Intelligent queuing optimization system based on data analysis
Through an intelligent queue optimization system based on data analysis, the historical queue data is used for business classification and correlation calculation, combined with word vector conversion technology, the matching of users to windows is optimized, and the window counter business backlog and long queue time in traditional queuing systems are solved. The problem of this has achieved a more efficient queuing process.
Patent Information
- Application Number
- CN202410917982.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-10
- Publication Date
- 2024-11-01
- Estimated Expiration
- 2044-07-10
AI Technical Summary
In the traditional queuing system, due to the different business processing capabilities and speeds of counters at different windows, the business backlog and the long queue time of users, bringing users a bad experience, and it is necessary to optimize the queuing process to improve efficiency.
Intelligent queue optimization system based on data analysis, collects historical queue data through the data acquisition module, and clusters data by the business classification module, and calculates sub-business correlation calculation module, and calculates user service satisfaction acquisition module. The adaptive rank module uses word vector transformation and cosine similarity to guide users to the most satisfactory window rank.
Through data analysis, improve the matching degree between users and window counters, optimize the queuing process, reduce user queuing time, and improve business processing efficiency.
Smart Images

Figure CN118469097B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of electronic digital data processing, and in particular to an intelligent queuing optimization system based on data analysis. Background Art
[0002] The core value of the intelligent queue optimization system lies in achieving optimal resource allocation through accurate data analysis, reducing user queue time, improving service quality and customer satisfaction. By analyzing historical queue data, personnel can be arranged reasonably, business processes can be optimized, differentiated services can be implemented, and users can be notified in advance, thereby reducing queue congestion and improving efficiency.
[0003] When users use the intelligent queuing system, they often need to go to a designated window counter to handle business. However, in the traditional queuing process, users queue up in the order in which they pick up their numbers, and the window counter to handle the business is not designated in advance. Users only go to an idle window counter to handle business in the order of their queuing. However, since the business personnel at different window counters have different abilities and speeds in handling different types of business, it will cause a backlog of window counter business and long waiting times for users, resulting in a bad experience for users. Therefore, it is necessary to optimize the user's queuing process. Summary of the invention
[0004] The present invention provides an intelligent queuing optimization system based on data analysis to solve the existing problems.
[0005] The intelligent queuing optimization system based on data analysis of the present invention adopts the following technical solutions:
[0006] An embodiment of the present invention provides an intelligent queuing optimization system based on data analysis, which includes the following modules:
[0007] A data acquisition module is used to acquire historical queuing data, wherein the historical queuing data at least includes user demand information, service type, window counter number, processing time, and processing order;
[0008] The service classification module is used to cluster the service types of the historical queue data acquired according to the service processing time to obtain several service clusters, and divide each service cluster into several sub-service clusters, each service cluster contains several services, each sub-service cluster contains several sub-services, and each sub-service is assigned a number;
[0009] The business handling correlation calculation module is used to draw the business handling correlation curve of all historical queue data according to the handling sequence and sub-business number, calculate the handling correlation of any two sub-businesses, and screen out the related sub-businesses of each sub-business;
[0010] The user service satisfaction acquisition module is used to obtain the attention of any user to each service based on the number of historical queuing data, the number of sub-service clusters, and the distance between sub-service clusters. According to the user's attention to each service, the processing time of the service and sub-service, and the number of services and sub-service clusters, the user's service satisfaction with each window counter is obtained;
[0011] The adaptive queuing module is used to obtain the user demand information of any user in the queue, convert the obtained user demand information and the user demand information of all users in the historical queuing data into word vectors to obtain the demand information vector, and guide the user to the window with the highest service satisfaction to queue up for business through the cosine similarity between the demand information vectors.
