A method and system for identifying duplicate power customer service tickets

Abstract

The application discloses a repeated power customer service order identification method and system, and relates to the technical field of power supply enterprise customer service system. The method comprises the following steps: obtaining customer service orders of a specified user within a preset time range, calculating the similarity of any two appeals based on text recognition, taking the customer service orders corresponding to the appeals with a similarity meeting a preset threshold as repeated appeal customer service orders, further extracting business orders within the time interval of the two repeated appeal customer service orders, constructing an information association matrix between the customer service orders and the business orders, and further judging whether the two repeated appeal customer service orders are repeated customer service orders based on the information association matrix. The combination of business orders and customer service orders solves the technical problem that the existing method only analyzes the text content of customer service orders, resulting in poor accuracy of repeated customer service order identification.

Description

Technical Field

[0001] This invention relates to the field of customer service systems for power supply companies, and in particular to a method for identifying duplicate power customer service work orders. Background Technology

[0002] Power supply companies are asset-intensive and data-intensive enterprises. They operate numerous power supply nodes and facilities, and as people's living standards improve, their service demands also rise, generating a massive amount of power service data. Currently, with the development of customer service systems, power supply companies are shifting their business services online. Online services have become a crucial channel for communication with customers, and customer service work orders, which record customer requests, form the basis of this communication. This process generates a massive volume of customer service work orders. Therefore, effectively processing these numerous work orders within the power supply company's customer service system is essential for improving customer experience and satisfaction.

[0003] Due to the numerous online channels provided by power supply companies, the significant differences in customer problem descriptions, and the involvement of multiple people from the same household, a large number of duplicate customer service work orders are generated in the power supply company's customer service system. Existing power supply company customer service systems focus on the text content of customer service work orders when identifying duplicates. They analyze the text content of two work orders from the same user, identifying work orders with high text similarity as duplicates. This method only determines whether a work order is a duplicate based on text content. When the same user expresses similar content, these similar work orders will also be incorrectly identified as duplicates, resulting in poor accuracy in identifying duplicate work orders. This leads to problems in responding to user needs and a decline in user experience or satisfaction. Summary of the Invention

[0004] This invention provides a method and system for identifying duplicate customer service work orders, which solves the technical problem that existing methods rely solely on the text content of customer service work orders, resulting in poor accuracy in identifying duplicate customer service work orders.

[0005] This invention provides a method for identifying duplicate electricity customer service work orders, comprising:

[0006] S1: Obtain customer service tickets for a specified user within a preset time range, and extract the corresponding requests for each customer service ticket;

[0007] S2: Calculate the similarity between any two requests through text recognition, and take the customer service work order corresponding to the request whose similarity meets the preset threshold as the duplicate request customer service work order;

[0008] S3: Determine whether there are any business work orders within the time interval between the establishment of the two duplicate customer service work orders. If not, determine that the two duplicate customer service work orders are duplicate customer service work orders; if so, proceed to step S4.

[0009] S4: Take the duplicate customer service work order with the earlier creation time among the two duplicate customer service work orders as the first duplicate customer service work order, and construct an information association matrix with the first duplicate customer service work order and the business work order.

[0010] S5: Determine whether the two duplicate customer service work orders are duplicate customer service work orders based on the information association matrix.

[0011] Preferably, the step of calculating the similarity between any two requests through text recognition and designating the customer service work orders corresponding to requests whose similarity meets a preset threshold as duplicate request customer service work orders specifically includes:

[0012] Establish a vector space model and use feature vectors to represent the requests in customer service work orders;

[0013] Calculate the cosine similarity of the angle between two feature vectors in the vector space. When the cosine similarity of the angle meets a preset threshold, it is determined that the requests corresponding to the two feature vectors are duplicates, and the customer service work orders corresponding to the two feature vectors are regarded as duplicate request customer service work orders.

[0014] Preferably, step S4 includes:

[0015] Extract all business processes from the business work order, and construct an information association matrix based on the first duplicate request customer service work order and all business processes in the business work order:

[0016] v = (i, j, k1)

[0017] Where i represents a business work order, j represents a business process, k1 represents the first repeat customer service work order, and v represents the relationship between the first repeat customer service work order k1 and the j-th process of business work order i.

[0018] Preferably, step S5 includes:

[0019] Calculate the impact value of the business work order on the first duplicate request customer service work order based on the aforementioned relationship;

[0020] Based on the impact value, determine whether the two duplicate customer service requests are duplicate customer service requests.

[0021] Preferably, calculating the impact value of the business work order on the first duplicate customer service work order based on the association relationship specifically includes:

[0022] The impact value of the business work order on the first repeat customer service work order is calculated based on a pre-established calculation formula:

[0023]

[0024]

[0025] Where G represents the set of unions of the impact values ​​of business work orders on customer service work orders; ∪ represents the union of the two work orders; i represents the business work orders within the time interval between the establishment of two duplicate customer service work orders; (q-p+1) is the number of business work orders i, where q≧p; j represents the business steps of business work order i, and (m-l+1) is the number of business steps j, where m≧l; k1 represents the first duplicate customer service work order; f((i,j,k1)) represents the impact value of the j-th step of the i-th business work order on the first duplicate customer service work order k1.

