A service allocation system for government service halls based on big data analytics

Abstract

This invention discloses a service allocation system for government service halls based on big data analysis, relating to the field of government service technology. The system includes a service category classification module, a real-time service allocation module, and a special service allocation module. The service category classification module accurately categorizes user requests for government services according to their type, significantly improving the efficiency and accuracy of service allocation. The separate processing of special and non-special service requests avoids resource waste and allocation conflicts. The real-time service allocation module introduces a real-time request queue and a real-time allocation model to ensure that non-special service requests receive a rapid response when service resources are available, improving user satisfaction. The special service allocation module optimizes resource allocation specifically for the needs of special users, achieving more refined and personalized services through separate windows and allocation mechanisms.

Description

Technical Field

[0001] This invention relates to the field of government service technology, and in particular to a government service hall service allocation system based on big data analysis. Background Technology

[0002] Current government service systems face several problems in service allocation, including uneven resource distribution, excessively long user waiting times, lack of intelligent scheduling, and inaccurate identification of user needs. Traditional government service systems typically rely on manual scheduling or fixed rules for service allocation, which makes resource allocation inflexible. This leads to some service windows being overloaded, resulting in long user wait times, while other windows are not fully utilized. This unreasonable resource allocation not only reduces service efficiency but also seriously affects user experience, leading to overall low efficiency and insufficient user satisfaction in government services.

[0003] To address this issue, it is necessary to analyze service demand using real-time information and historical data, dynamically adjust service resources in the government service hall, and implement specific measures such as introducing an intelligent dispatch system, optimizing queuing and appointment services, enhancing the personalization and diversification of services, and continuously optimizing through feedback mechanisms and contingency plans to improve resource utilization efficiency and overall service quality. Summary of the Invention

[0004] The present invention aims to provide a government service hall allocation system based on big data analysis to improve resource utilization efficiency and overall service quality.

[0005] A service allocation system for government service halls based on big data analytics includes:

[0006] The service category classification module includes a data acquisition unit and a category classification unit. The data acquisition unit acquires the government service requests from users to be assigned. These requests include a user identifier, service request content, service request time, and special user tags. The category classification unit analyzes the requests based on the government service allocation model to obtain the service categories for the users to be assigned. These categories include special service categories and non-special service categories. Based on these categories, the module allocates the government service requests. If the service category is non-special, the request is transferred to the real-time service allocation window; otherwise, it is transferred to the special service allocation window.

[0007] The real-time service allocation module includes a queue acquisition unit and a service allocation unit. The queue acquisition unit is used to input user requests for government services to be allocated into the real-time government service request queue within the real-time service allocation window, awaiting resource allocation. The real-time government service request queue contains user requests for government services Q. m m = 1, 2, ..., M; M is the total number of user government service requests to be allocated in the real-time government service request queue; the service allocation unit is used to allocate resources based on the real-time government service request queue and the real-time government service allocation model, and obtain the government service resource allocation result H. m According to the allocation results of government service resources H m Perform service allocation;

[0008] The special service allocation module includes a special service allocation unit; the special service allocation unit is used to allocate special resources for the government service requests of users to be allocated in the special service allocation window.

[0009] As a preferred technical solution of the present invention, the government service allocation and division model includes an information extraction layer, a category division layer and a result output layer;

[0010] The information extraction layer is used to convert the user identifier, service request content, service request time, and special user tags in the government service request of the user to be assigned into a feature set X of the government service request of the user to be assigned, X={X n |n=1, 2, ..., N}; where X n The feature items for the government service requests of users to be assigned are N, where N is the total number of feature items in the feature set X of the government service requests of users to be assigned.

[0011] The category segmentation layer is used to perform category analysis based on the feature set X of the government service request of the user to be assigned, and to obtain the service category of the user to be assigned.

[0012] The output layer is used to output the service categories to be assigned to users.

