Government affair service optimization method based on user behaviors

Abstract

The invention provides a user behavior-based government affair service optimization method, which comprises the following steps of: acquiring user behavior data, and monitoring and identifying the user behavior data to obtain a target equipment group list; service demand analysis and matching are carried out according to the target device group list, and service demand information is generated; generating a personalized service information text according to the service demand information; pushing the personalized service information text to the user equipment in the target equipment group list; and obtaining the push feedback data and the service result data, performing correlation analysis, and optimizing a service strategy according to an analysis result.

Description

Technical Field

[0001] This invention relates to the field of service optimization technology, and more specifically, to a method for optimizing government services based on user behavior. Background Technology

[0002] With the rapid development of information technology, government services are gradually transforming towards intelligence and precision. Traditional government services often adopt a broadcast-style, passive response model, such as releasing public information through television, radio, and portal websites. This results in problems such as inaccurate information coverage, poor response timeliness, and low user reach. Especially in scenarios such as emergency management and disaster early warning, traditional methods are unable to achieve real-time targeted services to specific areas and groups. For example, in meteorological disaster early warning, existing technologies usually issue warnings by administrative division, failing to accurately identify and reach individual users actually located in dangerous areas. This results in warning information not effectively serving those who truly need it, reducing the efficiency and effectiveness of public services. Existing services lack the ability to dynamically perceive and predict user behavior, making it impossible to intervene before users enter risk areas or to quantitatively evaluate and optimize service effectiveness, thus limiting further improvements in the intelligence level of government services. Therefore, there is an urgent need for a method to optimize government services based on real-time user behavior data, enabling precision, proactivity, and measurability. Summary of the Invention

[0003] In view of this, the present invention proposes a government service optimization method based on user behavior to solve the problems existing in the prior art.

[0004] To achieve the above objectives, this invention proposes a method for optimizing government services based on user behavior, comprising: Acquire user behavior data, monitor and identify the user behavior data, and obtain a list of target device groups; Based on the target device group list, service demand analysis and matching are performed to generate service demand information; Based on the service request information, generate personalized service information text; The personalized service information text is pushed to user devices in the target device group list; Obtain push notification feedback data and service result data, perform correlation analysis, and optimize service strategies based on the analysis results.

[0005] Optionally, the process of monitoring and identifying user behavior data includes: The user's location is obtained based on user behavior data, which includes interaction signals between the user device and the communication network. The user's location is compared with a pre-defined electronic geofence. When a user device is detected entering or staying within the electronic geofence, it is determined that it meets specific behavioral conditions, and the anonymous identifier of the user device is collected into the target device group list. The electronic geofence is bound to government business rules and is activated when a bound business event occurs.

[0006] Optionally, monitoring and identifying user behavior data may also include a prediction step: The system monitors potential user devices outside the electronic geofence and predicts the probability of a potential user device entering the electronic geofence based on its movement trajectory, speed, and direction vector information. When the predicted probability exceeds a preset threshold, a predicted entry event is generated, and the user device is added to the target device group list.

[0007] Optionally, the process of predicting the probability of a potential user device entering the electronic geofence includes: The confidence level is calculated based on the movement trajectory, speed, and direction vector information, where the confidence level is the probability value of a potential user device entering the electronic geofence. The comprehensive calculation process includes calculating a movement vector factor, a movement speed factor, a trajectory orientation factor, and a spatiotemporal proximity factor based on the movement trajectory, speed, and direction vector information. The movement vector factor represents the rate at which the user device approaches or moves away from the fence. The movement speed factor is the ratio of the user device's current movement speed to a preset maximum speed. The trajectory orientation factor represents the angle between the user device's movement direction and the fence direction. The spatiotemporal proximity factor is calculated based on the Euclidean distance between the user device's current location and the fence boundary and the estimated arrival time.

[0008] Optionally, the electronic geofence is set up through the following steps: delineating a geographic area using a visual map tool or by inputting an administrative division code, and converting it into precise coordinate boundaries using a geocoding service to generate an electronic geofence object with a unique ID; binding the electronic geofence with government business rules to form an event-fence-action logic rule; and automatically activating the electronic geofence when the bound trigger event occurs.