[0026] Preferably, determining whether the two duplicate customer service requests are duplicate customer service requests based on the influence value specifically includes:

[0027] Determine whether the set of impact values ​​G contains {Resolve}. If it does, then the two duplicate customer service requests are determined to be duplicate customer service requests; otherwise, the two duplicate customer service requests are determined to be duplicate customer service requests.

[0028] Preferably, after step S5, the method further includes:

[0029] When f((i,j,k1)) is {unresolved}, determine the business work order i and the corresponding business process j used to resolve the request in the first duplicate request customer service work order k1, and then visualize the business work order and the business process.

[0030] Preferably, the method further includes the following steps before step S2:

[0031] The customer service work orders are preprocessed, including removing customer service work orders corresponding to abnormal call numbers from among the duplicate call numbers. Abnormal call numbers include calls that are hung up without reason, internal testing calls, and other abnormal calls.

[0032] Preferably, the procedure before step S3 includes:

[0033] When the number of duplicate customer service work orders exceeds two, determine whether there are any business work orders within the time interval between the establishment time of the two duplicate customer service work orders with the earliest and latest establishment times. If there are, execute S3 for the two duplicate customer service work orders with close establishment times respectively; otherwise, determine all duplicate customer service work orders as duplicate customer service work orders.

[0034] This invention also provides a system for identifying duplicate electricity customer service work orders, comprising:

[0035] The acquisition module is used to acquire customer service tickets of a specified user within a preset time range; each customer service ticket contains a request.

[0036] The first identification module is used to calculate the similarity between any two requests through text recognition, and to identify the customer service work orders corresponding to the requests whose similarity meets a preset threshold as duplicate request customer service work orders.

[0037] The judgment module is used to determine whether there is a business work order within the time interval between the establishment of two duplicate customer service work orders. If not, the two duplicate customer service work orders are determined to be duplicate customer service work orders. If there is, the duplicate customer service work order with the earlier establishment time is taken as the first duplicate customer service work order, and an information association matrix is ​​constructed between the first duplicate customer service work order and the business work order.

[0038] The second identification module is used to determine whether the two duplicate customer service work orders are duplicate customer service work orders based on the information association matrix.

[0039] The present invention provides a method for identifying duplicate customer service work orders. This method acquires customer service work orders from a specified user within a preset time range, then uses text recognition to determine the similarity of user requests within these work orders. Work orders with similarity scores meeting a preset threshold are identified as duplicate request work orders. By acquiring business work orders established within the time interval between two duplicate request work orders, an information association matrix is ​​constructed between the earlier-established duplicate request work order and the business work order. This association matrix is ​​then used to determine whether the two duplicate request work orders are indeed duplicates. By combining the association matrix to identify the relationship between customer service work orders and business work orders, and thus combining them, the method solves the problem that existing methods for identifying duplicate customer service work orders only focus on the text of the work order itself and cannot accurately distinguish duplicate work orders, thereby improving the accuracy of duplicate customer service work order identification. Attached Figure Description

[0040] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 A flowchart illustrating a method for identifying duplicate power customer service work orders provided in an embodiment of the present invention;

[0042] Figure 2 This is a structural diagram of a system for identifying duplicate power customer service work orders provided in an embodiment of the present invention. Detailed Implementation

[0043] This invention provides a method for identifying duplicate customer service work orders, which solves the technical problem that existing methods rely solely on the text content of customer service work orders for identification, resulting in poor accuracy in identifying duplicate customer service work orders.

[0044] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only a part of the embodiments of this invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0045] Explanation of related terms:

[0046] Customer service work orders: various work orders that record customer inquiries, customer requests, and business processing instructions.

[0047] Work orders: Various work orders for handling user requests or internal events in the power system.

[0048] Some business work orders are created to handle requests from customer service work orders. The process involves creating a customer service work order first, and then creating a business work order to process the requests within that work order. When a request in an earlier customer service work order has already been resolved, a request in a later customer service work order may be similar to, but not identical to, the earlier request. Such later work orders are merely similar to, rather than duplicates, customer service work orders. Existing text recognition methods cannot distinguish between duplicate and similar customer service work orders, and may identify some similar work orders as duplicates.

[0049] For easier understanding, please refer to Figure 1 Embodiment 1 of this application provides a method for identifying duplicate electricity customer service work orders, including:

[0050] Step S1: Obtain customer service work orders for the specified user within a preset time range, and extract the corresponding requests for each customer service work order.

