Urban management event processing method and device

Abstract

The embodiment of the invention discloses a city management event processing method and device which are used for improving the processing efficiency of city management events, and the method comprises the steps: obtaining event data of city management events from a plurality of data domains, the event data comprising at least one of event time, event places and event related objects; constructing an event feature vector corresponding to each event based on the event data, wherein the elements of the event feature vector comprise event time, event places, event related objects, event evaluations and event keywords; performing similar event analysis and derivative event analysis on each event in the plurality of data fields based on the event feature vectors to construct an event topological relation graph; and performing decision processing based on the event topological relation graph.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of data supervision, and in particular to a city management event processing method and device. BACKGROUND

[0002] In the field of data supervision, the sources of city management events are complex and diverse, and multi-source heterogeneous data is difficult to be aggregated through a single processing rule, resulting in low efficiency of event supervision.

[0003] In actual application, after a city management event occurs, a series of related events may be triggered, which may be reported through different ways, causing the originally associated events to be separated from each other. Classification and processing of these separated events often require a lot of manpower and time cost, and even the same event may be repeatedly disposed of multiple times, resulting in waste of resources and low efficiency of event supervision.

[0004] How to improve the processing efficiency of city management events is a technical problem to be solved by the present application. SUMMARY

[0005] The purpose of the embodiments of the present application is to provide a city management event processing method and device to improve the processing efficiency of city management events.

[0006] In a first aspect, a city management event processing method is provided, comprising: obtaining event data of city management events from a plurality of data domains, the event data comprising at least one of event time, event location, and event involved object; constructing an event feature vector corresponding to each event based on the event data, elements of the event feature vector comprising event time, event location, event involved object, event evaluation, and event keyword; performing similar event analysis and derivative event analysis on each event in the plurality of data domains based on the event feature vector to construct an event topology relationship graph; performing decision processing based on the event topology relationship graph.

[0007] In a second aspect, a city management event processing device is provided, comprising: an obtaining module configured to obtain event data of city management events from a plurality of data domains, the event data comprising at least one of event time, event location, and event involved object; a constructing module configured to construct an event feature vector corresponding to each event based on the event data, elements of the event feature vector comprising event time, event location, event involved object, event evaluation, and event keyword; an analysis module, configured to perform similar event analysis and derivative event analysis on each event in the plurality of data domains based on the event feature vector, to construct an event topology relationship graph; a processing module, configured to perform decision processing based on the event topology relationship graph.

[0008] In a third aspect, an electronic device is provided, which includes a processor, a memory, and a computer program stored in the memory and executable on the processor, and when the computer program is executed by the processor, the steps of the method according to the first aspect are implemented.

[0009] In a fourth aspect, a computer readable storage medium is provided, which stores a computer program, and when the computer program is executed by a processor, the steps of the method according to the first aspect are implemented.

[0010] In a fifth aspect, a computer program product is provided, which includes a non-transitory computer readable storage medium storing a computer program, and the computer program is operable to cause a computer to perform some or all of the steps of the method according to the first aspect.

[0011] In the embodiments of the present application, first, event data of urban management events from a plurality of data domains is acquired, and the event data includes at least one of event time, event location, and event involved object; then, an event feature vector corresponding to each event is constructed based on the event data, and elements of the event feature vector include event time, event location, event involved object, event evaluation, and event keyword; next, similar event analysis and derivative event analysis are performed on each event in the plurality of data domains based on the event feature vector, to construct an event topology relationship graph; finally, decision processing is performed based on the event topology relationship graph. Through the scheme provided in the embodiments of the present application, feature extraction and event analysis can be performed on event data from a plurality of data domains to determine similar events and derivative events, and the relationship between events is explicitly determined through the construction of the event topology relationship graph, thereby effectively improving the efficiency of event decision processing. For event data from different data domains, the scheme provided in the embodiments of the present application represents the key information of events from multiple dimensions through the event feature vector in the manner of feature extraction, and then the relationship recognition effectiveness between events is improved according to the event feature vector, thereby improving the accuracy of the event topology relationship graph, to improve the processing efficiency of urban management events. BRIEF DESCRIPTION OF DRAWINGS