[0012] Furthermore, the specific method of calculating the handling relevance of any two sub-businesses and screening out the associated sub-businesses of each sub-business is as follows:
[0013] According to the number of all historical queue data, all historical queue data contains sub-business and The order of processing is The number of all data, all historical queue data including sub-business and The order of processing is The number of all data, all historical queue data including sub-business and The order of processing is Among the business handling correlation curves corresponding to all the data, all the business handling correlation curves are The average value of the distance, all historical queue data contains sub-services and The order of processing is Among the business handling correlation curves corresponding to all the data, all the business handling correlation curves are The average value of the distance is calculated Sub-business and The relevance between the sub-businesses;
[0014] If the handling correlation between any two sub-businesses is greater than a preset correlation threshold, one of the two sub-businesses is recorded as a correlated sub-business of the other sub-business.
[0015] Furthermore, the Sub-business and The specific calculation method of the handling correlation between sub-businesses is as follows:
[0016]
[0017] In the formula, Indicates the number of historical queue data Sub-business and The relevance between the sub-businesses; Indicates the number of all historical queue data; Indicates that all historical queue data contains sub-services and The order of processing is The number of all data; Indicates that all historical queue data contains sub-services and The order of processing is The number of all data; Indicates that all historical queue data contains sub-services and The order of processing is Among the business handling correlation curves corresponding to all the data, all the business handling correlation curves are The average value of the distance; Indicates that all historical queue data contains sub-services and The order of processing is Among the business handling correlation curves corresponding to all the data, all the business handling correlation curves are The average value of the distance; Represents an exponential function with a natural constant as base.
[0018] Furthermore, the method of obtaining the attention of any user to each service according to the number of historical queuing data, the number of sub-service clusters, and the distance between sub-service clusters, and obtaining the service satisfaction of any user to each window counter according to the attention of the user to each service, the processing time of the service and sub-service, and the number of service and sub-service clusters, includes the following specific methods:
[0019] Get the User in The set formed by the Euclidean distances between each sub-business cluster and other sub-business clusters is recorded as the first set of sub-business clusters. The number of sub-service clusters in the historical queue data of each user, the number of all historical queue data, In the historical queue data of users The number of data in each sub-service cluster, User in The variance of all Euclidean distances in the first set of sub-service clusters is calculated. The degree of attention paid to the business handled by each user;
[0020] According to the historical queue data The number of data of each user, the average number of data of all users in the historical queue data, The attention of each user to the business he / she handles, The number of related sub-businesses among the sub-businesses processed in the window counter, the number of all sub-business clusters in all business clusters, The number of business types processed at each window counter, the ... Window counter processing The average processing time of the sub-business cluster where the business type belongs, the average processing time of the historical queue data Window counter processing The average processing time of each business type is calculated. User for service satisfaction at each window counter.
[0021] Furthermore, the The specific calculation method of a user's business attention is as follows:
[0022]
[0023] In the formula, Indicates The degree of attention paid to the business handled by each user; Indicates The number of sub-service clusters in the historical queuing data of a user; Indicates the number of all historical queue data; Indicates In the historical queue data of users The number of data of each sub-service cluster; Indicates User in The variance of all Euclidean distances in the first set of sub-business clusters; Represents an exponential function.
[0024] Furthermore, the User for The specific calculation method for the service satisfaction of each window counter is:
[0025]
[0026] In the formula, Indicates User for Service satisfaction at each window counter; Indicates the number of historical queue data The amount of data for each user; Represents the average data quantity of all users in the historical queue data; Indicates The degree of attention paid by individual users to the services they handle; Indicates The number of related sub-businesses among the sub-businesses processed in each window counter; Indicates the number of all sub-service clusters in all service clusters; Indicates The number of business types handled at each window counter; Indicates the number of historical queue data Window counter processing The average processing time of the sub-business cluster where each business type belongs; Indicates that all users in the historical queue data are Window counter processing The average processing time for each business type.