[0013] As a preferred embodiment of the present invention, the specific steps for performing category analysis in the category classification layer include:

[0014] The category classification layer includes a multi-indicator evaluation layer and a comprehensive category classification layer;

[0015] In the multi-indicator evaluation layer, based on the feature item X of the government service request from the user to be assigned... n Category label assignment is performed to obtain the feature item label W for government service requests. n ;

[0016] In the comprehensive category classification layer, based on the feature item label W of all government service requests nPerform feature fusion to obtain the user service category to be assigned;

[0017] The specific steps for training a multi-metric evaluation layer include:

[0018] The multi-index evaluation layer uses the BP neural network model as the basic model; it collects several sets of government service request label pairs as training samples, each set of government service request label pairs as training samples contains target feature labels and corresponding government service request feature items; it combines several sets of government service request label pairs as training samples to obtain the government service request label pairs training set;

[0019] The BP neural network model is trained on the training set using government service request tags to obtain an initial multi-index evaluation layer. The initial multi-index evaluation layer is then evaluated. If the initial multi-index evaluation layer passes the model evaluation, it is used as the multi-index evaluation layer in the category division layer. Otherwise, the model is trained on the training set using government service request tags.

[0020] As a preferred embodiment of the present invention, the specific steps for feature fusion in the comprehensive category division layer include:

[0021] Based on the feature tag W of all government service requests n Construct a feature matrix of government service request tags;

[0022] Based on the feature matrix of government service request labels and the local attention mechanism, neighborhood features are captured to obtain the local government service request label features F. l ;

[0023] Based on the feature matrix of government service request labels and the global attention mechanism, feature correlation is calculated to obtain the global government service request label feature F. g ;

[0024] Using the formula F=[F l ;F g [The partial government service request tag feature F] l and global government service request tag feature F g The data is then combined to obtain integrated government service request tag features;

[0025] The integrated government service request label features are input into a linear layer for category mapping to obtain the user service category to be assigned.

[0026] The specific steps for training the comprehensive category-class splitting layer include:

[0027] Collect several sets of matrix category training samples. Each set of matrix category training samples contains a label feature matrix and the corresponding category classification result. Combine the several sets of matrix category training samples to obtain the matrix category training set.

[0028] The model is trained using the matrix category training set to obtain an initial comprehensive category partitioning layer. The initial comprehensive category partitioning layer is then evaluated. If the initial comprehensive category partitioning layer passes the model evaluation, it is used as the comprehensive category partitioning layer in the category partitioning layer; otherwise, the model is trained again using the matrix category training set.

[0029] As a preferred technical solution of the present invention, the real-time allocation model for government services includes a feature information acquisition layer, a dynamic resource allocation layer, and a resource allocation result output layer.

[0030] The feature information acquisition layer is used to obtain the government service request Q of the user to be assigned using big data technology. m Historical government service data is used to obtain derived government service data Y. m ;

[0031] The dynamic resource allocation layer is used to allocate government service requests from all users to be allocated (Q). m and derived government service data Y m Dynamic resource allocation is performed using a swarm optimization algorithm to obtain the government service resource allocation result H. m ;

[0032] The resource allocation result output layer is used to output the government service resource allocation results H m .

[0033] As a preferred embodiment of the present invention, the specific steps for performing dynamic resource allocation in the dynamic resource allocation layer include:

[0034] Obtain currently available government service resources to be allocated;

[0035] Based on all pending user government service requests Q m and derived government service data Y m Construct K entities for allocating government service resources, D k k=1,2,…,K; each individual D allocates government service resources. k It contains the allocation results of the currently pending government service resources based on the real-time request queue for government services; and allocates K government service resources to individual D. k Combining these elements, we obtain an iterative population for allocating government service resources; the number of iterations is set to t, t=1, 2, ..., t max ;t max This represents the maximum number of iterations.

[0036] Construct a simulated government service satisfaction model; the simulated government service satisfaction model includes a prediction and adjustment layer and a resource allocation evaluation layer; in the prediction and adjustment layer, based on derived government service data Y... mBy performing predictor matching, the simulated government predictor factor P is obtained. m In the resource allocation assessment layer, the simulated government policy prediction factor P is incorporated. m Requests for government services from users awaiting allocation (Q) m And the allocation of government service resources to individual D k A simulated satisfaction rating A was obtained by performing a simulated satisfaction rating calculation. m ;

[0037] Calculate all pending user government service requests Q m Corresponding simulated satisfaction rating A m The mean is used as the individual for allocating government service resources. k fitness S k ;

[0038] During the population iteration process, the non-dominated solutions generated in each iteration are stored to obtain the elite government service resource allocation population. A hypercube is then constructed for the elite government service resource allocation population to obtain the elite government service resource allocation hypercube C. t ;

[0039] Using formula Control the iterative direction of the population that allocates government service resources; among which, As the initial perturbation weights, For exponential control parameters, Hypercube C for allocating resources for elite government services t Random points within; Individual D for allocating government service resources k The position in the t-th iteration; Individual D for allocating government service resources k The displacement formula in the t-th iteration; These are random weighting coefficients, with values ​​in the range [0, 1].