[0051] Customer service plays a vital role in the development of power companies. Customer service work orders, which record customer requests, are a crucial foundation for communication with electricity customers. These work orders vary widely. Taking user A as an example, user A's customer service work orders can be categorized into several types, with each type containing multiple work orders. Step S1 can retrieve customer service work orders of the same type from a specified user within a specified time period. For example, retrieving user A's complaint-type customer service work orders from September 3rd to September 5th, 2022. It should be noted that a customer service work order typically contains only one request, but a request can contain one or more requirements; this application does not impose specific limitations.

[0052] In step S1, the customer service ticket includes information about the requesting party (customer information), the time of the request, and the specific content of the request. This information is stored in text format and together constitutes the text content of the customer service ticket.

[0053] Before executing step S2, the obtained customer service work orders can be preprocessed, including: removing customer service work orders corresponding to abnormal call numbers from among the duplicate call numbers. Abnormal call numbers include those that are hung up without reason, internal testing, and other abnormal calls. The above processing of the obtained customer service work orders can reduce the workload of the subsequent identification process.

[0054] Step S2: Calculate the similarity between any two requests through text recognition, and designate the customer service work orders corresponding to the requests whose similarity meets a preset threshold as duplicate request customer service work orders.

[0055] Due to the power supply department providing multiple online channels for handling services, the significant differences in the content of customer complaints, and the fact that multiple people from the same household may submit the same request, duplicate customer service work orders often occur. By performing text recognition on the requests in the text of customer service work orders, customer service work orders with similarity to the requests that meet a preset threshold can be identified as duplicate request customer service work orders.

[0056] In the process of text recognition, a vector space model can be established to use feature vectors to represent the requests in customer service work orders and to determine the similarity between requests.

[0057] In this step, the customer service work orders obtained in step S1 are analyzed to extract those whose request similarity meets a preset threshold. It is understandable that two, three, or more customer service work orders may have requests with similarity meeting the preset threshold. For example, if three customer service work orders all request "electric meter repair," then the similarity of the customer requests in these three work orders meets the preset threshold.

[0058] At the same time, after identifying the request, the work orders can be categorized according to the request, and the number of duplicate customer service work orders corresponding to the same request can be further counted, that is, the number of work orders in each category. For example, there are three duplicate customer service work orders for the request "electricity meter repair", and there are also multiple duplicate customer service work orders for the requests "electricity meter replacement" and "power switch replacement". The customer service work orders of the category with the largest number of requests will be processed first.

[0059] S3: Determine whether there are any business work orders within the time interval between the establishment of the two duplicate customer service work orders. If not, determine that the two duplicate customer service work orders are duplicate customer service work orders; if so, proceed to step S4.

[0060] In step S2, duplicate customer service work orders of the same user within a preset time range are extracted. There may be business between two duplicate customer service work orders, or there may be no business. When there is no business between two duplicate customer service work orders, it means that the request in the duplicate customer service work order created earlier must not have been resolved. Then these two duplicate customer service work orders are duplicate customer service work orders.

[0061] When there are more than two duplicate customer service tickets for the same type of request, i.e., more than two duplicate customer service tickets matching the preset conditions are identified, the two customer service tickets with the earliest and latest creation times are extracted first and second as duplicate customer service tickets, respectively. It is then determined whether there are any business orders within the interval between the earliest and latest creation times of these two duplicate customer service tickets. If not, it means that the request in the first duplicate customer service ticket cannot be resolved. In this case, all duplicate customer service tickets can be directly identified as duplicate customer service tickets without further pairwise judgment. However, when there are business orders within the interval between the earliest and latest creation times of these two duplicate customer service tickets, to determine which duplicate customer service ticket the business order is specifically used to resolve, it is necessary to pairwise judge the duplicate customer service tickets separately. Judging the two duplicate customer service tickets with adjacent creation times is the most direct and efficient method. Further, S3 is executed for the two duplicate customer service tickets with adjacent creation times.

[0062] Meanwhile, as described in step S2, after classifying work orders according to requests, the type of request with the most duplicate requests is usually processed first. For example, if the request "electricity meter repair" contains three work orders and the request "electricity meter replacement" contains four work orders, then the duplicate customer service work orders corresponding to the "electricity meter replacement" type of request will be processed first.

[0063] S4: Take the duplicate customer service work order with the earlier creation time as the first duplicate customer service work order, and construct an information association matrix with the first duplicate customer service work order and the business work order.

[0064] When business is generated in two duplicate customer service tickets, it means that the complaint in the earlier duplicate customer service ticket (the first duplicate customer service ticket) may or may not have been resolved.

[0065] If the request in the first duplicate request customer service ticket has already been resolved through a business ticket, then for duplicate request customer service tickets created after the business ticket creation time, although their request content meets the preset threshold with the first duplicate request customer service ticket, they can only be identified as customer service tickets with similar requests, and cannot be identified as duplicate customer service tickets.

[0066] In step S4, the purpose of extracting business work orders is to identify cases where the request meets a preset threshold but cannot be identified as a duplicate customer service work order.