[0012] The accompanying drawings, which are included to provide a further understanding of the present application and are incorporated in and constitute a part of this application, illustrate embodiments of the present application and serve to explain the present application. In the drawings: Figure 1ais one of flow schematic diagrams of a city management event processing method according to an embodiment of the present application; Figure 1b is a city management event event data composition schematic diagram of a city management event processing method according to an embodiment of the present application; Figure 2 is another flow schematic diagram of a city management event processing method according to an embodiment of the present application; Figure 3 is a third flow schematic diagram of a city management event processing method according to an embodiment of the present application; Figure 4a is a fourth flow schematic diagram of a city management event processing method according to an embodiment of the present application; Figure 4b is a fifth flow schematic diagram of a city management event processing method according to an embodiment of the present application; Figure 5a is a sixth flow schematic diagram of a city management event processing method according to an embodiment of the present application; Figure 5b is a seventh flow schematic diagram of a city management event processing method according to an embodiment of the present application; Figure 5c is a simple event context structure schematic diagram of a city management event processing method according to an embodiment of the present application; Figure 5d is a complex event context structure schematic diagram of a city management event processing method according to an embodiment of the present application; Figure 6 is a technical effect comparison schematic diagram between a city management event processing method according to an embodiment of the present application and a traditional method; Figure 7 is a structure schematic diagram of a city management event processing device according to an embodiment of the present application. DETAILED DESCRIPTION

[0013] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application. The drawing numbers in the present application are only used to distinguish each step in the scheme, and are not used to limit the execution order of each step, and the specific execution order is subject to the description in the specification.

[0014] In the urban management application scenario, various data such as news reports, social media, and network comments can provide information basis for urban management. In the event monitoring process, manual event reporting and manual event maintenance are heavily relied on, and the urban management system lacks automatic discovery, automatic tracking, multi-range detection, and automatic maintenance of events. If only the Internet text content is processed, information in other data domains such as social governance and urban management may be missed. Therefore, the multi-source heterogeneous data related to urban management is not fully utilized, and the multi-source heterogeneous data is difficult to be effectively processed by a single rule, which leads to low management efficiency of urban management events.

[0015] To solve the problems in the related art, an embodiment of the present application provides a method for processing urban management events, which can be used in a digital urban management system to automatically discover and monitor urban management events, and is especially suitable for urban governance application scenarios.

[0016] In the urban governance scenario, urban management events emerge in an endless stream, and the sources of various events are diverse, such as sensor monitoring data of production and living equipment, sensor monitoring data of river and lake water level and flow rate, pollutant monitoring data, urban traffic congestion data, citizen complaint data, and network user uploaded data. Among them, the data spread on the Internet is diverse in type and numerous in source, including text, pictures, videos, and other types of data, which can come from individual network users or official news media. Data from different sources may be triggered by an initial event, and continuously spread to form a kind of urban event, and show the context characteristics over time.

[0017] For example, for the problem of urban garbage disposal, some citizens report and complain about garbage disposal, and some citizens spread the topic of garbage disposal on the Internet. The origin of this event may be first monitored by urban management cameras or environmental monitoring equipment. How to quickly extract the context of event development from these multi-source heterogeneous data to automatically summarize related events and realize automatic event sorting is a technical problem to be solved in the field of urban management technology.

[0018] As shown in Figure 1a The scheme provided by an embodiment of the present application includes: S11: Obtain event data of urban management events from multiple data domains, the event data including at least one of event time, event location, and event involved object.

[0019] In the scheme provided by the embodiment of the present application, the event data of the city management event comes from multiple data domains, and the event data of different data domains can come from different data sources. The event data of each data domain can be different, and the scheme provided by the embodiment of the present application can be used to process multi-source heterogeneous event data.

[0020] In the scheme, the event refers to an event that is participated by multiple related roles at a certain time and a certain place, and has one or more actions. The event data obtained in the step includes at least one of the event time, the event place, and the event involved object. In actual application, the event data of different data domains often has different limitations, and the event data of a single data domain can not be able to completely represent all characteristics of the event. In the step, the event data of multiple data domains is obtained, which provides a rich data basis for subsequent event analysis and event context analysis.

[0021] Optionally, as shown in Figure 1b The event data of the city management event from the multiple data domains includes at least one of the following: The collection device data from the Internet of Things data domain includes abnormal alarm data sent by a collection device; The user equipment reporting data from the city user data domain includes city event data sent by a city user through a user equipment; The network information data from the Internet data domain includes network public opinion data associated with city management.

[0022] In the scheme provided by the embodiment of the present application, data of multiple data domains can be obtained, which can come from devices, user reporting or information dissemination platforms, and the event data of multiple data domains can be beneficial to realize smart city data analysis. In the embodiment of the present application, the time data is obtained from the Internet of Things domain, the life grid domain, and the social public opinion domain, and the sensor data, the life grid data, and the network data are analyzed and mined in a unified framework to discover and monitor the city management event.

[0023] The Internet of Things domain can also be referred to as a physical domain, which refers to the device attributes, device locations, and reporting data in the Internet of Things (IOT), which can include monitoring data and alarm data sent by various information sensing devices such as Internet of Things sensors and cameras in the Internet of Things system.