[0040] When the maximum number of iterations is reached, output the individual with the highest fitness for government service resource allocation, which is the optimal individual for government service resource allocation. Based on the allocation results of the optimal individual for government service resource allocation according to the real-time request queue for government services, output the government service resource allocation result H. m .

[0041] As a preferred technical solution of the present invention, the specific steps for special resource allocation for user government service requests include:

[0042] Obtain special government service resources to be allocated;

[0043] Based on the government service requests of users to be assigned, service resources are matched to obtain matched special government service resources; and special resource allocation is completed for the special government service resources to be assigned based on the matched special government service resources.

[0044] As a preferred embodiment of the present invention, the swarm optimization algorithm is the particle swarm optimization algorithm.

[0045] The present invention has the following advantages:

[0046] 1. This invention uses a service category segmentation module to accurately classify user government service requests according to their categories, significantly improving the efficiency and accuracy of service allocation. The separate processing of special service requests and non-special service requests avoids resource waste and allocation conflicts. The real-time service allocation module introduces a real-time request queue and a real-time allocation model to ensure that non-special service requests can receive a rapid response when service resources are available, improving user satisfaction. The special service allocation module is specifically designed to optimize resource allocation for the needs of special users, achieving more refined and personalized services through a separate window and allocation mechanism.

[0047] 2. This invention assigns category labels to each feature item through a multi-index evaluation layer, achieving fine-grained analysis at the feature level. The comprehensive category division layer further integrates global information through feature fusion, ensuring the accuracy and consistency of category division. The local attention mechanism captures neighborhood features, while the global attention mechanism measures global relevance, effectively reducing noise interference and improving feature representation capabilities. The category division task is split into two layers: multi-index evaluation and comprehensive category division. Each layer has independent functional modules and training processes, facilitating expansion and optimization. The functions of each layer are clear and complementary. The multi-index evaluation layer implements fine-grained label assignment, while the comprehensive category division layer is responsible for global feature fusion and category mapping. The multi-index evaluation layer, based on a BP neural network model, continuously adapts to different service request features and improves model generalization ability through dynamic training and iterative optimization on a large-scale government service request training set. The comprehensive category division layer enhances the ability to capture complex category relationships through the construction of matrix category training samples and the introduction of an attention mechanism, ensuring the intelligence and accuracy of the category division results. Attached Figure Description

[0048] Figure 1 This is a schematic diagram of the structure of a government service hall allocation system based on big data analysis, as used in an embodiment of the present invention. Detailed Implementation

[0049] To enable those skilled in the art to better understand the technical solutions of this invention, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of this invention.

[0050] A service allocation system for government service halls based on big data analytics, see [link / reference]. Figure 1 As shown, it includes:

[0051] The service category classification module includes a data acquisition unit and a category classification unit.

[0052] The data acquisition unit is used to acquire the government service requests of users to be assigned; the government service requests of users to be assigned include user identifiers, service request content, service request time, and special user tags.

[0053] User identifiers are codes or information used to uniquely identify each user, such as user ID, account name, ID card number, or other unique information. Their purpose is to ensure that each government service request can be clearly linked to a specific user, avoiding confusion or duplicate processing. Service request content is a description of the specific government service needs raised by the user, which may include service type (e.g., applying for documents, policy consultation, fee payment), service details, or problem descriptions. This is the core basis for the system to analyze user needs and match resources. Service request time is the specific time when the user submits the government service request. This data is used to manage service response priorities, statistically analyze service peak periods, or track the timeliness of the service process. Special user tags are tags used to mark certain users' special attributes or needs, such as: Priority service tag: for users who require priority processing, such as the elderly or disabled; VIP user tag: high-priority users who enjoy specific resources or services; Sensitive user tag: users involved in privacy protection or special matters; Special matter tag: users with special processing needs, etc.

[0054] The category division unit is used to analyze the user's government service request and the government service allocation model to obtain the user's service category. The user's service category includes special service category and non-special service category. Based on the user's service category, the user's government service request is allocated. If the user's service category is non-special service category, the user's government service request is transferred to the real-time service allocation window for service request; otherwise, the user's government service request is transferred to the special service allocation window for service request.