[0027] Furthermore, the method of guiding the user to the window with the highest service satisfaction to queue up for business through the cosine similarity between the demand information vectors includes the following specific methods:
[0028] The word vector corresponding to the user demand information in the user's historical queuing data is recorded as the historical demand information vector; the cosine similarity between the demand information vector of the qth user and the historical demand information vector is used to match the users, and A users with the largest cosine similarity with the qth user in the historical queuing data are screened out and recorded as similar users of the qth user, where A is a preset similar user number parameter, and the mean service satisfaction of all similar users of the qth user on each window counter is obtained, and the window counter with the largest mean service satisfaction is taken as the most suitable window counter corresponding to the qth user, thereby guiding the qth user to the most suitable window counter to queue up for business.
[0029] Furthermore, the specific method for obtaining the demand information vector is:
[0030] The Word2vec model is used to convert user demand information into word vectors, which are recorded as demand information vectors.
[0031] Furthermore, the specific method of drawing the service processing association curve of all historical queuing data according to the processing sequence and the sub-service number is as follows:
[0032] A business processing association curve of all historical queuing data is drawn, wherein the horizontal axis corresponding to the business processing association curve represents the processing order, and the vertical axis represents the number of the corresponding sub-business.
[0033] Furthermore, the business types of the acquired historical queuing data are clustered according to the business handling time to obtain several business clusters, and each business cluster is divided into several sub-business clusters, including the specific method of:
[0034] The K-Means clustering algorithm is used to cluster the business types of the historical queuing data according to the business processing time, and several clusters are obtained, which are recorded as business clusters, each of which contains several businesses;
[0035] Each business cluster is divided into several sub-business clusters, so that the processing time of the sub-businesses in any sub-business cluster is the same.
[0036] The beneficial effects of the technical solution of the present invention are: obtaining historical queuing data, classifying the obtained historical queuing data, drawing business processing association curves of all historical queuing data, calculating the processing association of any two sub-businesses, screening out sub-businesses with processing associations, obtaining user service satisfaction with each window counter according to the attention of different users to different businesses, obtaining the demand information of different users when handling business, converting the format of all demand information and the demand information of all users in the historical queuing data, determining the window counter with the highest user service satisfaction through cosine similarity, guiding users to the window with the highest service satisfaction to queue up for business, and improving the matching degree between users and window counters through analysis of historical queuing data, optimizing the queuing process, reducing the queuing time for users to handle business, and improving business handling efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0037] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0038] Figure 1 It is a structural block diagram of the intelligent queuing optimization system based on data analysis of the present invention. DETAILED DESCRIPTION
[0039] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation method, structure, features and effects of the intelligent queuing optimization system based on data analysis proposed by the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.
[0040] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
[0041] The specific scheme of the intelligent queuing optimization system based on data analysis provided by the present invention is described in detail below with reference to the accompanying drawings.
[0042] See also Figure 1 , which shows a structural block diagram of an intelligent queuing optimization system based on data analysis provided by an embodiment of the present invention, the system includes the following modules:
[0043] The data acquisition module is used to obtain historical queue data.
[0044] It should be noted that, since this embodiment optimizes the queuing system through data analysis, it is necessary to first obtain historical queuing data from the queuing system.
[0045] In order to implement the intelligent queuing optimization system based on data analysis proposed in this embodiment, it is first necessary to collect historical queuing data. The specific process is as follows:
[0046] The historical queue data of several users are obtained from the database of the existing queue system, wherein each user corresponds to several historical queue data, and the historical queue data specifically includes user ID number, user demand information, business type, window counter number, processing time, queuing time, processing order and other information.
[0047] The processing sequence indicates the order in which different services are processed when a user processes the services.
[0048] It should be noted that since each user queues for different times and the content of the services handled each time is not exactly the same, the relevant information of each user's queue is recorded as a historical queue data, so each user will correspond to several historical queue data.
[0049] So far, the historical queuing data is obtained through the above method.
[0050] The business classification module is used to cluster the business types of the historical queuing data obtained according to the business processing time to obtain several business clusters, and divide each business cluster into several sub-business clusters. Each business cluster contains several businesses, each sub-business cluster contains several sub-businesses, and each sub-business is assigned a number.