[0024] In the Internet of Things system, the digitalization of products can be represented using a thing model, which can be used to abstract and summarize the functions of products of different brands and different categories, so as to facilitate various parties to describe, control and understand the functions of products in a unified language. The thing model can include, for example, attribute, method and event information, as shown in the following table:

[0025] Among them, the attribute can be used to describe the device state, for example, including the code of the device, the type of the device, the model, the name, the description, the location information of the deployment, the person in charge, the responsible organization, etc. The method of the thing model is the ability or method that the device can be called by the outside. In this scheme, the key information obtained by the Internet of Things device can include the name, the time identifier, the device state, the event type (information, alarm, fault), the output parameter, the description, etc. In actual application, the specific content of the collected device data can be set according to the actual function of the device, so that the collected device data can fully represent the real features related to city management.

[0026] The city user data domain can also be referred to as a life grid domain, which can include data reported by citizens, and can specifically include the time, location, organization, and situation description of an event. Among them, citizens can report in various ways, such as through telephone, network client, questionnaire, etc. The event data of the life grid domain can include, for example, event name, classification, urgency, level, time, location, creator, organization name, cause of the incident, impact range, event status, situation description, resource request, measures taken, on-site photos, etc. In actual application, citizens can flexibly adjust the content of the report according to the actual event situation, so that the city event data sent by the city user through the user device can fully represent the real features related to city management.

[0027] The Internet data domain can propagate various forms of network information data, such as news reports, social data, video and picture comments, etc. The network information data in the Internet data domain includes a large amount of unstructured data, which can be in the form of text, picture, sound, video, etc. Through the scheme, data related to city management events can be extracted from the network information data, such as event title, publication time, publication location, author, content, included pictures, videos, character references, article references, topic references, etc.

[0028] In the scheme provided by the embodiments of the present application, the event data of the city management events of the multiple data domains are obtained, various multi-source heterogeneous fields are related and influenced to each other, can provide event details from different angles, can provide a data basis for subsequent event analysis and context analysis, and are beneficial to objectively reflect the whole picture of the event, thereby being beneficial to improving the efficiency of event processing.

[0029] Based on the event data of multiple data domains, multi-event correlation analysis can be facilitated. In actual applications, events in a certain domain are often related to events in other domains, and by analyzing the relationships between events, the root events and derivative events can be facilitated. Among them, the relationship between events can be analyzed from multiple perspectives, such as time sequence relationship, causal relationship, and inclusion relationship. From the life cycle of events, the generation, development, branching evolution, decline, and extinction of events often trigger the occurrence of cross-domain events. In actual system processing, if only events in a certain domain are analyzed, real and global information cannot be obtained, and the overall situation of events cannot be objectively presented. Therefore, the scheme provided in the embodiments of the present application can facilitate improving the effectiveness of event analysis and improving the efficiency of event processing by obtaining event data of multiple data domains such as physical domains, grid domains, and social domains.

[0030] S12: Construct an event feature vector corresponding to each event based on the event data, and elements of the event feature vector include event time, event location, event involved object, event evaluation, and event keyword.

[0031] In this step, based on the event data, an event feature vector corresponding to each event is constructed. The event feature vector is a multi-dimensional vector, and the elements of the event feature vector include event time, event location, event involved object, event evaluation, and event keyword. Among them, the event time, event location, and event involved object can be generated according to the event time, event location, and event involved object contained in the event data.

[0032] In addition, the event data can be evaluated based on the evaluation rules to generate event evaluation, and the event data can be segmented by text to generate event keywords, so as to construct the above-mentioned event feature vector as elements.

[0033] In this step, in view of the characteristics of diversified form and structure of event data, the event data from multiple data domains is uniformly parsed by feature extraction, and the multi-dimensional event feature vector fully reflects the event features.

[0034] S13: Based on the event feature vector, similar event analysis and derivative event analysis are performed on each event in the multiple data domains to construct an event topology relationship graph.

[0035] Since the event feature vector constructed in the above step can multi-dimensionally present the event features, the association relationship between events is analyzed based on the event feature vector in this step. Specifically, similarity comparison can be performed on different event feature vectors, so that events with a similarity greater than a first similarity are identified as similar events, events with a similarity less than or equal to the first similarity and greater than or equal to a second similarity are identified as derivative events, and events with a similarity less than the second similarity are identified as irrelevant events. The first similarity is greater than the second similarity. In actual application, the analysis rules of similar events and derivative events can be flexibly set according to actual needs.

[0036] Based on the similar event analysis and the derivative event analysis, the scheme provided in the embodiments of the present application can identify the association relationship between multiple events, so as to sort out the event context of multiple events. The event context refers to a tree structure formed by one or more associated events, representing the evolution relationship of events. The event context can be presented in the form of a topological relationship diagram.

[0037] S14: performing decision processing based on the event topological relationship diagram.

[0038] The event topological relationship diagram can fully present the association relationship between multiple events. In this step, the topological relationship diagram can be used to efficiently identify the root event, so as to facilitate batch decision processing of multiple associated events, and further improve the efficiency of city management event processing.