[0055] The government service allocation and classification model includes an information extraction layer, a category classification layer, and a result output layer.

[0056] The information extraction layer is used to convert the user identifier, service request content, service request time, and special user tags in the government service request of the user to be assigned into a feature set X of the government service request of the user to be assigned, X={X n |n=1, 2, ..., N}; where X nLet N be the feature items of the user's government service request to be assigned, and let N be the total number of feature items in the feature set X of the user's government service request to be assigned.

[0057] The category segmentation layer is used to perform category analysis based on the feature set X of the government service requests of the users to be assigned, and to obtain the service categories of the users to be assigned.

[0058] The output layer is used to output the service categories to be assigned to users.

[0059] The information extraction layer performs feature processing on key information of user government service requests to be allocated, such as user identifiers, request content, time, and special tags, to ensure that the basic data for category classification is efficient and accurate. The category classification layer analyzes based on feature sets, which can accurately identify special service categories and non-special service categories, thereby providing a reliable basis for subsequent resource allocation.

[0060] Requests for non-special service categories are directly assigned to the real-time allocation window and processed efficiently using the real-time allocation model, which greatly improves service response speed. Requests for special service categories enter a separate special service allocation window to achieve personalized resource allocation, avoid resource competition and interference, and meet the service needs of special users.

[0061] The specific steps for performing category analysis at the category segmentation level include:

[0062] The category classification layer includes a multi-indicator evaluation layer and a comprehensive category classification layer.

[0063] In the multi-indicator evaluation layer, based on the feature item X of the government service request from the user to be assigned... n Category label assignment is performed to obtain the feature item label W for government service requests. n。

[0064] In the comprehensive category classification layer, based on the feature item label W of all government service requests n Feature fusion is performed to obtain the user service category to be assigned.

[0065] The specific steps for training a multi-metric evaluation layer include:

[0066] The multi-index evaluation layer uses the BP neural network model as the basic model; it collects several sets of government service request label pairs as training samples, each set of government service request label pairs as training samples contains target feature labels and corresponding government service request feature items; it combines several sets of government service request label pairs as training samples to obtain the government service request label pairs training set.

[0067] The BP neural network model is trained on the training set using government service request tags to obtain an initial multi-index evaluation layer. The initial multi-index evaluation layer is then evaluated. If the initial multi-index evaluation layer passes the model evaluation, it is used as the multi-index evaluation layer in the category division layer. Otherwise, the model is trained on the training set using government service request tags.

[0068] The specific steps for feature fusion in the comprehensive category segmentation layer include:

[0069] Based on the feature tag W of all government service requests n Construct a feature matrix of government service request tags.

[0070] Based on the feature matrix of government service request labels and the local attention mechanism, neighborhood features are captured to obtain the local government service request label features F. l。

[0071] Based on the feature matrix of government service request labels and the global attention mechanism, feature correlation is calculated to obtain the global government service request label feature F. g。

[0072] Using the formula F=[F l ;F g [The partial government service request tag feature F] l and global government service request tag feature F g The data is then combined to obtain the integrated government service request tag features.

[0073] The integrated government service request label features are input into a linear layer for category mapping to obtain the user service category to be assigned.

[0074] In the category mapping step, the fused government service request label feature vector is input into a fully connected linear layer for mapping. This linear layer uses a weight matrix and a bias to complete the mapping from features to categories. Each row is transformed into a probability distribution through a softmax activation function, and finally the category prediction result corresponding to each sample is obtained. This step is adjusted through training to minimize the cross-entropy loss between the category prediction result and the true category label, so as to achieve accurate mapping of service categories.

[0075] The specific steps for training the comprehensive category-class splitting layer include:

[0076] Collect several sets of matrix category training samples. Each set of matrix category training samples contains a label feature matrix and the corresponding category classification result. Combine the several sets of matrix category training samples to obtain the matrix category training set.

[0077] The model is trained using the matrix category training set to obtain an initial comprehensive category partitioning layer. The initial comprehensive category partitioning layer is then evaluated. If the initial comprehensive category partitioning layer passes the model evaluation, it is used as the comprehensive category partitioning layer in the category partitioning layer; otherwise, the model is trained again using the matrix category training set.

