A social event processing method applied to the field of comprehensive governance

Abstract

This invention relates to a method for handling social incidents in the field of public security. It solves the problem of low efficiency in handling urban social incidents in existing technologies. It includes: S1, reporting incident content using the RE2 model and matching it with the text of incident content in a case library; incidents that meet the threshold matching condition are then processed in S3, referring to the incidents matched in the case library; S2, incidents that do not meet the threshold matching condition are identified according to a first-level component / event open classification model; S3, incidents from S1 or those determined to be first-level component / event categories in S2 are transmitted to the incident handling center, where a unified interface and data structure for receiving incidents are created; the RE2 model is used to compare the similarity of the received incidents with the problem description text in the second-level component / event management profile library to ensure the accuracy of incident classification. The advantage of this invention is that it effectively improves the efficiency of social incident handling.

Description

Technical Field

[0001] This invention relates to the field of social governance technology, specifically to a method for handling social incidents applied in the field of comprehensive social order management. Background Technology

[0002] The general process for handling incidents in the field of public security includes accepting the incident, responding to the handling results, following up, and conducting a satisfaction survey. The scope of incidents handled covers all aspects of society, including infrastructure, urban appearance, market supervision, street security, community management, and rural management. The departments handling incidents include the Political and Legal Affairs Commission, public security, market supervision, economic and information technology, administrative law enforcement, insurance, telecommunications, and other departments and units, as well as companies such as gas, water, telecommunications, and mobile. In some cases, reported social incidents require cross-departmental collaboration. In cases of multi-departmental collaboration, it is necessary to identify the responsible department, which will follow up and supervise the handling status of the incident.

[0003] Currently, the handling of various social events faces the following problems:

[0004] I. Inadequate reporting channels for incidents; people are unsure how to report problems after they are discovered.

[0005] Second, the business acceptance and handling process is not standardized, each department acts independently, and there is no unified method for assigning and accepting cases, which leads to frequent instances of mutual buck-passing and cases ending without resolution during the handling process.

[0006] Third, the lack of classification and grading of events makes it impossible to determine the priority of events when dealing with multiple events.

[0007] Fourth, there is a lack of necessary oversight and feedback on the incidents; those in charge are unaware of the progress of the incidents, and the public is unaware of the outcomes.

[0008] Most of the existing solutions for handling social problems focus on a specific step in the process, and the process is relatively simple. They do not classify, grade, or divert events. Moreover, many relevant departments currently have their own unique event collection and processing systems. The diversity of these systems leads to incompatibility issues, which affects the real-time nature of problem-solving and reduces the efficiency of handling social events. Summary of the Invention

[0009] The purpose of this invention is to address the above-mentioned problems by providing a social incident handling method that is rationally designed and highly efficient for application in the field of public security.

[0010] To achieve the above objectives, the present invention adopts the following technical solution: a method for handling social incidents in the field of public security, characterized in that the method specifically includes the following steps:

[0011] S1. The event content is reported using the RE2 model and matched with the event content text in the case library for similarity. Events that meet the threshold matching conditions are entered into S3 for assignment processing, referring to the events matched in the case library. By using the RE2 model to report events and compare them with cases in the case library for assignment, the event category can be determined, which can effectively improve the accuracy of data.

[0012] S2. Events that do not meet the threshold matching conditions are identified according to the first-level component / event open-set text classification model.

[0013] S3. Events from S1 or those identified as first-level components / events in S2 are transmitted to the event reception center to create a unified interface and data structure for receiving events. The unified interface and data structure can make the reception of social events uniform and facilitate management.

[0014] S4. Based on the primary component / event category of the current event, retrieve all secondary component / event management images under the current primary component / event category from the secondary component / event management image library. Use the RE2 model to detect the similarity between the current event content and the problem description text in the secondary component / event management. Use the RE2 model to compare the problem description text similarity between the received event and the images in the secondary component / event management image library to ensure the accuracy of event classification.

[0015] S5. Create a virtual network for the event, and add the responsible department to the grid as a grid member according to the secondary component / event management profile. Then, determine whether the current event is a public welfare event based on the binary classification algorithm model. If it is a public welfare event, extract the information of the parties involved from the current event using the named entity extraction method and add it to the virtual grid.