[0039] The scheme provided in the embodiments of the present application can be applied in the city management and governance scenario. For multi-source heterogeneous data such as physical device information, life grid information, and network public opinion information, an event feature vector is constructed according to the characteristics of the data, so as to perform feature analysis and calculation on the multi-source heterogeneous data, to efficiently analyze the association relationship between events, and to achieve the purpose of automatic discovery and automatic supervision of city management events.

[0040] Through the scheme provided in the embodiments of the present application, feature extraction and event analysis can be performed on event data from multiple data sources to determine similar events and derivative events. The relationship between events is explicitly determined through the construction of an event topological relationship diagram, so as to effectively improve the efficiency of event decision processing. For event data from different data domains, the scheme provided in the embodiments of the present application represents the key information of events in multiple dimensions through event feature vectors by means of feature extraction, and then improves the effectiveness of relationship identification between events according to the event feature vectors, so as to improve the accuracy of the event topological relationship diagram and the efficiency of city management event processing.

[0041] Based on the scheme provided in the above embodiments, as shown in Figure 2 In step S12, the event feature vector corresponding to each event is constructed based on the event data, which includes: S21: performing content recognition on the non-text type data in the event data to obtain full-text event data of each event.

[0042] In the embodiments of the present application, preprocessing is performed on the non-text type data in the event data. Among them, the content recognition method is selected according to the actual file type of the event data. For example, image content recognition is used for pictures, and video content recognition is used for videos, so that various types of non-text type data are recognized and processed into text type event data.

[0043] In the data preprocessing process, data with missing main content can be removed, and missing content that can be repaired can be repaired to improve the overall quality of the event data. Optionally, an artificial intelligence model is used to recognize the content of the non-text type data, and the element information contained in the non-text content is extracted to obtain full-text event data.

[0044] S22: performing word segmentation on the full-text event data of each event to construct a keyword vector corresponding to each event, wherein the keyword vector includes a plurality of event keywords obtained by performing word segmentation on the full-text event data.

[0045] In this step, word segmentation is performed on the full-text event data of each event, and the processed data can be represented as a keyword vector:

[0046] Among them,represents the first keyword of the event, represents the second keyword, represents the last keyword of the sentence, and L is the number of keywords in the data. Optionally, after performing word segmentation on the full-text event data, the word segmentation result is filtered to extract keywords with semantics and remove meaningless auxiliary words, so as to highlight the event features expressed by the event keywords and improve the effectiveness of the keyword vector.

[0047] S23: performing statistical rule analysis on the plurality of event keywords obtained by performing word segmentation on the full-text event data of each event to construct an evaluation index vector corresponding to each event, wherein the evaluation index vector includes a plurality of evaluation indexes and corresponding event evaluations.

[0048] The event keywords in the above step can represent the features of the event. In this step, statistical rule analysis is performed on the plurality of event keywords to statistically obtain evaluation indexes that can be used to realize event evaluation and event evaluations corresponding to each evaluation index.

[0049]

[0050] In the embodiments of the present application, the evaluation indexes and event evaluations can be represented in the form of a pointer dictionary. Alternatively, the evaluation indexes are referred to as indexes in the index dictionary, and the event evaluations are referred to as indexes in the pointer dictionary. Different regular expressions are used according to different data formats, and mathematical statistical analysis rules are used as follows: If (rule 1) - elseif (rule 2) - elseif (rule 3) - … - else (rule n) The constructed indexes and indexes are represented in the form of a dictionary as follows:

[0051] wherein, is the index name, is the corresponding index value. The indexes can include general index information of events such as time, place, person, and institution, and urban management indexes including event handling vehicles, event handling personnel, reporting sources, event types, and disposal states.

[0052] Alternatively, in this step, the rule analysis mode can be flexibly selected according to the structure of the event data. For example, for standard structured event data, rule analysis is performed according to the structure of the event data to obtain corresponding indexes and indexes. For unstructured data, indexes and indexes are extracted through named entity recognition, image recognition, and the like.

[0053] S24: Constructing an event feature vector corresponding to each event based on the keyword vector and the evaluation index vector.

[0054] In the scheme provided in the embodiments of the present application, the event data is parsed to construct a keyword vector and an evaluation index vector. The event key features can be reflected from two dimensions of text keywords and feature evaluation indexes. Further, the event feature vector corresponding to the event is constructed based on the keyword vector and the evaluation index vector, which can reflect the features of the event from multiple dimensions and fully and objectively represent the characteristics of the event, thereby providing a data basis for subsequent analysis of the correlation between events.