[0067] Step S4, based on Step S3, extracts the business work orders within the creation time range of the two duplicate customer service work orders. Specifically, the duplicate customer service work order with the earlier creation time is designated as the first duplicate customer service work order, and the duplicate customer service work order with the later creation time is designated as the second duplicate customer service work order; obtain all business work orders within the creation time range from the first duplicate customer service work order to the second duplicate customer service work order; wherein, the business work order includes the business process.

[0068] Understandably, since a business work order may contain work orders that resolve the request in the first duplicate request customer service work order, and may also contain work orders unrelated to resolving the request in the first duplicate request customer service work order, there may be multiple work orders between the creation time of the first duplicate request customer service work order and the second duplicate request customer service work order. Therefore, in this step, an information association matrix is ​​actually constructed between multiple business work orders and the first duplicate request customer service work order. Since each business work order typically contains several business steps, this information association matrix actually reflects the relationship between all business steps in multiple business work orders and the first duplicate request customer service work order.

[0069] S5: Determine whether the two duplicate customer service work orders are duplicate customer service work orders based on the information association matrix.

[0070] The information association matrix constructed in step S4 based on business work orders and customer service work orders can be understood as the relationship between the business steps in the business work order and the requests in the customer service work order. The completion of a certain business step in the business work order will affect the requests in the customer service work order. This effect is reflected in the value of the information association matrix, which has three possible values: resolved, unresolved, and irrelevant.

[0071] Here, "resolved" means that the customer service request in the business work order is resolved after a certain business step in the business work order is completed; "unresolved" means that the customer service request in the business work order is not resolved after a certain business step in the business work order is completed; "irrelevant" means that the completion of a certain business step in the business work order has no impact on resolving the customer service request in the business work order. In subsequent implementations, the value of the above information association matrix is ​​recorded as the impact value of the business work order on the customer service work order.

[0072] By establishing an information association matrix, the impact value of each business order on the first duplicate customer service order is calculated. Then, based on the impact value, it is determined whether the duplicate customer service order is a similar customer service order or a duplicate customer service order. It can be understood that when a business order in the information association matrix is ​​used to resolve the request in the first duplicate customer service order, and the process of that business order has just been completed, it means that the request in the first duplicate customer service order has been resolved (impact value: "resolved"). Although the text recognition content of the second duplicate customer service order meets the preset condition (the similarity between the two requests meets the preset threshold) relative to the first duplicate customer service order, it is no longer judged as a duplicate customer service order, but rather as a similar customer service order. If the process of that business order has not yet been completed, it means that the request in the first duplicate customer service order has not yet been resolved (impact value: "unresolved"). In this case, the second duplicate customer service order is a duplicate customer service order relative to the first duplicate customer service order.

[0073] As can be seen from the above process, similar customer service tickets and duplicate customer service tickets contain the same requests. However, the difference lies in the fact that in similar customer service tickets, the earlier-created ticket has already been resolved, while the later-created ticket will trigger a new business ticket. That is, the request in the later-created ticket needs to be executed again. On the other hand, duplicate customer service tickets indicate that none of the corresponding customer service tickets have been resolved. If a business ticket has already been triggered, it does not need to be triggered again. This reduces the misidentification of duplicate customer service tickets in the existing technology, avoids the omission or delay in processing similar customer service tickets, improves user experience and satisfaction, and also reduces the workload of the system's customer service staff.

[0074] When a service order is not used to resolve a request in the first duplicate customer service order, regardless of whether the service order is completed or not, it is irrelevant to resolving the request in the first duplicate customer service order, and the impact value is "irrelevant".

[0075] For ease of understanding, please refer to the following example for the above explanation regarding duplicate and similar customer service tickets:

[0076] Customer A called on September 14th to inquire about the monthly electricity bill details up to midnight on September 14th. This generated a customer service ticket 1 for Customer A on September 14th, requesting to inquire about the monthly electricity bill details up to midnight on September 14th. On September 15th, Customer A called again to inquire about the monthly electricity bill details up to midnight on September 14th. This generated a customer service ticket 2 for Customer A on September 15th, requesting to inquire about the monthly electricity bill details up to midnight on September 14th.

[0077] The two customer service work orders above have the same request content, and the processing result for both requests is: output customer A's monthly electricity usage details up to 00:00 on September 14.

[0078] If the request in Customer Service Ticket 1 has not been completed before Customer Service Ticket 2 arrives, it means that the requests in Customer Service Ticket 1 and Customer Service Ticket 2 are completely identical, both requesting a query of the monthly electricity usage details up to 00:00 on September 14th. Customer Service Ticket 1 and Customer Service Ticket 2 only need to output a list with the content "Customer A's monthly electricity usage details up to 00:00 on September 14th". In this case, Customer Service Ticket 1 and Customer Service Ticket 2 are duplicate customer service tickets.