[0016] S6. Follow up on the handling results and conduct satisfaction surveys. Satisfaction surveys help optimize the handling of social incidents and store high-satisfaction handling methods for easy reference.

[0017] In the aforementioned method for handling social incidents applied to the field of public security, step S1, the establishment of the RE2 model specifically includes the following steps:

[0018] S11. The input includes the current event event1 and the case library. For each event content event2 in the case library, perform the following steps:

[0019] S12. Convert the text events1 and event2 into vectors in embedding form;

[0020] S13. The embedding vector sequence is fed into three blocks for continuous processing. Each block has the same structure, consisting of an encoder layer, an alignment layer, and a fusion layer. The encoder layer calculates the context features of the sequence, the alignment layer calculates the relevance features of the text event1 and event2, and the output of the fusion layer is the input of the next block or the input of the pooling layer.

[0021] S14. The pooling layer converts the output of the block into a fixed-length vector.

[0022] After calculating the features of S15, event1, and event2, a binary classification model is constructed in the prediction layer. Category 0 indicates that event1 and event2 are not similar, and category 1 indicates that event1 and event2 are similar.

[0023] In the aforementioned method for handling social incidents applied to the field of comprehensive social management, in step S2, the first-level component / event open classification model has 5 categories: public facilities, road traffic facilities, urban environment implementation, landscaping facilities, and extended components; there are 10 categories of first-level events: urban environment, publicity and advertising, construction management, street order, street security, market supervision, emergencies, community management, rural management, and extended events; and the total number of second-level component / event categories is 235-245. A second-level component / event management profile database is established for each second-level component / event. The content of the second-level component / event management profile database includes: the category of the first-level component / event, the responsible department, the problem description, and the first-level handling requirements.

[0024] In the above-mentioned method for handling social events in the field of comprehensive social order management, in step S3, a primary component / event classification model is established through the Open-set Text Classification algorithm. The primary component / event classification model has 12+1 categories, where 12 are meaningful and fixed classification categories, including the above-mentioned 10 primary event categories and 5 primary component categories, and 1 is an open category, which includes new and potential component / event categories. The unified interface and data structure include a unified interface request method, data encryption method, departmental area coding, and callback method.

[0025] In the aforementioned method for handling social events in the field of comprehensive social governance, the Open-set TextClassification algorithm is built on CNN. During testing, the last layer of the CNN extracts the sentence features of the input class to obtain the Activation Vector (AV). Then, it calculates the k nearest distances from the Activation Vector of each test sample to each known class and denotes the Activation Vectors of these k samples as k-ClassActivation Vectors (k-CAV). The distance between the AV of the test sample and k-CAV is calculated. The sum of the k probability values ​​of each training class is denoted as the total closed set probability. Finally, OSP = 1 - total closed set probability is calculated. The maximum total closed set probability of all training classes is compared with OSP to determine whether the test class is an unknown class or a known class.

[0026] In the above-mentioned method for handling social events in the field of comprehensive social order, in step S4, the responsible department of the profile that meets the maximum threshold condition is the recommended handling department for the event. If the threshold condition cannot be met, the event is manually assigned, and the staff can determine whether to create a new secondary component / event management profile for this type of event.

[0027] In the aforementioned method for handling social events applied to the field of public security, step S5, the establishment of the binary classification algorithm model specifically includes the following steps:

[0028] S51. A classification training corpus S is formed from content other than the textual information in the event text. The prediction variables for corpus S are manually labeled, and y is denoted as y. i =+1 indicates a public welfare event, y i =-1 indicates a non-public event;

[0029] S52. Divide the sentence set S into Strain and Stest in an 8:2 ratio. For each sentence in Strain and Stest, calculate the following formula:

[0030] S53. Combine all sentence vectors in the Strain to obtain matrix X, and calculate the first principal component u of X using TruncatedSVD.

[0031] S54. Perform the following calculation on each sentence vector Vs in Strain and Stest: V s =V s -u*u T *v s ;

[0032] S55, For and Where y i ∈{+1,-1}, and i∈(1,2,…,m)∪(1,2,…,n), the model is trained and evaluated using the Support Vector Machine (SVM) classification algorithm and the Bayesian parameter optimizer.