[0009] Optional, the service requirements analysis and matching process includes: The system matches user devices in the target device group list with pre-defined rules in the government knowledge base; it triggers corresponding service requests based on the matching results and calls the corresponding standardized service content templates from the service content template library.

[0010] Optionally, the process of generating personalized service information text includes: The standardized service content template invoked will be automatically integrated with the current context information, which includes the specific area name, disaster type, and real-time time.

[0011] Optional, the correlation analysis process includes: Collect push feedback data, which includes message delivery rate, exposure rate, and user interaction data. Collect service outcome data, which includes data on casualties, property damage, changes in traffic indicators, and social media sentiment. The push feedback data and service result data are correlated in a unified spatiotemporal dimension, and the service effect is quantitatively evaluated through horizontal and vertical comparisons. Analyze the correlation between different service strategies and performance indicators, generate optimization instructions, and provide feedback to optimize service strategies.

[0012] On the other hand, the present invention also provides a government service optimization system based on user behavior for performing the above-described method.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: By dynamically matching electronic geofences with users' real-time locations and combining this with anonymous device identification technology, precise service delivery to specific areas, times, and groups of people is achieved. This effectively avoids information overload and public nuisance, significantly improving the relevance and user acceptance of government services. The introduction of a user behavior prediction mechanism can identify users' behavioral intentions before they actually enter risk areas, triggering early warnings and service preparations in advance. This represents a shift from a "passive response" to a "proactive intervention" service model, particularly suitable for time-sensitive scenarios such as emergency management and public safety. By collecting and analyzing push feedback data and service outcome data across multiple dimensions, service effectiveness can be quantitatively evaluated, effective strategies identified, and automatically fed back into service rules, content generation, and push strategies, forming a continuously self-optimizing intelligent service loop. The entire process uses anonymous device identifiers for user behavior tracking and service triggering, without involving personal identity information. This ensures user privacy and security while enabling personalized service content generation and delivery based on behavioral characteristics. The method of this invention is not limited to meteorological disaster early warning, but can be widely applied to various government service scenarios such as policy promotion, traffic guidance, and health tips. It has good versatility and replicability, and helps to promote the overall intelligent upgrade of the government service system. Attached Figure Description

[0014] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. In the drawings: Figure 1 This is a schematic diagram of the method flow in an embodiment of the present invention. Detailed Implementation

[0015] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the disclosure to those skilled in the art. It should be noted that, unless otherwise specified, the embodiments and features described herein can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0016] This embodiment proposes a government service optimization method based on user behavior, such as... Figure 1 As shown, it includes: S1. Obtain relevant user behavior data, and monitor and identify the user behavior data to obtain a target list; By monitoring the interaction signals between the user's mobile device and the signaling base station in the cellular network, and calculating the user's location in real time based on the interaction signals, the system can sense and record user device behavior events such as staying, entering or leaving a pre-set electronic geofence based on the user's location. The electronic geofence is the area range where a warning is issued or a meteorological disaster occurs, such as "a red rainstorm warning area has appeared in XX Street, XX City".

[0017] In behavioral events, all anonymous device identifiers that meet specific behavioral conditions are automatically aggregated to form a dynamic, pending target list, i.e., a target device group list. The behavioral conditions include staying in or entering an area under a warning or experiencing a meteorological disaster, such as entering a red rainstorm warning area in XX street, XX city during the warning period. The anonymous device identifier is the user device's identification information, such as an encrypted IMSI number.

[0018] The setup process for electronic geofences is as follows: In the government service management backend, authorized users (such as emergency management personnel) use visual map tools or directly input administrative division codes (such as the administrative codes of streets and communities) to delineate one or more polygonal or circular areas. Simultaneously, the system automatically converts the administrative region name (such as "XX City XX Street") into a precise set of latitude and longitude coordinates through geocoding services, forming a geographic boundary to generate a geofence object with a unique ID. Its core attributes include: fence ID, fence name (such as "XX Street Rainstorm Red Warning Zone"), and the set of geographic coordinate boundaries.