[0079] If the request in Customer Service Ticket 1 has been completed before Customer Service Ticket 2 arrives, it means that Customer Service Ticket 1 has already generated a list containing "Customer A's monthly electricity usage details up to 00:00 on September 14th". At this point, Customer Service Ticket 2 is a completely new customer service ticket compared to Customer Service Ticket 1, and Customer Service Ticket 2 will also generate a separate list containing "Customer A's monthly electricity usage details up to 00:00 on September 14th". The processing of Customer Service Ticket 2 can be understood as: "re-generating" Customer A's monthly electricity usage details up to 00:00 on September 14th. Although the request content in Customer Service Ticket 1 and Customer Service Ticket 2 is the same, the two customer service tickets need to generate their own separate lists. Therefore, Customer Service Ticket 1 and Customer Service Ticket 2 are no longer duplicate customer service tickets but similar customer service tickets.

[0080] This application combines business work orders with customer service work orders. By analyzing the impact of the business processing status between two duplicate customer service work orders on the customer service work order itself, it identifies duplicate customer service work orders. This solves the technical problem that existing methods for identifying duplicate customer service work orders only focus on the customer service work order itself and cannot distinguish between duplicate and similar customer service work orders, resulting in poor accuracy in identifying duplicate customer service work orders. By determining the relationship between customer service work orders and business work orders through an association matrix, and combining business work orders and customer service work orders to identify duplicate customer service work orders, the accuracy of duplicate customer service work order identification is improved.

[0081] In one specific implementation, step S2 calculates the similarity between any two requests using text recognition, and identifies the customer service work orders corresponding to requests with similarity scores meeting a preset threshold as duplicate request customer service work orders. Specifically, this can be done as follows:

[0082] A vector space model is established, and feature vectors are used to represent the requests in customer service work orders. Furthermore, the cosine similarity of the angle between two feature vectors in the vector space is calculated. When the cosine similarity of the angle meets a preset threshold, it is determined that the requests corresponding to the two feature vectors are duplicates, and the customer service work orders corresponding to the two feature vectors are regarded as duplicate request customer service work orders.

[0083] Understandably, customer service tickets contain customer requests, which are recorded in text form. By establishing a vector space model, feature vectors are used to represent customer requests. Then, the cosine similarity between the angles between two vectors in the space is used to measure the similarity between the requests in two customer service tickets. When the similarity meets a preset threshold, the two requests are considered duplicate requests. Furthermore, the customer service tickets corresponding to two duplicate requests are called duplicate request customer service tickets.

[0084] The method of using the cosine similarity of the angle between two vectors representing the request to measure the similarity between two customer service work orders and to determine similar customer service work orders can be implemented using the technology in the license document "CN105335496B".

[0085] In one specific implementation, step S4 may include:

[0086] Extract all business processes from the business work order, and construct an information association matrix based on the first duplicate request customer service work order and all business processes in the business work order:

[0087] v = (i, j, k1)

[0088] Where i represents a business work order, j represents a business process, k1 represents the first repeat customer service work order, and v represents the relationship between the first repeat customer service work order k1 and the j-th process of business work order i.

[0089] Business work orders typically involve handling various types of work orders for user requests or internal events within the power system. Business work orders encompass multiple types and usually correspond to customer service work orders. In step S4, after determining that one of the two customer service work orders is a duplicate request, the business work order i within the time interval between the establishment of the two duplicate request customer service work orders is obtained based on their creation time. The aforementioned business work order refers to all business work orders for the user within the interval.

[0090] Since two duplicate customer service requests may have multiple business orders within the time interval between their creation, and each business order contains several business steps, although the requests in the two duplicate customer service requests are identical, the business steps in the business order generated between customer service request k1 and customer service request k2 can only resolve the request in customer service request k1. Therefore, the business order is associated with customer service request k1. This association is represented as: v = (i, j, k1), indicating the association between customer service request k1 and the j-th step of business order i. The aforementioned customer service requests k1 and k2 correspond to the first and second duplicate customer service requests in the aforementioned embodiment one, and will not be explained further thereafter.

[0091] It's important to note that in electricity customer service, there exists a type of customer service work order that can be resolved with a single action, without requiring a separate business work order. For example, if a customer service work order requests to express gratitude to a specific customer service representative, the corresponding business work order type would be a "thank you letter," and the steps would only include: expressing gratitude to the designated customer service representative. This implementation also uses business work orders for unified representation, and the information association matrix can be represented as: v = (i, 1, k).

[0092] Understandably, the completion of a certain step in a business order will affect the request in a customer service order. This impact is reflected in the value of the information association matrix, which has three possible values: resolved, unresolved, and irrelevant. That is, the completion of a certain step in a business order resolves the request in a customer service order, does not resolve the request in a customer service order (but the business step is related to resolving the request), or is irrelevant to resolving the request in a customer service order (the business step is not related to resolving the request).

[0093] Once the relationship between customer service work order k1 and business work order is established, the impact value of business work order on customer service work order k1 can be calculated to further determine whether the two duplicate customer service work orders are duplicate customer service work orders.