[0033] In the above-mentioned social event handling method applied to the field of public security, in step S6, the results of the satisfaction survey are divided into: very satisfied, satisfied, basically satisfied, dissatisfied and very dissatisfied. Events with satisfaction results of very satisfied and satisfied are included in the event handling case library. For future events, relevant collaborative departments and handling processes are quickly recommended to users based on the historical data of the event handling case library.

[0034] In the aforementioned method for handling social incidents applied to the field of public security, the incident handling center includes incident sources such as citizen service hotlines, petitions, and other related industry management information systems. It standardizes the data structure for incidents that can be accessed, allows the receipt of additional fields beyond the basic fields, and provides a unified method for callbacks during the incident handling process. It offers diverse methods for handling social incidents, facilitating timely feedback and enabling simultaneous processing of information from various sources.

[0035] In one of the aforementioned methods for handling social incidents in the field of public security, the virtual network designates the primary person in charge or the responsible unit for the incident as a grid leader. As the social incident progresses and the number of personnel in the handling department increases, the grid automatically grows. The grid leader is responsible for coordinating relevant resources and tracking the final resolution of the incident. For incidents involving people's livelihoods, relevant parties are added to the virtual grid, and the progress of the entire incident can be understood and monitored through PCs and mobile devices, with evaluations conducted after the incident is resolved.

[0036] Compared with existing technologies, the advantages of this invention are: reasonable design, high efficiency in handling social events, and enhanced coordination for social events requiring multi-departmental collaboration. It defines a two-level classification system for urban components and events, defines component and event management profiles for subcategories, and employs a recommendation method for receiving departments based on a combination of RE2 and manual assignment. An event reception center is established to receive events from different departmental systems and serves as a channel for grassroots systems to request resources from higher levels. The invention incorporates two concepts: a virtual grid and a command center. Event leaders and parties involved in public welfare events are designated as grid members. For events requiring cross-departmental collaboration, an online command center is initiated, inviting relevant departments to participate in multi-departmental collaborative processing. Based on follow-up visits and satisfaction surveys, the processing of positively received events is saved in a case model library. When future events occur, the library's historical data will quickly recommend relevant collaborating departments and processing procedures to users, effectively improving the efficiency of handling social events. Attached Figure Description

[0037] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0038] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0039] like Figure 1 As shown, a method for handling social incidents in the field of public security is proposed. This method specifically includes the following steps:

[0040] S1. The event content is reported using the RE2 model and matched with the event content text in the case library. Events that meet the threshold matching conditions are entered into S3 for dispatch processing, referring to the events matched in the case library.

[0041] S2. Events that do not meet the threshold matching conditions are identified according to the first-level component / event open-set text classification model.

[0042] S3. Events from S1 or events identified as first-level components / events in S2 are transmitted to the event handling center to create a unified interface and data structure for accessing events.

[0043] S4. Based on the primary component / event category of the current event, retrieve all secondary component / event management images under the current primary component / event category from the secondary component / event management image library, and use the RE2 model to detect the similarity between the current event content and the problem description text in the secondary component / event management.

[0044] S5. Create a virtual network for the event, and add the responsible department to the grid as a grid member according to the secondary component / event management profile. Then, determine whether the current event is a public welfare event based on the binary classification algorithm model. If it is a public welfare event, extract the information of the parties involved from the current event using the named entity extraction method and add it to the virtual grid.

[0045] S6. Follow up on the handling results of the incident and conduct a satisfaction survey.

[0046] In step S1, the establishment of the RE2 model specifically includes the following steps:

[0047] S11. The input includes the current event event1 and the case library. For each event content event2 in the case library, perform the following steps:

[0048] S12. Convert the text events1 and event2 into vectors in embedding form;

[0049] S13. The embedding vector sequence is fed into three blocks for continuous processing. Each block has the same structure, consisting of an encoder layer, an alignment layer, and a fusion layer. The encoder layer calculates the context features of the sequence, the alignment layer calculates the relevance features of the text event1 and event2, and the output of the fusion layer is the input of the next block or the input of the pooling layer.

[0050] S14. The pooling layer converts the output of the block into a fixed-length vector.

[0051] After calculating the features of S15, event1, and event2, a binary classification model is constructed in the prediction layer. Category 0 indicates that event1 and event2 are not similar, and category 1 indicates that event1 and event2 are similar.