[0019] The geofence created in the previous step is then bound to specific government business rules. This binding includes a trigger event and a trigger action. The trigger event is: a weather warning platform issues a warning for a certain area, or a red rainstorm warning is issued. After the event is triggered, the associated geofence is automatically marked as a red rainstorm warning zone. This generates an "event-fence-action" logical rule. When the bound event occurs, the system automatically sets the geofence's status to active.

[0020] Once a bound business event (such as the issuance of a red alert for heavy rain) is detected by the system through the API interface, the rules engine will immediately execute the activation action. The activated fence ID and its coordinate boundaries are pushed to the location data processing platform in real time. The platform will use this fence boundary as the basis for judgment and begin to "determine the entry and exit" of incoming real-time user location signaling data.

[0021] The electronic fence is activated upon the occurrence of a triggering event. When the weather warning is lifted, the system automatically disables the fence. The location data processing platform will then cease monitoring the fence.

[0022] If the disaster area changes, authorized users can manually adjust the fence boundaries, and the updated coordinates will take effect immediately.

[0023] Under specific behavioral conditions, the user's location is recorded in real time. When the user's location is within the electronic geofence of the warning or meteorological disaster for a fixed period of time, it is marked as "staying". When the user's location was outside the electronic geofence at the previous time point of the warning or meteorological disaster and is inside the electronic geofence at the current time point, it is marked as "entering".

[0024] In order to provide better service to relevant users, we will predict the behavioral events that users may enter the pre-set electronic geofence based on the perception information of relevant users, and check whether the predicted behavioral events meet specific behavioral conditions. If they do, we will also add the user's device to the target device group list. Devices within a fixed distance extending outward from the electronic geofence are detected. The system continuously receives and records the sequence of anonymous user locations between the geofence and its extended boundary, forming real-time movement trajectories. Each trajectory point contains vector information such as location, time, speed, and direction. This constitutes a "potential target" list, containing all devices moving towards the active geofence that have not yet entered it. The prediction model is activated when a device's movement trajectory indicates a continuous approach to a geofence. Prediction calculations are based on the following key parameters: Distance: The Euclidean distance between the device's current location and the fence boundary. Movement vector: The device's current speed and direction. The directionality of movement is determined by calculating the angle between the movement direction and the directional angle pointing towards the fence center or entrance. Under disaster warning and occurrence conditions, the prediction model is triggered when the device's movement speed allows it to reach the geofence boundary within the Euclidean distance, and the angle in the movement vector is less than a certain threshold.

[0025] The prediction model combines the above parameters to calculate the probability (confidence level) that the device will enter the target fence in a short window in the future (e.g., within 5-10 minutes). The confidence score is calculated using the aforementioned movement speed, angle, and distance. Specifically, the distance between the current user's location and the electronic fence is calculated based on the movement speed and angle in the movement vector, taking into account the current time and the next time step. The difference between the distance at the previous time step and the current time step is then evaluated as positive or negative. A positive value assigns a movement vector factor score of 0.3, while a negative value assigns 1. A maximum movement speed is set, and the ratio of the current movement speed to the maximum speed is used as the movement speed factor score. A value of 1 is assigned if the maximum speed is exceeded. A maximum angle is also set, and the ratio of the difference between the maximum angles to the maximum angles is used as the angle factor score. A value of 0 is assigned if the maximum angle is exceeded. Finally, the ratio of the movement speed to the Euclidean distance at a specific point in time within the disaster warning or occurrence period is calculated as the movement distance factor score. The movement vector factor score, movement speed factor score, angle factor score, and movement distance factor score are then weighted and summed (e.g., weights of 0.25, 0.25, 0.25, and 0.25), and the weighted sum is used as the confidence score. When this probability value exceeds a preset confidence threshold (e.g., 85%), a predicted entry event for the current user device is generated, and the device is recorded in the target list for prior notification.

