, characterized in that: The storage medium stores a computer program, wherein the computer program is configured to execute the method according to any one of claims 1 to 5 when executed.

8. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to run the computer program to perform the method according to any one of claims 1 to 5.

Images