[0051] It should be noted that the historical queuing data includes different types of services handled by different users, and often only major business options are provided for users to choose from. There is no subdivision of the different sub-services included in each major business category, and the user's satisfaction with the service of each window counter is also related to the business handled. Therefore, it is necessary to first divide the business types according to the historical queuing data.
[0052] Specifically, first, the K-Means clustering algorithm is used, with the preset K value being N. The business types of the acquired historical queuing data are clustered according to the business processing time, and several clusters are obtained, which are recorded as business clusters. Each business cluster contains several business types, and the business processing time of all types in the cluster is similar.
[0053] Then, the service types are numbered, and each service cluster is divided into several sub-service clusters, so that the processing time of all services in any sub-service cluster is the same, wherein each sub-service cluster contains several sub-services, and each sub-service is assigned a number.
[0054] It should be noted that since different types of services have different processing times, the user's queuing time will also be different. The user's queuing time further affects the user's satisfaction with the service of the window counter. Therefore, the services in the historical queuing data are first clustered based on the service processing time.
[0055] It should be noted that, in this embodiment, the value of N is preset to 8 based on experience, and can be adjusted according to actual conditions, and is not specifically limited in this embodiment.
[0056] So far, several service clusters and several sub-service clusters are obtained through the above method.
[0057] The business processing correlation calculation module is used to draw the business processing correlation curve of all historical queuing data according to the processing sequence and sub-business number, calculate the processing correlation between any two sub-businesses, and screen out the related sub-businesses of each sub-business.
[0058] It should be noted that when users handle multiple services of different types, they often handle them in a certain order, that is, there is a certain correlation between different service types. For example, when users handle services, they will involve service types. Sub-business and business type Sub-business ,and and Having a certain order of precedence indicates that the two businesses have a certain correlation. Therefore, when optimizing the queuing system, the correlation between different types of businesses needs to be considered.
[0059] Specifically, a business handling correlation curve of all historical queuing data is drawn, wherein the horizontal axis corresponding to the business handling correlation curve represents the time of business handling, that is, the order of handling different businesses when handling the business, and the vertical axis represents the number of the corresponding sub-business;
[0060] So the Sub-business and The calculation method of the handling correlation between sub-businesses is:
[0061]
[0062] In the formula, Indicates the number of historical queue data Sub-business and The relevance between the sub-businesses; Indicates the number of all historical queue data; Indicates that all historical queue data contains sub-services and The order of processing is The number of all data; Indicates that all historical queue data contains sub-services and The order of processing is The number of all data; Indicates that all historical queue data contains sub-services and The order of processing is Among the business handling correlation curves corresponding to all the data, all the business handling correlation curves are The average value of the distance; Indicates that all historical queue data contains sub-services and The order of processing is Among the business handling correlation curves corresponding to all the data, all the business handling correlation curves are The average value of the distance; Represents an exponential function with a natural constant as base.
[0063] It should be noted that Indicates that the historical queue data contains sub-services and The order of processing is The number of all data, if the value is larger, it means that the sub-business and The higher the frequency of co-occurrence in historical queue data, the stronger the correlation between the two; Indicates that the sub-business and The order of handling determines the relevance of the two. If the value is smaller, it means that the sub-business and The more regular the processing order in the historical queue data, the stronger the correlation between the two.
[0064] It should be noted that the The distance is obtained by using a DTW (Dynamic Time Warping) algorithm, wherein the DTW algorithm is an existing algorithm, and thus is not described in detail in this embodiment.
[0065] According to the above steps, the handling correlation between any two sub-businesses is calculated, and the preset correlation threshold is M. If the handling correlation between any two sub-businesses is greater than the preset correlation threshold, one of the two sub-businesses is recorded as a related sub-business of the other sub-business.
[0066] It should be noted that, in this embodiment, the value of M is preset to 0.7 based on experience, and can be adjusted according to actual conditions, and is not specifically limited in this embodiment.