[0055] Based on the scheme provided in the above embodiments, as shown in Figure 3 In step S23, the multiple event keywords obtained by performing word segmentation on the full-text event data of each event are subjected to statistical rule analysis to construct an evaluation index vector corresponding to each event, including: S31: Performing word frequency statistics on the event keywords of each event to determine multiple common evaluation indexes of each event.

[0056] In this step, word frequency statistics is performed on the event keywords, and keywords with high word frequency are keywords commonly possessed by events. Semantic analysis is performed on keywords with high word frequency, and keywords that can be used as evaluation indexes are extracted to be used as common evaluation indexes of events. If the number of keywords that can be used as evaluation indexes is large, a preset number of keywords can be selected as common evaluation indexes.

[0057] In actual application, the common evaluation indexes can be related to event attributes, such as event time, event location, and event involved objects. Through multiple common evaluation indexes, detailed information of events can be fully presented, and horizontal comparison between events can be facilitated, so that the correlation between events can be determined.

[0058] S32: Perform evaluation on the full-text event data of each event respectively to obtain event evaluation corresponding to the multiple common evaluation indexes respectively.

[0059] In this step, for any event to be evaluated, event evaluation corresponding to the event to be evaluated is determined based on multiple common evaluation indexes. For example, common evaluation indexes and corresponding event evaluation are represented in the form of a dictionary as follows:

[0060] Among them, is a common evaluation index, is the corresponding event evaluation.

[0061] S33: Construct an evaluation index vector by taking the multiple common evaluation indexes and corresponding event evaluation as elements.

[0062] In this step, the multiple common evaluation indexes and corresponding event evaluation are combined as elements to construct an evaluation index vector. The evaluation index vector can be represented as .

[0063] Through the scheme provided in the embodiments of the present application, common evaluation indexes are efficiently determined through text segmentation and word frequency statistics. The common evaluation indexes can be used to effectively evaluate each event, and the evaluation index vector constructed can be used for horizontal comparison between different events, which is beneficial to efficiently determining the correlation between events.

[0064] Based on the scheme provided in the above embodiments, optionally, as shown in Figure 4a before the step S24, that is, before constructing the event feature vector corresponding to each event based on the keyword vector and the evaluation index vector, the method further includes: S41: Determine the timeliness feature value corresponding to each event based on the event time of each event through a decay function.

[0065] The effectiveness of the urban management event is closely related to the timeliness, and the longer the event time is from the current time, the lower the timeliness is. In the scheme provided in the embodiments of the present application, the timeliness of the event is calculated using a decay function, and the timeliness characteristic value is as follows:

[0066] wherein, represents the current time, represents the time when the event occurs, which can be the event time in the event data.

[0067] In the step S24, the event feature vector corresponding to each event is constructed based on the keyword vector and the evaluation index vector, including: S42: The keyword vector, the evaluation index vector, the timeliness characteristic value, the event time, the event location, and the event involved object are taken as elements to construct the event feature vector corresponding to each event.

[0068] In this step, the event features carried in the event data and the features constructed by analysis are taken as elements to construct the event feature vector. For example, the event feature vector (event) can be constructed as follows:

[0069] Wherein, t, loc, figures, org, aging, index, and keys represent the event time, the event location, the event involved person, the event involved organization, the timeliness characteristic value, the evaluation index vector (also referred to as index and index number), and the keyword vector of the event event. The event feature vector model constructed in the scheme can effectively realize the objective and sufficient representation of the urban management event, and serve as an effective basis for subsequent urban management event processing.

[0070] Referring to the flowchart shown in Figure 4b In the scheme provided in the embodiments of the present application, first, data preprocessing is performed on the event data to improve the quality of the processed event data. Then, the keyword vector is constructed, and the index and the index number are constructed. The index and the index number are collectively constructed as the evaluation index vector. Subsequently, the data timeliness is calculated to generate the timeliness characteristic value. Finally, the features in multiple dimensions are taken as elements to construct the event feature vector.

[0071] Through the scheme provided in the embodiments of the present application, the multi-dimensional features of the event can be expressed in the form of the event feature vector, and the event features embodied in multiple different data domains can be fully exhibited, so that the event feature vector can objectively and comprehensively exhibit the event details.

[0072] Based on the scheme provided in the above embodiments, optionally, as Figure 5aAs shown in the above step S13, the similar event analysis and the derivative event analysis are performed on each event in the plurality of data domains based on the event feature vector to construct an event topology relationship graph, including: S51: performing clustering on each event based on the event feature vector to obtain a plurality of event clustering clusters, and each event belonging to the same event clustering cluster is a similar event.

[0073] S52: determining the derivative relationship of each event clustering cluster based on the event feature vector of the center position of each event clustering cluster.

[0074] S53: generating a node representing a similar event based on the plurality of event clustering clusters, and generating a connection line between nodes based on the derivative relationship between the plurality of event clustering clusters to construct an event topology relationship graph.