[0094] The impact of the business work order on the customer service work order k1 is calculated as follows:

[0095]

[0096]

[0097] Where G represents the set of unions of the impact values ​​of business work orders on customer service work orders; ∪ represents the union; i represents the business work orders of the first duplicate customer service work order (k1) and the second duplicate customer service work order (k2) within the interval; (q-p+1) is the number of business work orders i, where q≧p; j represents the business steps of the first duplicate customer service work order and the second duplicate customer service work order within the interval; (m-l+1) is the number of business steps j, where m≧l; k1 represents the first duplicate customer service work order; f((i,j,k1)) represents the impact value of the j-th step of the i-th business work order on customer service work order k1.

[0098] The values ​​of f((i,j,k1)) include three types: resolved, unresolved, and irrelevant.

[0099] Here, "resolved" means that the request in customer service work order k1 was resolved after the completion of the j-th business step of the i-th business work order; "unresolved" means that the request in customer service work order k1 was not resolved after the completion of the j-th business step of the i-th business work order; and "irrelevant" means that the completion of the j-th business step of the i-th business work order was irrelevant to resolving the request in customer service work order k1.

[0100] For easier understanding, please refer to the following example:

[0101] Taking customer service work orders k1 and k2 as examples, the time interval between k1 and k2 is t. Within the time interval t, two business work orders related to customer A are generated. Business work order 1 has 1 step, and business work order 2 has 3 steps. The impact of the business work orders on customer service work order k1 includes the following four situations:

[0102] ① After completing the first step of business work order 1, the processing result of customer service work order k1 (the request in it) is one of {resolved, unresolved, irrelevant};

[0103] ② After completing the first step of business work order 2, the processing result of customer service work order k1 (the request in it) is one of {resolved, unresolved, irrelevant};

[0104] ③ After completing the second step of business work order 2, the processing result of customer service work order k1 (the request in it) is one of {resolved, unresolved, irrelevant};

[0105] ④ After completing the third step in business work order 2, the processing result of customer service work order k1 (the request) is one of {resolved, unresolved, irrelevant}.

[0106] If any of the above four situations has an impact value of {resolved}, it means that the request in customer service ticket k1 has been resolved within the above time period; however, if none of the above four situations has an impact value of {resolved}, it means that the request in customer service ticket k1 has not been resolved within the above time period.

[0107] Furthermore, this implementation method determines whether two duplicate customer service tickets are duplicate customer service tickets in the following way:

[0108] {Completed}∩G={Completed} indicates that these two duplicate customer service tickets are similar customer service tickets.

[0109] {Completed}∩ This indicates that these two duplicate customer service tickets are duplicate customer service tickets.

[0110] Understandably, if any of the four scenarios above has an impact value of {resolved}, then set G must contain an element named "resolved". This means that the request in customer service ticket k1 has been resolved. Based on the aforementioned analysis, if the request in the earlier customer service ticket among two duplicate request customer service tickets has been resolved, then the two duplicate request customer service tickets are similar customer service tickets. However, if none of the four scenarios above has an impact value of {resolved}, then set G will not contain an element named "resolved", meaning that the request in customer service ticket k1 has not been resolved, indicating that the two duplicate request customer service tickets are duplicate customer service tickets.

[0111] By traversing all business processes within the time interval between the creation of two duplicate customer service work orders, analyzing the impact of business processes on the requests in the customer service work orders, and determining whether two duplicate customer service work orders are duplicate customer service work orders, this solves the technical problem of existing methods that only determine duplicate customer service work orders by analyzing the text content of customer service work orders, and cannot distinguish between duplicate customer service work orders and similar customer service work orders.

[0112] It's important to note that a customer service ticket may contain only one request, but a single request does not necessarily mean that a single business process can resolve the issue. For example, if the request in a customer service ticket is: a residential building's wiring needs repair, the corresponding business process might include: retrieving repair materials, assigning repair personnel, scheduling a repair time with the resident, the repair personnel arriving at the resident's home, and the repair being completed.

[0113] In two duplicate customer service work orders, by analyzing the impact of business work orders on duplicate customer service work orders, and tracing back to the business work orders when f((i,j,k))={unresolved}, we can find out which business work order i's business step was not completed in this customer service work order (k2). By tracing back the reasons for the similarity of work orders in similar customer service work orders and displaying them, we can deepen the depth of customer service.

[0114] The present invention provides a method for identifying duplicate customer service work orders. This method acquires customer service work orders from a specified user within a preset time range, then uses text recognition to determine the similarity of user requests within these work orders. Work orders with a similarity threshold are identified as duplicate requests. Furthermore, by acquiring business work orders established within the time interval between two duplicate request work orders, an information association matrix is ​​constructed between the earlier duplicate request work order and the business work order. This association matrix is ​​then used to determine whether the two duplicate request work orders are indeed duplicates. By combining the association matrix to identify the relationship between customer service work orders and business work orders, and focusing solely on the text of the work order itself, the method improves the accuracy of duplicate customer service order identification.