[0067] At this point, the business types with handling relevance are obtained through the above method.
[0068] The user service satisfaction acquisition module is used to obtain the attention of any user to each service based on the number of historical queuing data, the number of sub-service clusters, and the distance between sub-service clusters; and to obtain the service satisfaction of any user with each window counter based on the user's attention to each service, the processing time of the service and sub-service, and the number of service and sub-service clusters.
[0069] It should be noted that in the process of obtaining user satisfaction with the services of each window counter, due to different business types, users' focus is different. Therefore, it is necessary to obtain user satisfaction with the services of each window counter separately according to the user's focus on different businesses.
[0070] Specifically, since each user has different concerns about services, for example, some users frequently handle services, and the services they handle are often similar, that is, the services that the user cares about are relatively similar. The user's concern about services affects the service satisfaction of services handled at different window counters. Therefore, through the The service satisfaction of each window counter is quantified by the user's attention to the business he / she handles.
[0071] Get the User in The Euclidean distance between all historical queue data in the sub-service cluster is recorded as the first set of sub-service clusters, The calculation method of a user's attention to the business he handles is:
[0072]
[0073] In the formula, Indicates The degree of attention paid to the business handled by each user; Indicates The number of sub-service clusters in the historical queuing data of a user; Indicates the number of all historical queue data; Indicates In the historical queue data of users The number of data of each sub-service cluster; Indicates User in The variance of all Euclidean distances in the first set of sub-business clusters; Represents an exponential function.
[0074] It should be noted that Indicates The weight value of a sub-service cluster is represented by the amount of data in the sub-service cluster. The larger the value, the higher the user's attention to the sub-service cluster, and the greater the weight of the sub-service cluster in the user's service attention; Indicates the distribution discreteness of the sub-service cluster. If the value is smaller, it means that the data distribution in the sub-service cluster is more uniform, that is, the randomness of the sub-service cluster is smaller, and the corresponding service attention is also smaller.
[0075] Then calculate the User for The service satisfaction of each window counter is calculated as follows:
[0076]
[0077] In the formula, Indicates User for Service satisfaction at each window counter; Indicates the number of historical queue data The amount of data for each user; Represents the average data quantity of all users in the historical queue data; Indicates The degree of attention paid by individual users to the services they handle; Indicates The number of related sub-businesses among the sub-businesses processed in each window counter; Indicates the number of all sub-service clusters in all service clusters; Indicates The number of business types handled at each window counter; Indicates the number of historical queue data Window counter processing The average processing time of the sub-business cluster where each business type belongs; Indicates the number of historical queue data Window counter processing The average processing time for each business type.
[0078] It should be noted that Indicates User for The service satisfaction weight of the window counter is The more attention a user pays to the business he / she handles, the more attention the user pays to the business he / she handles, and the more attention the user pays to the ability of the window counter to handle the business. Conversely, the user pays more attention to the time it takes to handle the business at the window counter. Indicates The business capability of a window counter. If the window counter involves more business types, that is, the larger the value, the stronger the business processing capability of the window counter; Indicates The ability of a window counter to handle various types of business. If the time for handling the business is shorter, it means that the window counter has a stronger ability to handle the business.
[0079] So far, the user's service satisfaction with each window counter is obtained through the above method.
[0080] The adaptive queuing module is used to obtain the user demand information of any user in the queue, convert the obtained user demand information and the user demand information of all users in the historical queuing data into word vectors to obtain the demand information vector, and guide the user to the window with the highest service satisfaction to queue up for business through the cosine similarity between the demand information vectors.
[0081] Specifically, obtain the user demand information of any user in the queue. Since the obtained user demand information is usually text data, it is necessary to use the Word2vec model to convert the text information into word vectors to obtain the demand information vectors of different users. Use the same method to convert the demand information of all users in the historical queue data into word vectors, which are recorded as historical demand information vectors.