[0075] Next, the present scheme will be further described in conjunction with an example, referring to Figure 5b the flowchart shown.

[0076] In the scheme provided by the embodiments of the present application, first, event data of a plurality of data domains is acquired, which can include physical domain data, grid domain data, social domain data, etc. Then, the event data is processed for analysis, which can include city management event reporting detection. The detection can acquire the latest reported city management event data in real time through triggering, polling, etc. After detecting the occurrence of newly reported city management event data, the event data is analyzed for feature construction to represent the multi-dimensional features of the event in the form of an event feature vector. Further, multi-domain similar event analysis is performed according to the event feature vector, and multi-domain derivative event analysis is performed according to the event feature vector to determine the correlation between events. The event context is established based on the event correlation, which can be represented in the form of an event topology relationship graph.

[0077] Optionally, to ensure the completeness of the event data processing, it is also possible to determine whether the event data is disposed. If the end condition is met, the monitoring is ended. If the end condition is not met, the city management event reporting is continuously detected.

[0078] Specifically, in the step of city management event reporting detection, the reported events occurring in the physical domain, the grid domain, and the social domain can be detected as seed events for event discovery and monitoring. Optionally, it can include abnormal event alarms and abnormal behavior alarm events issued by various Internet of Things sensor devices in the Internet of Things system; reported events received by the life grid system, the service hotline system, the city management system, the fire safety system, and the Internet field reported public opinion events, etc.

[0079] In the feature construction step, the event data is feature constructed, which can specifically include data preprocessing, keyword vector construction, index and index construction, data timeliness calculation, event feature vector construction, forming a feature vector of each event, for example, represented as: .

[0080] In the multi-domain data similarity analysis step, for the events that have been detected, in the data aggregation system, by using the similarity analysis method, the related events of the event in the physical domain, the grid domain, the social domain, etc. are identified.

[0081] For example, a city fire event, if captured by the camera and sensor of the physical domain, the public opinion on the event in the Internet can be identified from the social domain, and the related data of the event by the organizations such as fire safety in the grid domain.

[0082] The events with similarity greater than the threshold are judged as similar events, otherwise are judged as dissimilar events.

[0083] Event and The similarity is as follows:

[0084] Wherein, is the time-dependent relationship, The similarity of the event location, the event involved object, the event evaluation, and the keyword is calculated respectively. Among them, the similarity of the keyword can be calculated using the following cosine distance:

[0085] Wherein, is the vector value of event , is the vector value of event .

[0086] The time-dependent relationship can be determined in the following way:

[0087] In the derived event analysis step, the clustering algorithm can be used to cluster the above events. Optionally, the vector value of the cluster center point represents the set of corresponding cluster events, and the data in a cluster is regarded as the same event, and the cluster exceeding the threshold is regarded as a derived event.

[0088] In the step of event line construction, the events are arranged in time sequence according to the above-mentioned cluster vector features and time features, to form an event line tree structure diagram, and the events are associated with each other through event identification. The event line (EL) can be represented as a tree structure, for example, Figure 5c A simple event line structure is shown, in which, represents five associated events arranged in time sequence. For another example, Figure 5d A complex event line structure is shown, which includes multiple branches ( ). In practical applications, a node can represent a single event or multiple events combined based on high similarity, and the directed lines connected between the nodes are used to represent the association relationship between the nodes, and the indication direction of the directed lines is used to represent the time sequence of the events. The above-mentioned tree event line can also be called a time topology relationship diagram.

[0089] Alternatively, the event line can also be represented as:

[0090] wherein, represents the i th event, represents the son event of, and n represents the total number of events in the event line.

[0091] In the scheme provided by the embodiments of the present application, the structure of the event line is a tree structure, which reflects the complete life cycle process of the generation, development, evolution and extinction of the event. In order to facilitate the technical personnel to view and subsequent urban management event processing, the event line can be sent to the large screen system or the urban management applet system for display after being constructed.

[0092] Next, the effectiveness of the scheme provided by the embodiments of the present application is verified through the clustering model evaluation index, and the adjacency list of the clustering result is as follows:

[0093] Suppose there are events in the data set, which are , , , , and clusters are obtained through the event discovery method, which are , , , , the number of events belonging to event and clustered into cluster is .

[0094] event and cluster The precision of

[0095] event and cluster The recall of

[0096] wherein, , .