[0115] For easier understanding, please refer to Figure 2 This application provides a system for identifying duplicate electricity customer service work orders, including:

[0116] Module 1 is used to retrieve customer service work orders for a specified user within a preset time range; each customer service work order contains a request.

[0117] The first identification module 2 is used to calculate the similarity between any two requests through text recognition, and to designate the customer service work orders corresponding to the requests whose similarity meets a preset threshold as duplicate request customer service work orders.

[0118] The first judgment module 3 is used to determine whether there is a business work order within the time interval between the establishment of two duplicate customer service work orders. If not, the two duplicate customer service work orders are determined to be duplicate customer service work orders. If there is, the duplicate customer service work order with the earlier establishment time is taken as the first duplicate customer service work order, and an information association matrix is ​​constructed between the first duplicate customer service work order and the business work order.

[0119] The second identification module 4 is used to determine whether the two duplicate customer service work orders are duplicate customer service work orders based on the information association matrix.

[0120] In one specific embodiment, the first identification module 2 includes:

[0121] The first module is used to build a vector space model and use feature vectors to represent the requests in customer service work orders.

[0122] The first calculation module is used to calculate the cosine similarity of the angle between two feature vectors in the vector space. When the cosine similarity of the angle meets a preset threshold, it is determined that the requests corresponding to the two feature vectors are duplicated, and the customer service work orders corresponding to the two feature vectors are regarded as duplicate request customer service work orders.

[0123] In one specific embodiment, the first determination module 3 includes:

[0124] Extraction module: Used to extract all business processes in the business work order, and construct an information association matrix based on the first duplicate request customer service work order and all business processes in the business work order.

[0125] v = (i, j, k1)

[0126] Where i represents a business work order, j represents a business process, k1 represents the first repeat customer service work order, and v represents the relationship between customer service work order k1 and the j-th process of business work order i.

[0127] In one specific embodiment, the second identification module 4 includes:

[0128] The impact value calculation module is used to calculate the impact value of the business work order on the first duplicate request customer service work order based on the correlation. Specifically:

[0129] The impact value of the business work order on the first repeat customer service work order is calculated based on a pre-established calculation formula:

[0130]

[0131]

[0132] Where G represents the set of unions of the impact values ​​of business work orders on customer service work orders; ∪ represents the union of the two duplicate customer service work orders; i represents the business work order within the time interval between the establishment of the two duplicate customer service work orders; (qp) is the number of business work orders i; j represents the business process of business work order i; (ml) is the number of business processes j; k1 represents the first duplicate customer service work order; f((i,j,k1)) represents the impact value of the j-th process of the i-th business work order on the first duplicate customer service work order k1.

[0133] The second judgment module is used to determine whether the two duplicate customer service requests are duplicate customer service requests based on the impact values. Specifically: it determines whether the impact value set G contains {resolve}. If so, it determines that the two duplicate customer service requests are not duplicate customer service requests; otherwise, it determines that the two duplicate customer service requests are duplicate customer service requests.

[0134] Based on the foregoing embodiments, this application provides another preferred embodiment, wherein the identification system further includes a preprocessing module, which operates before the first identification module and is used to preprocess the acquired customer service work orders, including: removing customer service work orders corresponding to abnormal call numbers from among the duplicate call numbers, the abnormal call numbers including unexcused hang-ups, internal testing, and abnormal calls, etc.

[0135] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0136] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between apparatuses or units through some interfaces, and may be electrical, mechanical, or other forms.

[0137] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0138] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0139] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0140] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying duplicate electricity customer service work orders, characterized in that, Comprise: S1: obtain the customer service order of the specified user in the preset time range, and extract the corresponding appeal of each customer service order; S2: calculate the similarity of any two appeals by text recognition, and take the customer service order corresponding to the appeal whose similarity meets the preset threshold as the repeated appeal customer service order; S3: determine whether there is a business order in the time interval of the two repeated appeal customer service orders, if not, determine that the two repeated appeal customer service orders are repeated customer service orders; if yes, execute step S4; S4: take the repeated appeal customer service order with earlier establishment time as the first repeated appeal customer service order, and construct an information association matrix with the first repeated appeal customer service order and the business order; S5: determine whether the two repeated appeal customer service orders are repeated customer service orders based on the information association matrix; Step S4 comprises: extract all business links in the business order, and construct an information association matrix based on all business links in the first repeated appeal customer service order and the business order: v= (i, j, k1) Wherein, i represents a business order, j represents a business link, k1 represents a first repeated appeal customer service order, and v represents the association relationship between the first repeated appeal customer service order k1 and the jth link of the business order i; Step S5 comprises: calculate the influence value of the business order on the first repeated appeal customer service order based on the association relationship; determine whether the two repeated appeal customer service orders are repeated customer service orders based on the influence value; The calculation of the influence value of the business order on the first repeated appeal customer service order based on the association relationship comprises: calculate the influence value of the business order on the first repeated appeal customer service order based on the pre-established calculation formula: wherein G represents a set of union of influence values of business work orders on customer service work orders; represents a joint union; i represents a business work order within a time interval of two repeated complaint customer service work orders; (q-p+1) is the number of business work orders i, wherein q≥p; j represents a business link of the business work order i, (m-l+1) is the number of business links j, wherein m≥l; k1 represents a first repeated complaint customer service work order; represents an influence value of the jth link of the ith business work order on the first repeated complaint customer service work order k1; The determination of whether the two repeated appeal customer service orders are repeated customer service orders based on the influence value comprises: determine whether {solution} is contained in the influence value set G, if yes, determine that the two repeated appeal customer service orders are not repeated customer service orders; if not, determine that the two repeated appeal customer service orders are repeated customer service orders; After step S5, it further comprises: When When {unresolved}, determine the business order i and the corresponding business stage j that resolve the complaint in the first repeated complaint customer service order k1, and visually display the business order and the business stage.