[0082] For the cosine similarity between the demand information vector of the qth user and the historical demand information vector, matching is performed between users, and A users with the largest cosine similarity with the qth user in the historical queuing data are screened out and recorded as similar users of the qth user, where A is the preset similar user number parameter. The mean service satisfaction of all similar users of the qth user on each window counter is obtained, and the window counter with the largest mean service satisfaction is taken as the most suitable window counter corresponding to the qth user, so as to guide the qth user to the most suitable window counter to queue up for business.
[0083] It should be noted that the Word2vec model is an existing technology and will not be described in detail here.
[0084] At this point, this embodiment is completed.
[0085] It should be noted that the The model is only used to represent negative correlation and constrain the output of the model to be in In the specific implementation, it can be replaced by other models with the same purpose. This embodiment is only based on The model is described as an example without any specific limitation. is the input to the model.
[0086] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.
Claims
1. Intelligent queuing optimization system based on data analysis, characterized by: The system includes the following modules: A data acquisition module, used to acquire historical queuing data, wherein the historical queuing data at least includes user demand information, service type, window counter number, processing time and processing order of the user; The service classification module is used to cluster the service types of the historical queue data acquired according to the service processing time to obtain several service clusters, and divide each service cluster into several sub-service clusters, each service cluster contains several services, each sub-service cluster contains several sub-services, and each sub-service is assigned a number; The business handling correlation calculation module is used to draw the business handling correlation curve of all historical queue data according to the handling sequence and sub-business number, calculate the handling correlation of any two sub-businesses, and screen out the related sub-businesses of each sub-business; The user service satisfaction acquisition module is used to obtain the attention of any user to each service based on the number of historical queuing data, the number of sub-service clusters, and the distance between sub-service clusters. According to the user's attention to each service, the processing time of the service and sub-service, and the number of services and sub-service clusters, the user's service satisfaction with each window counter is obtained; The adaptive queuing module is used to obtain the user demand information of any queuing user, convert the obtained user demand information and the user demand information of all users in the historical queuing data into word vectors, obtain the demand information vector, and guide the user to the window with the highest service satisfaction to queue up for business through the cosine similarity between the demand information vectors; The method of obtaining the attention of any user to each service according to the number of historical queuing data, the number of sub-service clusters, and the distance between sub-service clusters, and obtaining the service satisfaction of any user to each window counter according to the attention of the user to each service, the processing time of the service and sub-service, and the number of service and sub-service clusters, includes the following specific methods: Get the User in The set formed by the Euclidean distances between each sub-business cluster and other sub-business clusters is recorded as the first set of sub-business clusters. The number of sub-service clusters in the historical queue data of each user, the number of all historical queue data, In the historical queue data of users The number of data in each sub-service cluster, User in The variance of all Euclidean distances in the first set of sub-service clusters is calculated. The degree of attention paid to the business handled by each user; According to the historical queue data The number of data of each user, the average number of data of all users in the historical queue data, The attention of each user to the business he / she handles, The number of related sub-businesses among the sub-businesses processed in the counter, the number of all sub-business clusters in all business clusters, The number of business types processed at each window counter, the ... Window counter processing The average processing time of the sub-business cluster where the business type belongs, the average processing time of the historical queue data Window counter processing The average processing time of each business type is calculated. User for Service satisfaction at each window counter; The said The specific calculation method of a user's business attention is as follows: In the formula, Indicates The degree of attention paid to the business handled by each user; Indicates The number of sub-service clusters in the historical queuing data of a user; Indicates the number of all historical queue data; Indicates In the historical queue data of users The number of data of each sub-service cluster; Indicates User in The variance of all Euclidean distances in the first set of sub-business clusters; represents the exponential function; The said User for The specific calculation method for the service satisfaction of each window counter is: In the formula, Indicates User for Service satisfaction at each window counter; Indicates the number of historical queue data The amount of data for each user; Represents the average data quantity of all users in the historical queue data; Indicates The degree of attention paid by individual users