[0052] In step S2, the open classification model for primary components / events includes five primary component categories: public facilities, road traffic facilities, urban environment implementation, landscaping facilities, and extended components; and ten primary event categories: urban environment, advertising, construction management, street order, street security, market supervision, emergencies, community management, rural management, and extended events. The total number of secondary component / event categories is 235-245. A secondary component / event management profile database is established for each secondary component / event. This database includes the primary component / event category, responsible department, problem description, and primary handling requirements. Primary components include urban infrastructure within public areas of urban management. Primary components have two levels: a major category and a minor category, with each major category consisting of several minor categories. Primary events are related to human behavior or phenomena and behaviors that affect or damage the urban environment and order due to natural factors. Primary event categories include two major categories and two levels of minor categories, with each major category consisting of several minor categories.

[0053] In step S3, a primary component / event classification model is established using the Open-set Text Classification algorithm. The primary component / event classification model has 12+1 categories, where 12 are meaningful and fixed categories, including the aforementioned 10 primary event categories and 5 primary component categories. The remaining category is an open category, encompassing new and potential component / event categories. A unified interface and data structure are implemented, including unified interface request methods, data encryption methods, departmental area codes, and callback methods. As the city's development scale expands and new things emerge, some management content needs to be expanded and improved. The principle of expansion is: following the method of immutable major categories and extensible minor categories, new components and events are added under expanded component and expanded event categories.

[0054] Furthermore, the Open-set Text Classification algorithm is built on CNN. During testing, the last layer of the CNN extracts the sentence features of the input class to obtain the Activation Vector (AV). Then, it calculates the k nearest distances from the Activation Vector of each test example to each known class and denotes the Activation Vectors of these k examples as k-Class Activation Vectors (k-CAV). The distance between the AV of the test example and the k-CAV is calculated. The sum of the k probability values ​​of each training class is denoted as the total closed set probability. Finally, OSP = 1 - total closed set probability is calculated. The maximum total closed set probability of all training classes is compared with OSP to determine whether the test class is an unknown class or a known class.

[0055] In step S4, the responsible department for the profile that meets the maximum threshold condition is the recommended handling department for the event. If the threshold condition cannot be met, it is manually assigned, and staff can determine whether to create a new secondary component / event management profile for this type of event.

[0056] In step S5, the establishment of the binary classification algorithm model specifically includes the following steps:

[0057] S51. A classification training corpus S is formed from content other than the textual information in the event text. The prediction variables for corpus S are manually labeled, and y is denoted as y. i =+1 indicates a public welfare event, y i =-1 indicates a non-public event;

[0058] S52. Divide the sentence set S into Strain and Stest in an 8:2 ratio. For each sentence in Strain and Stest, calculate the following formula:

[0059] S53. Combine all sentence vectors in the Strain to obtain matrix X, and calculate the first principal component u of X using TruncatedSVD.

[0060] S54. Perform the following calculation on each sentence vector Vs in Strain and Stest: s =v s -u*u T *v s ;

[0061] S55, For and Where yi Given i ∈ {+1, -1} and i ∈ (1, 2, ..., m) ∪ (1, 2, ..., n), the model is trained and evaluated using a Support Vector Machine (SVM) classification algorithm and a Bayesian parameter optimizer. Furthermore, this invention provides an extension method that allows authorized third-party systems to access the event handling method, facilitating customized enrichment of the department's processing workflow for different levels of processing departments and different event types. To achieve this objective, this invention designs a set of extension interfaces and access authentication schemes:

[0062] 1. The system provides a unified framework specification based on three aspects: unified interface style, unified interface call, and unified access control. Third parties can develop and integrate their own business systems seamlessly by following the specification.

[0063] 2. The system adopts a front-end and back-end separation approach. The front-end provides a JavaScript plugin. After the plugin is introduced on a third-party page, the page style will be automatically standardized, tracking points will be inserted, and the back-end interface will be connected. After the front-end page is developed and embedded into this system, a dynamic login token can be obtained during the opening process.

[0064] 3. The system provides backend SDKs for various commonly used development languages, encapsulating security encryption logic and open interfaces. Third parties only need to configure the host, key, and secret key to easily access and call the interfaces provided by this system. The backend of the third party can obtain the currently logged-in user, permissions, and business information in real time based on the dynamic token obtained by the frontend, thereby continuing the third party's own internal business processes.