[0078] The multi-index evaluation layer assigns category labels to each feature item, achieving fine-grained analysis at the feature level. The comprehensive category division layer further integrates global information through feature fusion, ensuring the accuracy and consistency of category division. The local attention mechanism captures neighborhood features, while the global attention mechanism measures global relevance, effectively reducing noise interference and improving feature representation capabilities. The category division task is split into two layers: multi-index evaluation and comprehensive category division. Each layer has independent functional modules and training processes, facilitating expansion and optimization. The functions of each layer are clear and complementary. The multi-index evaluation layer implements fine-grained label assignment, while the comprehensive category division layer is responsible for global feature fusion and category mapping. The multi-index evaluation layer, based on a BP neural network model, continuously adapts to different service request characteristics and improves the model's generalization ability through dynamic training and iterative optimization on a large-scale government service request training set. The comprehensive category division layer enhances the ability to capture complex category relationships through the construction of matrix category training samples and the introduction of an attention mechanism, ensuring the intelligence and accuracy of the category division results.

[0079] The category mapping process after feature fusion introduces a linear layer to ensure the continuity and stability of category output, thereby improving the final effect of service category division.

[0080] The real-time service allocation module includes a queue acquisition unit and a service allocation unit.

[0081] The queue acquisition unit is used to input the government service requests of users to be allocated into the government service real-time request queue in the service real-time allocation window to wait for resource allocation; the government service real-time request queue contains the government service requests Q of users to be allocated. m m = 1, 2, ..., M; M is the total number of user government service requests to be allocated in the real-time government service request queue.

[0082] The specific value of M is set by professional technicians based on the actual situation, representing the maximum number of batch processing requests in a single batch.

[0083] The real-time service allocation module establishes a real-time request queue for government services through the queue acquisition unit, ensuring that all user requests awaiting allocation are queued in an orderly manner according to a first-in-first-out logic, avoiding conflicts and chaos in resource allocation; effectively reducing allocation latency, improving the utilization efficiency of service resources, and achieving efficient response to user needs; the dynamic management capability of the real-time request queue can be adjusted according to fluctuations in the number of requests, ensuring the stability and continuity of service allocation regardless of changes in queue length, adapting to the processing needs of large batches of requests during peak periods, and balancing the load on the service system.

[0084] The service allocation unit is used to allocate resources based on the real-time request queue and real-time allocation model of government services, and obtain the government service resource allocation result H. m According to the allocation results of government service resources H m Perform service allocation.

[0085] The real-time allocation model for government services includes a feature information acquisition layer, a dynamic resource allocation layer, and a resource allocation result output layer.

[0086] The feature information acquisition layer is used to obtain the government service request Q of the user to be assigned using big data technology. m Historical government service data is used to obtain derived government service data Y. m。

[0087] The dynamic resource allocation layer is used to allocate government service requests from all users to be allocated (Q). m and derived government service data Y m Dynamic resource allocation is performed using a swarm optimization algorithm to obtain the government service resource allocation result H. m The swarm optimization algorithm is the particle swarm optimization algorithm.

[0088] The resource allocation result output layer is used to output the government service resource allocation results H m。

[0089] The feature information acquisition layer utilizes big data technology to mine historical government service data from users, generating derived data that provides comprehensive data support for dynamic resource allocation, helping to improve the accuracy and targeting of resource allocation. The dynamic resource allocation layer, based on a group optimization algorithm, performs dynamic calculations according to current request characteristics and historical data, adapting to changes in resource supply and demand in complex scenarios. Through the analysis of historical government service data, the model can predict resource demand trends, thereby adjusting allocation strategies in advance and reducing the risk of resource shortages during peak periods. The generation of derived data increases the depth of understanding of user behavior patterns and demand patterns, providing more decision-making basis for dynamic allocation. The dynamic resource allocation layer can process diverse user requests in real time, adapt to changes in resource demand, and adjust allocation strategies based on real-time data, ensuring the efficiency and flexibility of the allocation process.

[0090] The specific steps for dynamic resource allocation in the dynamic resource allocation layer include:

[0091] Obtain currently available government service resources to be allocated;

[0092] Based on all pending user government service requests Q m and derived government service data Y m Construct K entities for allocating government service resources, D k k=1,2,…,K; each individual D allocates government service resources. k It contains the allocation results of the currently pending government service resources based on the real-time request queue for government services; and allocates K government service resources to individual D. k Combining these elements, we obtain an iterative population for allocating government service resources; the number of iterations is set to t, t=1, 2, ..., t max ;t max This represents the maximum number of iterations.