[0026] S2. Perform service analysis and matching based on the target list to generate service requirement information; In the target device group list, the occurrence of the behavioral condition is automatically matched with pre-defined rules in the government knowledge base. The rule engine determines that this behavioral condition triggers the service request of "meteorological disaster emergency warning". Then, it calls the standardized service content template ("Emergency Guide for Red Alert for Heavy Rain") bound to this request from the service content template library and prepares to execute the associated "emergency SMS push" service action. During this process, the aforementioned service request that has triggered the meteorological disaster emergency warning begins to prepare the corresponding service request information after being triggered. This service request information includes the standardized service content template bound to the request and the corresponding warning and meteorological disaster information. At this point, a definite logical association has been established between the user's behavior and the specific service content and service action.

[0027] In the rules engine, when severe weather occurs, its template is pre-associated with the aforementioned severe weather types. Based on the severe weather, the corresponding template is automatically called and the severe weather content is recorded.

[0028] S3. Based on the above service requirement information, generate the corresponding personalized service information text; The standardized content template invoked in the previous step is automatically integrated with the current context information to generate the final personalized service information text. The current context information includes the warning and the occurrence of meteorological disasters, such as the specific area name, disaster type, and real-time time. The generated personalized service information text is as follows: For example, replacing "[Area]" in the template with "XX Street" generates a complete, accurate, and highly targeted warning SMS message: "[XX City Emergency Management Bureau] reminds you: XX Street has issued a red rainstorm warning. Please immediately prepare for disaster prevention and mitigation." This step ensures that the service information not only conforms to standards but also dynamically adapts to specific scenarios, achieving a technological transformation from "standardized service" to "personalized service."

[0029] S4. Push personalized service information text to the devices in the target list with precise service; After generating personalized service content, the message to be sent, along with the target device group list in S1, is submitted to the operator's SMS delivery gateway via the application programming interface (API). The delivery gateway strictly follows the received list, only sending push requests to device numbers within the list. This ensures that the information is accurately delivered to the target behavioral group (people within the warning area) without interfering with users outside the area, achieving "precision targeting" of the service. The target user's mobile device receives the service SMS via the mobile communication network and is notified to view it via a terminal notification. At this point, the government service has completed its final reach from the government to the user. The risk-avoidance behavior taken by the user after reading the information signifies the completion of the technical process of this government service, and the achievement of the service objective.

[0030] S5. End after push notification. Collect feedback data and service result data from this push notification and perform correlation analysis. Feedback data includes SMS delivery rate and user feedback tags, while service result data includes disaster loss reports for the area.

[0031] The specific details of the above correlation analysis are as follows: The system collects feedback data from multiple sources. User push feedback data includes carrier gateway receipts, terminal interaction data, message exposure rate, subsequent operation conversion rate, and ignore / close rate. The carrier gateway receipts are integrated with the SMS gateway, automatically collecting the delivery status of each push message (success, failure, device off, etc.) and calculating the delivery rate. Terminal interaction data (e.g., via government apps) is used for messages pushed through apps, collecting: message exposure rate (representing the proportion of users who click to view the message); subsequent operation conversion rate (representing the proportion of users who click links or buttons in the message or perform subsequent queries / operations); and ignore / close rate (representing the proportion of users who directly close the notification).

[0032] Simultaneously, service outcome data (a macro-level measure of the final service effectiveness) is collected. This data includes: analyzing the number of casualties and property damage caused by the disaster within the warning area, and comparing this with data from similar historical disasters where services were not provided; collecting data on changes in average vehicle speed, congestion index, and the number of traffic accident reports within the warning area; collecting data on changes in the number of community hospital visits for specific diseases after specific health alerts are pushed out; and utilizing publicly available social media data, employing sentiment analysis techniques to assess changes in public sentiment (positive, negative, neutral) towards related topics after receiving service information. Finally, heat map data of pedestrian flow in specific areas is obtained to analyze whether the population density in those areas decreased as expected after the evacuation notice was issued.