[0097] The F-measure of event and cluster is defined by the corresponding precision and recall:

[0098] event The F-measure of event is defined as the maximum of , wherein The corresponding indicates the cluster label to which the event belongs, and and are defined as the precision and recall of event . Thus, the three evaluation criteria of event , and are respectively represented as

[0099] The three evaluation criteria of the algorithm applied to the entire data set are calculated by weighted average of the evaluation criteria of each event:

[0100]

[0101] Regarding the event mining effect of the present scheme, the following takes five urban management event related data as test data, and takes the precision, recall and F-measure as evaluation criteria to evaluate the effect. The urban management event test data is:

[0102] Figure 6 As an effect comparison diagram, the method provided by the embodiment of the present application can achieve better effect compared with the traditional method. From the test results, since the characteristics of the Internet of Things data and the characteristics of the life grid data are considered in the present scheme, the recall rate of the urban management event is high, and the present scheme can objectively and fully show the various characteristic details of the urban management event.

[0103] The scheme provided by the embodiments of the present application can be applied to the city management and emergency management business in the city management event discovery and monitoring scene of multi-source heterogeneous data, and can analyze and mine event data from multiple data domains. First, according to the characteristics of Internet of Things data, life grid data and Internet data, data features are extracted and calculated by a feature selection method, including time, place, person, institution, timeliness, reference, data index and keyword dictionary, and keyword set. Then, a vector model of event information is constructed according to these features. Next, similarity analysis and calculation of multi-domain data and derived event analysis are performed, and the same event and different events are automatically distinguished, and the parent-child derived relationship of the same event is mined. Finally, an event context is established using a tree structure to reflect the complete life cycle process of event generation, development, evolution and extinction, so as to achieve the closed-loop process of city management event discovery and monitoring.

[0104] The scheme can effectively process multi-source heterogeneous data, extract features from multi-domain, multi-device and multi-data source data of different forms, and express event details, thereby improving the discovery and monitoring effect of city management events. The scheme can analyze event features from multiple dimensions, show the influence of events on city management, and provide more comprehensive data support for city management, which has high application prospect.

[0105] In order to solve the problems in the related art, the embodiments of the present application also provide a city management event processing device 70, as shown in Figure 7 The device comprises: An acquisition module 71 acquires event data of city management events from multiple data domains, and the event data includes at least one of event time, event place and event involved object. A construction module 72 constructs an event feature vector corresponding to each event based on the event data, and the elements of the event feature vector include event time, event place, event involved object, event evaluation and event keyword. An analysis module 73 performs similarity event analysis and derived event analysis on each event in the multiple data domains based on the event feature vector to construct an event topology relationship graph. A processing module 74 performs decision processing based on the event topology relationship graph.

[0106] The device provided by the embodiment of the present application can perform feature extraction and event analysis on event data from multiple data domains to determine similar events and derived events, and can determine the relationship between events by constructing an event topology graph, thereby effectively improving the efficiency of event decision processing. For event data from different data domains, the scheme provided by the embodiment of the present application can represent the key information of events in multiple dimensions by means of feature extraction in the form of an event feature vector, and can improve the effectiveness of relationship recognition between events according to the event feature vector, thereby improving the accuracy of the event topology graph and improving the processing efficiency of urban management events.

[0107] In the device provided by the embodiment of the present application, the above modules can also implement the method steps provided by the method embodiments. Alternatively, the device provided by the embodiment of the present application can also include other modules in addition to the above modules to implement the method steps provided by the method embodiments. The device provided by the embodiment of the present application can achieve the technical effects achieved by the method embodiments.

[0108] Preferably, the embodiment of the present application further provides an electronic device, including a processor, a memory, a computer program stored in the memory and executable on the processor, and the computer program is executed by the processor to implement the processes of the above-mentioned one of the urban management event processing method embodiments, and can achieve the same technical effects. To avoid repetition, this will not be repeated here.

[0109] The embodiment of the present application further provides a computer readable storage medium, and the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the processes of the above-mentioned one of the urban management event processing method embodiments, and can achieve the same technical effects. To avoid repetition, this will not be repeated here. The computer readable storage medium includes, for example, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0110] The embodiment of the present application further provides a computer program product, and the computer program product includes a non-transitory computer readable storage medium storing a computer program, and the computer program is operable to cause a computer to execute some or all of the steps of the above-mentioned one of the urban management event processing method embodiments, and can achieve the same technical effects. To avoid repetition, this will not be repeated here.

[0111] Those skilled in the art will appreciate that embodiments of the present application can be readily used as a method, a system, or a computer program product. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer readable storage media (including, but not limited to, disk memory, CD-ROMs, optical storage devices, etc.) embodying computer readable program code.

[0112] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks or in combination with the flowchart block or blocks.

[0113] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the function specified in the flowchart block or blocks or combination thereof.

[0114] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks or in combination with the flowchart block or blocks.

[0115] In one typical configuration, the computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0116] The memory can include non-persistent memory and / or persistent memory, for example, read only memory (ROM) and / or flash memory, as RAM. The memory is an example of computer readable media.

[0117] Computer-readable media includes permanent and non-permanent, movable and non-movable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible to a computing device. According to the definition herein, computer-readable media does not include transitory media such as modulated data signals and carriers.