[0093] Construct a simulated government service satisfaction model; the simulated government service satisfaction model includes a prediction and adjustment layer and a resource allocation evaluation layer; in the prediction and adjustment layer, based on derived government service data Y... m By performing predictor matching, the simulated government predictor factor P is obtained. m In the resource allocation assessment layer, the simulated government policy prediction factor P is incorporated. m Requests for government services from users awaiting allocation (Q) m And the allocation of government service resources to individual D k A simulated satisfaction rating A was obtained by performing a simulated satisfaction rating calculation. m。

[0094] A simulated government service satisfaction model can be built based on a random forest model or a support vector machine (SVM) model. Historical government service data is collected and organized, including user characteristics, service request content, service resource allocation records, and corresponding user satisfaction scores. The raw data is standardized or normalized to ensure consistent numerical ranges for the model input data. Key features related to user satisfaction are extracted, such as response time, service completion rate, and resource matching degree. Feature selection methods, such as information gain or feature importance ranking, are used to select the most relevant features to improve the model's prediction accuracy and efficiency. The processed data is divided into training and validation sets. The training set is used to train the model and optimize hyperparameters. The validation set is used to test model performance, and the error index between the predicted results and actual satisfaction is calculated. Government service request features and allocation schemes are input into the model to predict satisfaction scores. Based on the model's predicted satisfaction scores, resource allocation strategies are optimized to improve the overall user experience.

[0095] Calculate all pending user government service requests Qm Corresponding simulated satisfaction rating A m The mean is used as the individual for allocating government service resources. k fitness S k ;

[0096] During the population iteration process, the non-dominated solutions generated in each iteration are stored to obtain the elite government service resource allocation population. A hypercube is then constructed for the elite government service resource allocation population to obtain the elite government service resource allocation hypercube C. t ;

[0097] Using formula Control the iterative direction of the population that allocates government service resources; among which, As the initial perturbation weights, For exponential control parameters, Hypercube C for allocating resources for elite government services t Random points within; Individual D for allocating government service resources k The position in the t-th iteration; Individual D for allocating government service resources k The displacement formula in the t-th iteration; These are random weighting coefficients, with values ​​in the range [0, 1].

[0098] When the maximum number of iterations is reached, output the individual with the highest fitness for government service resource allocation, which is the optimal individual for government service resource allocation. Based on the allocation results of the optimal individual for government service resource allocation according to the real-time request queue for government services, output the government service resource allocation result H. m。

[0099] This system introduces an elite government service resource allocation population and a hypercube, combined with a non-dominated solution storage mechanism, to effectively avoid local optima traps and ensure that the population iteration process continuously approaches the global optimum. It enhances the algorithm's search diversity and exploration capabilities by controlling the iteration direction using random weight coefficients and perturbation weights. By matching derived government service data through a prediction and adjustment layer, it simulates future resource demands and dynamically adapts to changes in different scenarios and user requests, improving allocation flexibility. In the resource allocation evaluation layer, a comprehensive evaluation is conducted by combining user requests, prediction factors, and allocation individuals to ensure that the resource allocation strategy meets real-time needs. Fitness calculation is based on simulated satisfaction scores to ensure that the resource allocation scheme takes user satisfaction as the core objective. By calculating the average satisfaction level, the allocation results of individuals are combined with the global objective, improving the overall user experience. The dynamic resource allocation layer, through the introduction of a population optimization algorithm, satisfaction-driven fitness calculation, and an elite population storage mechanism, achieves dynamic optimization and globally optimal decision-making in resource allocation. Its modular design ensures the system's scalability and flexibility, while prioritizing user satisfaction further improves the accuracy of the allocation strategy and service quality, providing strong support for an intelligent government service resource allocation system.

[0100] The special service allocation module includes special service allocation units;

[0101] The special service allocation unit is used in the special service allocation window to allocate special resources for government service requests from users to be allocated.