to the services they handle; Indicates The number of related sub-businesses among the sub-businesses processed in each window counter; Indicates the number of all sub-service clusters in all service clusters; Indicates The number of business types handled at each window counter; Indicates the number of historical queue data Window counter processing The average processing time of the sub-business cluster where each business type belongs; Indicates that all users in the historical queue data are Window counter processing Average processing time for each business type; The specific method of calculating the handling relevance of any two sub-businesses and screening out the associated sub-businesses of each sub-business is as follows: According to the number of all historical queue data, all historical queue data contains sub-business and The order of processing is The number of all data, all historical queue data including sub-business and The order of processing is The number of all data, all historical queue data including sub-business and The order of processing is Among the business handling correlation curves corresponding to all the data, all the business handling correlation curves are The average value of the distance, all historical queue data contains sub-services and The order of processing is Among the business handling correlation curves corresponding to all the data, all the business handling correlation curves are The average value of the distance is calculated Sub-business and The relevance between the sub-businesses; If the handling correlation between any two sub-businesses is greater than the preset correlation threshold, one of the two sub-businesses is recorded as a related sub-business of the other sub-business; The said Sub-business and The specific calculation method of the handling correlation between sub-businesses is as follows: In the formula, Indicates the number of historical queue data Sub-business and The relevance between the sub-businesses; Indicates the number of all historical queue data; Indicates that all historical queue data contains sub-services and The order of processing is The number of all data; Indicates that all historical queue data contains sub-services and The order of processing is The number of all data; Indicates that all historical queue data contains sub-services and The order of processing is Among the business handling correlation curves corresponding to all the data, all the business handling correlation curves are The average value of the distance; Indicates that all historical queue data contains sub-services and The order of processing is Among the business handling correlation curves corresponding to all the data, all the business handling correlation curves are The average value of the distance; Represents an exponential function with a natural constant as base.
2. The intelligent queuing optimization system based on data analysis according to claim 1 is characterized in that: The specific method of guiding the user to the window with the highest service satisfaction to queue up for business through the cosine similarity between the demand information vectors includes: The word vector corresponding to the user demand information in the user's historical queuing data is recorded as the historical demand information vector; the cosine similarity between the demand information vector of the qth user and the historical demand information vector is used to match the users, and A users with the largest cosine similarity with the qth user in the historical queuing data are screened out and recorded as similar users of the qth user, where A is a preset similar user number parameter, and the mean service satisfaction of all similar users of the qth user on each window counter is obtained, and the window counter with the largest mean service satisfaction is taken as the most suitable window counter corresponding to the qth user, thereby guiding the qth user to the most suitable window counter to queue up for business.
3. The intelligent queuing optimization system based on data analysis according to claim 1 is characterized in that: The specific method for obtaining the demand information vector is: The Word2vec model is used to convert user demand information into word vectors, which are recorded as demand information vectors.
4. The intelligent queuing optimization system based on data analysis according to claim 1 is characterized in that: The specific method of drawing the business processing association curve of all historical queuing data according to the processing sequence and the sub-business number includes: A business processing association curve of all historical queuing data is drawn, wherein the horizontal axis corresponding to the business processing association curve represents the processing order, and the vertical axis represents the number of the corresponding sub-business.
5. The intelligent queuing optimization system based on data analysis according to claim 1 is characterized in that: The business types of the acquired historical queuing data are clustered according to the business processing time to obtain several business clusters, and each business cluster is divided into several sub-business clusters, including the specific method of: The K-Means clustering algorithm is used to cluster the business types of the historical queuing data according to the business processing time, and several clusters are obtained, which are recorded as business clusters, each of which contains several businesses; Each business cluster is divided into several sub-business clusters, so that the processing time of the sub-businesses in any sub-business cluster is the same.
Citation Information
Patent Citations
Queue number generating method and terminal equipment
CN109377633A
Satisfaction analysis method and device for bank counter service
CN113469562A
Data processing method and device
CN115394002A
Queuing method and device, and system, client, and server
WO2021103631A1