[0118] It should also be noted that the terms "comprising", "containing", or any other variant thereof are intended to cover a non-exclusive inclusion, such that a process, method, article or apparatus that comprises a list of elements does not only include those elements, but can also include other elements not expressly listed or inherent to such process, method, article or apparatus. Without more limitations, the element defined by the statement "comprising a" does not exclude the presence of additional identical elements in the process, method, article or apparatus that includes the element.

[0119] Those skilled in the art will appreciate that embodiments of the present application can be provided as a method, system or computer program product. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) containing computer usable program code.

[0120] The above description is only an embodiment of the present application and is not intended to limit the present application. Those skilled in the art can make various modifications and changes to the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the scope of the claims of the present application.

Claims

1. A method for managing an event in a city, characterized by, The method comprises: obtaining event data of urban management events from multiple data domains, the event data comprising at least one of event time, event location, and event involved object; constructing an event feature vector corresponding to each event based on the event data, elements of the event feature vector comprising event time, event location, event involved object, event evaluation, and event keyword; performing similar event analysis and derivative event analysis on each event in the multiple data domains based on the event feature vector to construct an event topology relationship graph; performing decision processing based on the event topology relationship graph.

2. The method of claim 1, wherein, The method of constructing an event feature vector corresponding to each event based on the event data comprises: performing content recognition on non-text type data in the event data to obtain full-text event data of each event; performing word segmentation on the full-text event data of each event to construct a keyword vector corresponding to each event, the keyword vector comprising multiple event keywords obtained by performing word segmentation on the full-text event data; performing statistical rule analysis on the multiple event keywords obtained by performing word segmentation on the full-text event data of each event to construct an evaluation index vector corresponding to each event, the evaluation index vector comprising multiple evaluation indexes and corresponding event evaluations; constructing an event feature vector corresponding to each event based on the keyword vector and the evaluation index vector.

3. The method of claim 2, wherein, The method of performing statistical rule analysis on the multiple event keywords obtained by performing word segmentation on the full-text event data of each event to construct an evaluation index vector corresponding to each event comprises: performing word frequency statistics on the event keywords of each event to determine multiple common evaluation indexes of each event; performing evaluation on the full-text event data of each event respectively to obtain event evaluations corresponding to the multiple common evaluation indexes respectively; constructing an evaluation index vector with the multiple common evaluation indexes and corresponding event evaluations as elements.

4. The method of claim 2, wherein, Before constructing an event feature vector corresponding to each event based on the keyword vector and the evaluation index vector, the method further comprises: determining a timeliness feature value corresponding to each event based on the event time of each event by a decay function; wherein the method of constructing an event feature vector corresponding to each event based on the keyword vector and the evaluation index vector comprises: constructing an event feature vector corresponding to each event with the keyword vector, the evaluation index vector, the timeliness feature value, the event time, the event location, and the event involved object as elements.

5. The method according to any one of claims 1 to 4, wherein The method of performing similar event analysis and derivative event analysis on each event in the multiple data domains based on the event feature vector to construct an event topology relationship graph comprises: performing clustering on each event based on the event feature vector to obtain multiple event clustering clusters, each event in the same event clustering cluster being similar events; determining a derivative relationship of each event clustering cluster based on the event feature vector of the center position of the event clustering cluster; generating nodes representing similar events based on the multiple event clustering clusters and generating inter-node connections based on the derivative relationship between the multiple event clustering clusters to construct an event topology relationship graph.

6. The method according to any one of claims 1 to 4, wherein The event data of urban management events from multiple data domains comprises at least one of the following: The collection device data from the Internet of Things data domain includes abnormal alarm data sent by a collection device. The user equipment report data from the urban user data domain includes urban event data sent by an urban user through a user equipment. The network information data from the Internet data domain includes network public opinion data associated with urban management.

7. An urban management event processing apparatus characterized by comprising: The method comprises: An acquisition module acquires event data of urban management events from multiple data domains, the event data including at least one of event time, event location, and event involved object; A construction module constructs an event feature vector corresponding to each event based on the event data, elements of the event feature vector including event time, event location, event involved object, event evaluation, and event keyword; An analysis module performs similar event analysis and derivative event analysis on each event in the multiple data domains based on the event feature vector to construct an event topology relationship graph; A processing module performs decision processing based on the event topology relationship graph.

8. An electronic device, comprising: The computer program is stored in the memory and executable on the processor, and when executed by the processor, the steps of the method according to any one of claims 1 to 6 are implemented. The computer program is stored in the memory and executable on the processor, and when executed by the processor, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium, characterized in that, The computer program product comprises a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to perform the steps of the method according to any one of claims 1 to 6.

10. A computer program product, characterised in that, ​

Images