[0102] The specific steps for allocating special resources to users requesting government services include:

[0103] Obtain special government service resources to be allocated;

[0104] Service resources are matched based on the government service requests of users to be assigned, resulting in matched special government service resources. By accurately matching the specific attributes of the government service requests of users to be assigned with special government service resources, it is ensured that the resources can meet the personalized needs of users and improve service quality. The resource matching process is designed for special service requests to avoid ineffective or wasteful resource allocation. A special resource allocation is completed for each special government service request to be assigned.

[0105] A separate special service allocation window and matching mechanism are designed specifically for handling special government service requests, avoiding resource contention with ordinary requests and effectively improving the processing efficiency of special requests. The separate processing flow reduces the complexity of the overall system resource scheduling and ensures rapid response to special services. The special service allocation module focuses on meeting the requirements of users with high priority or special needs, significantly improving user satisfaction. The targeted allocation mechanism avoids user dissatisfaction caused by resource competition or delays.

[0106] It should be understood that those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims. Parts not described in detail in this specification are prior art known to those skilled in the art.

Claims

1. A service allocation system for government service halls based on big data analytics, characterized in that: include: The service category classification module includes a data acquisition unit and a category classification unit; the data acquisition unit is used to acquire government service requests from users to be assigned. The government service requests from users to be assigned include user identifiers, service request content, service request time, and special user tags; The category division unit is used to analyze the government service requests of users to be assigned and the government service allocation division model to obtain the service categories of users to be assigned. Among them, the service categories of users to be assigned include special service categories and non-special service categories. Based on the service categories of users to be assigned, the government service requests of users to be assigned are allocated. If the service category of users to be assigned is a non-special service category, the government service requests of users to be assigned are transferred to the real-time service allocation window for service request; otherwise, the government service requests of users to be assigned are transferred to the special service allocation window for service request. The real-time service allocation module includes a queue acquisition unit and a service allocation unit. The queue acquisition unit is used to input user requests for government services to be allocated into the real-time government service request queue within the real-time service allocation window, awaiting resource allocation. The real-time government service request queue contains user requests for government services Q. m m = 1, 2, ..., M; M is the total number of user government service requests to be allocated in the real-time government service request queue; the service allocation unit is used to allocate resources based on the real-time government service request queue and the real-time government service allocation model, and obtain the government service resource allocation result H. m According to the allocation results of government service resources H m Perform service allocation; The special service allocation module includes a special service allocation unit; the special service allocation unit is used to allocate special resources for the government service requests of users to be allocated in the special service allocation window; The government service allocation and classification model includes an information extraction layer, a category classification layer, and a result output layer. The category segmentation layer is used to perform category analysis based on the feature set X of the government service request of the user to be assigned, and to obtain the service category of the user to be assigned. The specific steps for performing category analysis at the category segmentation level include: The category classification layer includes a multi-indicator evaluation layer and a comprehensive category classification layer; In the multi-indicator evaluation layer, based on the feature item X of the government service request from the user to be assigned... n Category label assignment is performed to obtain the feature item label W for government service requests. n ; In the comprehensive category classification layer, based on the feature item label W of all government service requests n Perform feature fusion to obtain the user service category to be assigned; The specific steps for training a multi-metric evaluation layer include: The multi-index evaluation layer uses the BP neural network model as the basic model; it collects several sets of government service request label pairs as training samples, each set of government service request label pairs as training samples contains target feature labels and corresponding government service request feature items; it combines several sets of government service request label pairs as training samples to obtain the government service request label pairs training set; The BP neural network model is trained on the training set using government service request tags to obtain an initial multi-index evaluation layer. The initial multi-index evaluation layer is then evaluated. If the initial multi-index evaluation layer passes the model evaluation, it is used as the multi-index evaluation layer in the category division layer. Otherwise, the model is trained on the training set using government service request tags. The specific steps for feature fusion in the comprehensive category segmentation layer include: Based on the feature tags W of all government service requests n Construct a feature matrix of government service request tags; Based on the feature matrix of government service request labels and the local attention mechanism, neighborhood features are captured to obtain the local government service request label features F. l ; Based on the feature matrix of government service request labels and the global attention mechanism, feature correlation is calculated to obtain the global government service request label feature F. g ; Using the formula F=[F l ;F g [The partial government service request tag feature F] l and global government service request tag feature F g The data is then combined to obtain integrated government service request tag features; The integrated government service request label features are input into a linear layer for category mapping to obtain the user service category to be assigned.