[0065] 4. Provide encryption logic instructions for development languages ​​that do not provide SDK support to integrate the technology themselves;

[0066] Based on the above solutions, this invention achieves both high flexibility in expansion and relatively strict standardization.

[0067] In detail, in step S6, the satisfaction survey results are categorized as: very satisfied, satisfied, basically satisfied, dissatisfied, and very dissatisfied. Events with very satisfied or satisfied results are added to the event handling case library. For future events, the system quickly recommends relevant collaborating departments and processing procedures to the user based on historical data from the case library. This assists users in quickly creating an effective command center to resolve issues as soon as possible. Furthermore, for cross-departmental or cross-administrative region events, the command center invites responsible units from the same or different administrative regions to establish a multi-departmental collaborative command center, reducing buck-passing and shirking responsibility, promoting the smooth resolution of social events, and improving the efficiency of social event handling. For each social event completed by the command center, the system automatically categorizes it and saves it to the case library along with the processing procedures of the collaborating departments within the command center. When future events occur, the system uses a recommendation algorithm to quickly recommend relevant collaborating departments and processing procedures to the user based on historical data from the case library, assisting the user in quickly creating an effective command center.

[0068] Furthermore, the incident handling center includes incident sources from citizen service hotlines, petitions, and other related industry management information systems. It standardizes the data structure of incidents that can be accessed, allows the acceptance of additional fields beyond the basic fields, and provides a unified way to call back the incident handling process.

[0069] Preferably, the virtual network designates the primary person in charge or responsible unit for an event as the grid leader. As the event progresses and the number of personnel in the handling department increases, the grid automatically grows. The grid leader is responsible for coordinating relevant resources and tracking the final resolution of the event. For social events involving people's livelihoods, the relevant parties involved in the social event are added to the virtual grid. They can understand and monitor the progress of the entire event through PC and mobile terminals, and evaluate the event after it is resolved.

[0070] In summary, the principle of this embodiment is as follows: by building an event acceptance center with a unified interface and data structure for accessing events, social events are accessed, and the departments responsible for handling these social events are intelligently recommended to facilitate quick allocation by staff. Furthermore, the actual allocation results are used as training material to reverse train the recommendation algorithm, thereby improving the accuracy of the recommendations.

[0071] Secondly, a virtual network is established during the handling of social events, and an online command center is initiated as needed to request multi-departmental collaboration. Relevant personnel can participate in the entire collaborative handling process of the event through PC or mobile terminals. The final handling results will be fed back to all relevant personnel in the event handling process in real time. In order to meet the requirements of real-time and efficient communication in the command center and to ensure smooth video and voice collaborative communication, a dedicated hardware server for the command center is built.

[0072] Finally, after the social event is handled, a satisfaction survey is conducted. Events with very satisfactory or satisfactory results are added to the event handling case library. The recommendation algorithm will be automatically trained on a regular basis based on new cases to generate new models and improve the accuracy of event handling recommendations. In the future, social events will be quickly recommended to users with relevant collaborating departments and handling processes based on historical data from the event handling case library.

[0073] The specific embodiments described herein are merely illustrative of the spirit of the invention. Those skilled in the art to which this invention pertains may make various modifications or additions to the described specific embodiments or use similar methods to substitute them, without departing from the spirit of the invention or exceeding the scope defined by the appended claims.