[0033] Data association and fusion are performed between user push notification data and statistical service results data: All collected user push notification data and statistical service results data are unified to the same time window and geographical scope. For example, analyzing the various results data for "XX Street" within 6 hours after a rainstorm warning SMS is sent. A structured analysis record is created for each service push activity. This record associates the push data with the results data: Its analysis records include a primary key: Service Activity ID, and its corresponding feature data includes push channels, push content, target audience size, message delivery rate, click-through rate, etc. Its corresponding final result data includes the number of casualties in the region, estimated property damage, traffic congestion index change rate, and the proportion of negative sentiment on social media, etc.

[0034] After association, the association analysis model is calculated, specifically: The aforementioned related datasets were analyzed using statistical and machine learning models to conduct a quantitative evaluation of the effects.

[0035] The assessment includes both horizontal and vertical comparisons. The horizontal comparison analyzes the differences between the target area that received the push notification and adjacent control areas with similar disaster situations that did not receive the notification. The vertical comparison analyzes the differences between the data after this push notification and similar events in the region's history when the service was not pushed. This allows for quantitative assessment conclusions such as "This push notification may have reduced casualties by X%" and "Property damage decreased by Y%".

[0036] Following the evaluation, relevant strategy improvement analyses were conducted: using multivariate correlation analysis, the correlation between different push strategies (such as push timing, urgency of content wording, and push channels) and final performance indicators (such as casualty reduction rate and click-through rate) was analyzed. Insights were generated, such as "pushing a message 30-60 minutes before a disaster is more effective than pushing it 2 hours in advance" and "SMS messages containing specific action guidelines have twice the subsequent conversion rate of ordinary reminder SMS messages."

[0037] After analysis, specific optimization strategies are generated and fed back. The analysis findings were translated into specific, actionable optimization instructions, which were automatically fed back into the core processes. The S1 analysis showed that sending alerts to users who briefly passed through the area could effectively reduce incidents. It was recommended to monitor users who might exhibit such behavior, such as by monitoring navigation software or the triggering conditions for predicted entry events using the current user's device. For example, if a larger (pre-set) range of users was detected with speeds exceeding a threshold, these users could be marked as potentially exhibiting such behavior, and predictions could be made based on the fixed maximum values ​​and weights in the prediction model.

[0038] After feedback to S2: Add more templates and further integrate them with disaster weather information, such as including specific instructions like 'Immediately find a sturdy shelter,' which will result in a higher conversion rate. It is recommended to prioritize these types of templates in the emergency template library. After feedback to S4: User profiles can be optimized, and the APP channel can be prioritized for younger users.

[0039] It should also be noted that, for S1, the device predicted to enter the event records the anonymous device identifier, the predicted fence ID, the event type (predicted entry), the prediction confidence level, and the estimated arrival time. For predictions with extremely high confidence, the device can be added to the "Prediction Target List" before the user actually enters. Preparatory information is sent in advance based on the prediction target category, using the above method and specific service requirement information (a combination of specific templates and meteorological information). For example: "You are heading towards a red rainstorm warning area, expected to arrive in 10 minutes. Please find a safe place to take shelter immediately." If the device actually enters the fence within the ETA time window, the prediction is marked as successful, and this type of trajectory feature is reinforced, such as further adjusting the weight allocation of its confidence calculation. If the device does not enter (e.g., changes direction, stops), it is marked as a failure and removed from the "Target Device Group List" and the "Prediction Target List." The error data from this prediction will be sent back to the prediction model for training and optimizing future prediction accuracy.

[0040] Meanwhile, historical behavior patterns are analyzed: if the anonymous device has a traceable historical trajectory, the model will analyze its historical behavior, such as whether it regularly commutes, frequently visits the area or similar areas. If such patterns exist, the device is strongly labeled and associated with the area. After an area warning or disaster occurs, the strongly labeled device is added to the target list and the prediction target list. Simultaneously, the system prepares personalized service content for these devices in advance and establishes push channels, keeping them in a "standby" state. Once an actual "entry" event is received, the push command can be sent with zero delay.