2. The government service hall allocation system based on big data analysis according to claim 1, characterized in that, The information extraction layer is used to convert the user identifier, service request content, service request time, and special user tags in the government service request of the user to be assigned into a feature set X of the government service request of the user to be assigned, X={X n |n=1, 2, ..., N}; where X n The feature items for the government service requests of users to be assigned are N, where N is the total number of feature items in the feature set X of the government service requests of users to be assigned. The output layer is used to output the service categories to be assigned to users.

3. The government service hall allocation system based on big data analysis according to claim 2, characterized in that, The specific steps for training the comprehensive category-class splitting layer include: Collect several sets of matrix category training samples. Each set of matrix category training samples contains a label feature matrix and the corresponding category classification result. Combine the several sets of matrix category training samples to obtain the matrix category training set. The model is trained using the matrix category training set to obtain an initial comprehensive category partitioning layer. The initial comprehensive category partitioning layer is then evaluated. If the initial comprehensive category partitioning layer passes the model evaluation, it is used as the comprehensive category partitioning layer in the category partitioning layer; otherwise, the model is trained again using the matrix category training set.

4. The government service hall allocation system based on big data analysis according to claim 3, characterized in that, The real-time allocation model for government services includes a feature information acquisition layer, a dynamic resource allocation layer, and a resource allocation result output layer. The feature information acquisition layer is used to obtain the government service request Q of the user to be assigned using big data technology. m Historical government service data is used to obtain derived government service data Y. m ; The dynamic resource allocation layer is used to allocate government service requests from all users to be allocated (Q). m and derived government service data Y m Dynamic resource allocation is performed using a swarm optimization algorithm to obtain the government service resource allocation result H. m ; The resource allocation result output layer is used to output the government service resource allocation results H m .

5. A government service hall allocation system based on big data analysis according to claim 4, characterized in that, The specific steps for dynamic resource allocation in the dynamic resource allocation layer include: Obtain currently available government service resources to be allocated; Based on all pending user requests for government services Q m and derived government service data Y m Construct K entities for allocating government service resources, D k k=1,2,…,K; each individual D allocates government service resources. k It contains the allocation results of the currently pending government service resources based on the real-time request queue for government services; and it allocates K government service resources to individual D. k Combining these elements, we obtain an iterative population for allocating government service resources; the number of iterations is set to t, t=1, 2, ..., t max ;t max This represents the maximum number of iterations. Construct a simulated government service satisfaction model; the simulated government service satisfaction model includes a prediction and adjustment layer and a resource allocation evaluation layer; in the prediction and adjustment layer, based on derived government service data Y... m By performing predictor matching, the simulated government predictor factor P is obtained. m In the resource allocation assessment layer, the simulated government policy prediction factor P is incorporated. m Requests for government services from users awaiting allocation (Q) m and the allocation of government service resources to individual D k A simulated satisfaction rating A was obtained by performing a simulated satisfaction rating calculation. m ; Calculate all pending user government service requests Q m Corresponding simulated satisfaction rating A m The mean is used as the individual for allocating government service resources. k fitness S k ; During the population iteration process, the non-dominated solutions generated in each iteration are stored to obtain the elite government service resource allocation population. A hypercube is then constructed for the elite government service resource allocation population to obtain the elite government service resource allocation hypercube C. t ; Using formula Control the iterative direction of the population that allocates government service resources; among which, As the initial perturbation weights, For exponential control parameters, Hypercube C for allocating resources for elite government services t Random points within; Individual D for allocating government service resources k The position in the t-th iteration; Individual D for allocating government service resources k The displacement formula in the t-th iteration; These are random weighting coefficients, with values ​​in the range [0, 1]. When the maximum number of iterations is reached, output the individual with the highest fitness for government service resource allocation, which is the optimal individual for government service resource allocation. Based on the allocation results of the optimal individual for government service resource allocation according to the real-time request queue for government services, output the government service resource allocation result H. m .

6. A government service hall allocation system based on big data analysis according to claim 5, characterized in that, The specific steps for allocating special resources to users requesting government services include: Obtain special government service resources to be allocated; Based on the government service requests of users to be assigned, service resources are matched to obtain matched special government service resources; and special resource allocation is completed for the special government service resources to be assigned based on the matched special government service resources.

7. A government service hall allocation system based on big data analysis according to claim 6, characterized in that, The swarm optimization algorithm is the particle swarm optimization algorithm.

Images