2. The identification method according to claim 1, characterized in that, The calculation of the similarity of any two appeals by text recognition, taking the customer service order corresponding to the appeal whose similarity meets the preset threshold as the repeated appeal customer service order, comprises: establish a vector space model, and use a feature vector to represent the appeal in the customer service order; calculate the cosine similarity of two feature vectors in the vector space, when the cosine similarity meets the preset threshold, it is determined that the appeals corresponding to the two feature vectors are repeated, and the customer service orders corresponding to the two feature vectors are taken as repeated appeal customer service orders.

3. The identification method according to claim 1, characterized in that, Before step S2, it further comprises: preprocessing the customer service order, including eliminating the customer service order corresponding to the abnormal incoming call number in the repeated incoming call number, wherein the abnormal incoming call number includes no reason for hanging up, internal dialing test and abnormal incoming call.

4. The identification method according to claim 1, characterized in that, Before step S3, it further comprises: When the number of the repeated complaint customer service work orders exceeds 2, it is judged whether there is a business work order in the interval of the establishment time of the two repeated complaint customer service work orders with the earliest establishment time and the latest establishment time, if yes, the two repeated complaint customer service work orders with the closest establishment time are executed S3; if no, all the repeated complaint customer service work orders are determined as repeated customer service work orders.

5. A system for identifying recurring power service tickets, the system comprising: Comprise: The acquisition module is used for acquiring customer service work orders of a specified user in a preset time range; wherein, one customer service work order contains one complaint; The first identification module is used for calculating the similarity of any two complaints through text identification, and taking the customer service work orders corresponding to the complaint pairs with the similarity meeting a preset threshold as repeated complaint customer service work orders; The judgment module is used for judging whether there is a business work order in the interval of the establishment time of the two repeated complaint customer service work orders, if no, determining the two repeated complaint customer service work orders as repeated customer service work orders; if yes, taking the repeated complaint customer service work order with the earlier establishment time as a first repeated complaint customer service work order, and constructing an information correlation matrix with the first repeated complaint customer service work order and the business work order; The second identification module is used for judging whether the two repeated complaint customer service work orders are repeated customer service work orders based on the information correlation matrix; The judgment module is specifically used for extracting all business links in the business work order, and constructing an information correlation matrix based on all business links in the first repeated complaint customer service work order and the business work order: v= (i, j, k1) Wherein, i represents a business work order, j represents a business link, k1 represents a first repeated complaint customer service work order, and v represents the correlation relationship between the first repeated complaint customer service work order k1 and the jth link of the business work order i; The second identification module is specifically used for calculating the influence value of the business work order on the first repeated complaint customer service work order based on the correlation relationship, and judging whether the two repeated complaint customer service work orders are repeated customer service work orders based on the influence value; The calculation of the influence value of the business work order on the first repeated complaint customer service work order based on the correlation relationship specifically comprises: The influence value of the business work order on the first repeated complaint customer service work order is calculated based on a pre-established calculation formula: wherein G represents a set of union of influence values of business work orders on customer service work orders; represents a joint union; i represents a business work order within a time interval of two repeated complaint customer service work orders; (q-p+1) is the number of business work orders i, wherein q≥p; j represents a business link of the business work order i, (m-l+1) is the number of business links j, wherein m≥l; k1 represents a first repeated complaint customer service work order; represents an influence value of the jth link of the ith business work order on the first repeated complaint customer service work order k1; The judgment of whether the two repeated complaint customer service work orders are repeated customer service work orders based on the influence value specifically comprises: It is judged whether {solution} is contained in the influence value set G, if yes, it is determined that the two repeated complaint customer service work orders are not repeated customer service work orders; if no, it is determined that the two repeated complaint customer service work orders are repeated customer service work orders; After judging whether the two repeated complaint customer service work orders are repeated customer service work orders based on the information correlation matrix, it further comprises: When When {unresolved}, determine the business order i and the corresponding business stage j for resolving the appeal in the first repeated appeal customer service order k1, and visually display the business order and the business stage.

Images