[0041] On the other hand, the present invention also provides a government service optimization system based on user behavior, comprising: The first module is used to acquire relevant user behavior data, monitor and identify the user behavior data, and obtain a target list; The second module is used to perform service analysis and matching based on the target list and generate service requirement information. The third module is used to generate corresponding personalized service information text based on the above service requirement information; The fourth module is used to push personalized service information text to devices in the target list with precise service delivery. The fifth module is used to conclude the process after the push notification. It collects feedback data and service result data from this push notification and performs correlation analysis.

[0042] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for optimizing government services based on user behavior, characterized in that: include: Acquire user behavior data, monitor and identify the user behavior data, and obtain a list of target device groups; Based on the target device group list, service demand analysis and matching are performed to generate service demand information; Based on the service request information, generate personalized service information text; The personalized service information text is pushed to user devices in the target device group list; Obtain push notification feedback data and service result data, perform correlation analysis, and optimize service strategies based on the analysis results.

2. The method according to claim 1, characterized in that, The process of monitoring and identifying user behavior data includes: The user's location is obtained based on user behavior data, which includes interaction signals between the user device and the communication network. The user's location is compared with a pre-defined electronic geofence. When a user device is detected entering or staying within the electronic geofence, it is determined that it meets specific behavioral conditions, and the anonymous identifier of the user device is collected into the target device group list. The electronic geofence is bound to government business rules and is activated when a bound business event occurs.

3. The method according to claim 2, characterized in that, Monitoring and identifying user behavior data also includes a prediction step: The system monitors potential user devices outside the electronic geofence and predicts the probability of a potential user device entering the electronic geofence based on its movement trajectory, speed, and direction vector information. When the predicted probability exceeds a preset threshold, a predicted entry event is generated, and the user device is added to the target device group list.

4. The method according to claim 2, characterized in that, The process of predicting the probability of a potential user device entering the electronic geofence includes: The confidence level is calculated based on the movement trajectory, speed, and direction vector information, where the confidence level is the probability value of a potential user device entering the electronic geofence. The comprehensive calculation process includes calculating a movement vector factor, a movement speed factor, a trajectory orientation factor, and a spatiotemporal proximity factor based on the movement trajectory, speed, and direction vector information. The movement vector factor represents the rate at which the user device approaches or moves away from the fence. The movement speed factor is the ratio of the user device's current movement speed to a preset maximum speed. The trajectory orientation factor represents the angle between the user device's movement direction and the fence direction. The spatiotemporal proximity factor is calculated based on the Euclidean distance between the user device's current location and the fence boundary and the estimated arrival time.

5. The method according to claim 2, characterized in that, The electronic geofence is set up through the following steps: delineating a geographic area using a visual map tool or by inputting an administrative division code, and converting it into precise coordinate boundaries using a geocoding service to generate an electronic geofence with a unique ID; The electronic geofence is bound to government business rules to form an event-fence-action logic rule; when the bound trigger event occurs, the electronic geofence is automatically activated.

6. The method according to claim 1, characterized in that, The process of service demand analysis and matching includes: The system matches user devices in the target device group list with pre-defined rules in the government knowledge base; it triggers corresponding service requests based on the matching results and calls the corresponding standardized service content templates from the service content template library.

7. The method according to claim 1, characterized in that, The process of generating personalized service information text includes: The standardized service content template invoked will be automatically integrated with the current context information, which includes the specific area name, disaster type, and real-time time.

8. The method according to claim 1, characterized in that, The process of correlation analysis includes: Collect push feedback data, which includes message delivery rate, exposure rate, and user interaction data. Collect service outcome data, which includes data on casualties, property damage, changes in traffic indicators, and social media sentiment. The push feedback data and service result data are correlated in a unified spatiotemporal dimension, and the service effect is quantitatively evaluated through horizontal and vertical comparisons. Analyze the correlation between different service strategies and performance indicators, generate optimization instructions, and provide feedback to optimize service strategies.

9. A government service optimization system based on user behavior, characterized in that: Used to perform the method described in any one of claims 1-8.

Images