Claims

1. A method for handling social incidents applied in the field of public security, characterized in that, This method specifically includes the following steps: S1. The event content is reported using the RE2 model and matched with the event content text in the case library. Events that meet the threshold matching conditions are entered into S3 for dispatch processing, referring to the events matched in the case library. S2. Events that do not meet the threshold matching conditions are identified according to the first-level component / event open classification model, namely the Open-setText Classification model. S3. Events from S1 or those identified as first-level components / events in S2 are transmitted to the event handling center to create a unified interface and data structure for accessing events. S4. Based on the primary component / event category of the current event, retrieve all secondary component / event management images under the current primary component / event category from the secondary component / event management image library, and use the RE2 model to detect the similarity between the current event content and the problem description text in the secondary component / event management. S5. Create a virtual network for the event, and add the responsible department to the grid as a grid member according to the secondary component / event management profile. Then, determine whether the current event is a public welfare event based on the binary classification algorithm model. If it is a public welfare event, extract the information of the parties involved from the current event using the named entity extraction method and add it to the virtual grid. S6. Follow up on the handling results of the incident and conduct a satisfaction survey; In step S1, the establishment of the RE2 model specifically includes the following steps: S11. Input includes the current event event1 and the case library. For each event content event2 in the case library, perform the following steps: S12. Convert the text events1 and event2 into vectors in embedding form; S13. The embedding vector sequence is fed into three blocks for continuous processing. Each block has the same structure, consisting of an encoder layer, an alignment layer, and a fusion layer. The encoder layer calculates the context features of the sequence, the alignment layer calculates the relevance features of the text event1 and event2, and the output of the fusion layer is the input of the next block or the input of the pooling layer. S14. The pooling layer converts the output of the block into a fixed-length vector. After calculating the features of S15, event1, and event2, a binary classification model is constructed in the prediction layer. Category 0 indicates that event1 and event2 are not similar, and category 1 indicates that event1 and event2 are similar.

2. The social incident handling method applied to the field of comprehensive social management according to claim 1, characterized in that, In step S2, the first-level component / event open classification model has 5 first-level component categories: public facilities, road traffic facilities, urban environment implementation, landscaping facilities, and extended components; and 10 first-level event categories: urban environment, advertising, construction management, street order, street security, market supervision, emergencies, community management, rural management, and extended events. The total number of second-level component / event categories is 235-245. Establish a management profile library for each secondary component / event. The content of the secondary component / event management profile library includes: the category of the primary component / event to which it belongs, the responsible department, the problem description, and the primary handling requirements.

3. The social incident handling method applied to the field of comprehensive social order management according to claim 2, characterized in that, In step S3, a primary component / event classification model is established using the Open-set Text Classification algorithm. The primary component / event classification model has 12+1 categories, where 12 are meaningful and fixed categories, including the aforementioned 10 primary event categories and 5 primary component categories, and 1 is an open category, which includes new and potential component / event categories. The unified interface and data structure include a unified interface request method, data encryption method, departmental area encoding, and callback method.

4. The social incident handling method applied to the field of comprehensive social order management according to claim 3, characterized in that, The Open-set Text Classification algorithm described above is built on CNN. During testing, the last layer of the CNN extracts the sentence features of the input class to obtain the Activation Vector (AV). Then, it calculates the k nearest distances from the Activation Vector of each test example to each known class and denotes the Activation Vectors of these k examples as k-Class Activation Vectors (k-CAV). The distance between the AV of the test example and the k-CAV is calculated. The sum of the k probability values ​​of each training class is denoted as the total closed set probability. Finally, OSP = 1 - total closed set probability is calculated. The maximum total closed set probability of all training classes is compared with OSP to determine whether the test class is an unknown class or a known class.

5. The social incident handling method applied to the field of comprehensive social order management according to claim 4, characterized in that, In step S4, the responsible department for the profile that meets the maximum threshold condition is the recommended handling department for the event. If the threshold condition cannot be met, it is manually assigned, and staff can determine whether to create a new secondary component / event management profile for this type of event.

6. The social incident handling method applied to the field of comprehensive social management according to claim 1, characterized in that, In step S6, the results of the satisfaction survey are divided into: very satisfied, satisfied, basically satisfied, dissatisfied, and very dissatisfied. Events with very satisfied or satisfied results are included in the case library. For future events, relevant collaborating departments and processing procedures are quickly recommended to users based on historical data in the case library.

7. The social incident handling method applied to the field of comprehensive social management according to claim 6, characterized in that, The incident handling center includes incident sources from citizen service hotlines, petitions, and other related industry management information systems. It standardizes the data structure of incidents that can be accessed, allows the acceptance of additional fields beyond the basic fields, and provides a unified way to call back the incident handling process.

8. The social incident handling method applied to the field of comprehensive social management according to claim 7, characterized in that, The virtual network sets the first person in charge or the responsible unit of an event as the grid leader. As the event progresses and the number of personnel in the handling department increases, the grid automatically grows. The grid leader is responsible for coordinating relevant resources and tracking the final resolution of the event. For incidents involving people's livelihood, relevant parties are added to a virtual grid, and the progress of the entire incident is monitored and supervised through PC and mobile devices, and an evaluation is conducted after the incident is resolved.

Images