Quality of service mutual evaluation method and device, electronic equipment and storage medium

Abstract

The present application relates to the field of artificial intelligence, and provides a service quality mutual evaluation method and device, electronic equipment and storage medium, comprising: initiating a service quality mutual evaluation task to at least two service personnel; obtaining an encrypted binary vector transmitted by each service personnel after executing the service quality mutual evaluation task; decrypting the encrypted binary vector to obtain a decrypted binary vector; determining a directed graph based on each decrypted binary vector; determining a target sorting mode of nodes in the directed graph, and determining a service quality mutual evaluation ranking result of the service personnel based on the target sorting mode and the directed graph. The service quality mutual evaluation method provided by the present application can realize the service quality evaluation mode by using the mutual evaluation task of two service personnel as a group while protecting privacy and security, optimize the service quality evaluation system of the service personnel, and also determine the mutual evaluation ranking result through the node sorting mode, effectively expand the use range of the ranking result, and improve the accuracy and efficiency of the service quality mutual evaluation.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of artificial intelligence, and in particular to a business quality mutual evaluation method and device, electronic equipment and storage medium. BACKGROUND

[0002] The business quality evaluation of a communication service refers to a process of evaluating and monitoring the quality of a communication service provided by a communication service provider. Existing business quality application methods mainly include the following: 1. Model-based business quality evaluation method: according to business requirements and application scenarios, a quality evaluation model is established using information such as user usage volume in a subsequent period of time, and there are white box mechanisms (evaluation methods of the evaluation model are explained to business personnel in advance) and black box mechanisms (evaluation methods of the evaluation model are not explained to business personnel in advance); 2. Consensus-based business quality evaluation method: front-line personnel can form an effective business quality evaluation method in long-term production practice, and the development quality of the business can be well determined by the front-line personnel, which can also promote mutual learning, decentralize evaluation power, and significantly improve the satisfaction of the front-line personnel.

[0003] However, for the model-based business quality evaluation method, if the evaluation mode is disclosed, the evaluation fairness will be affected; if the evaluation mode is not disclosed, the front-line marketing personnel will not have a good perception; for the consensus-based business quality evaluation method, the marketing evaluation personnel are prone to collude with each other, and the evaluation dimensions of the marketing evaluation personnel are inconsistent, resulting in inaccurate business quality evaluation, and the evaluation process of the marketing evaluation personnel is relatively lengthy, which will lead to low evaluation efficiency if there are a large number of evaluation tasks. SUMMARY

[0004] The embodiments of the present application provide a business quality mutual evaluation method and device, electronic equipment and storage medium to solve the problems of low accuracy and efficiency of business quality mutual evaluation.

[0005] In a first aspect, the embodiments of the present application provide a business quality mutual evaluation method, comprising:

[0006] Initiating a business quality mutual evaluation task to at least two sales personnel;

[0007] Obtaining an encrypted binary vector transmitted by each of the sales personnel after performing the business quality mutual evaluation task; the binary vector is used to represent the business quality advantage-disadvantage relationship between two sales personnel;

[0008] Decrypting the encrypted binary vector to obtain a decrypted binary vector;

[0009] Determining a directed graph based on each of the decrypted binary vectors;

[0010] Determine the target sorting method for the nodes in the directed graph, and based on the target sorting method and the directed graph, determine the peer evaluation ranking result of the salesperson's business quality.

[0011] In one embodiment, determining the target sorting method for nodes in the directed graph, and determining the peer review ranking result of the salesperson's business quality based on the target sorting method and the directed graph, includes:

[0012] If the directed graph is a directed acyclic graph, then delete the nodes with an out-degree of zero in the directed graph and determine the nodes with an out-degree of zero as the first and second nodes;

[0013] Iteratively execute the step of deleting nodes with an out-degree of zero in the directed graph, and determine the sorting information of all nodes in the directed graph except for the first and second nodes;

[0014] Based on the sorting information of each node in the directed graph, the peer evaluation ranking results of the salesperson's business quality are determined.

[0015] In one embodiment, determining the target sorting method for nodes in the directed graph, and determining the peer review ranking result of the salesperson's business quality based on the target sorting method and the directed graph, further includes:

[0016] If the directed graph is a directed cyclic graph, then the connectivity of each node in the directed graph is calculated by traversing it; the connectivity of a node refers to the number of times the same node is passed through all paths in the directed graph.

[0017] Based on the connectivity of each node in the directed graph, determine the node in the directed graph with the largest decrease in connectivity;

[0018] Based on the node with the largest decrease in connectivity, the directed graph is divided into at least two subgraphs; each subgraph includes a superior subgraph and a inferior subgraph; the superior subgraph is the subgraph pointed to by the node with the largest decrease in connectivity; the inferior subgraph is the subgraph that points to the node with the largest decrease in connectivity.

[0019] Based on each of the subgraphs, the peer review ranking results of the salesperson's business quality are determined.

[0020] In one embodiment, when determining the peer review ranking of the salesperson's business quality based on each of the subgraphs, the following steps are performed for each of the subgraphs:

[0021] Iterate through and calculate the connectivity of each node in the current subgraph;

[0022] Based on the connectivity of each node in the current subgraph, determine the node in the current subgraph with the largest decrease in connectivity;

[0023] Based on the node with the largest decrease in connectivity in the current subgraph, a target subgraph is determined; the target subgraph is a directed acyclic graph.

[0024] Based on the target subgraph of the current subgraph and the target subgraphs of other subgraphs, the salesperson's business quality peer review ranking result is determined.

[0025] In one embodiment, determining the directed graph based on each of the decrypted binary vectors includes:

[0026] Obtain the business completion information of each salesperson in the business quality mutual evaluation task;

[0027] The business completion information is input into a binary discriminant model to obtain a binary discriminant vector output by the binary discriminant model; the binary discriminant model is used to determine the business quality relationship between two salespersons.

[0028] The loss function is calculated based on the binary discriminant vector and the decryption binary vector to obtain the loss value of the binary discriminant model;

[0029] If the loss value is less than a preset threshold, a random number is generated;

[0030] Based on the random number, a directed graph is determined.

[0031] In one embodiment, determining the directed graph based on the random number includes:

[0032] If the random number is greater than or equal to a preset random value, then a directed graph is determined based on the binary discriminant vector.

[0033] In one embodiment, determining the directed graph based on the random number further includes:

[0034] If the random number is less than the preset random value, then the business quality mutual evaluation task will be re-initiated to at least two salespersons;

[0035] Obtain the new encrypted binary vector transmitted by each of the aforementioned salespersons after they have executed the re-initiated business quality mutual evaluation task;

[0036] The new encrypted binary vector is decrypted to obtain a new decrypted binary vector;

[0037] Based on the new decryption binary vector, a directed graph is determined.

[0038] Secondly, embodiments of the present invention provide a business quality mutual evaluation device, comprising:

[0039] The task initiation module is used to initiate business quality mutual evaluation tasks to at least two salespersons;

[0040] The acquisition module is used to acquire the encrypted binary vector transmitted by each of the salespersons after performing the business quality mutual evaluation task; the binary vector is used to represent the business quality superiority or inferiority relationship between two salespersons.

[0041] The decryption module is used to decrypt the encrypted binary vector to obtain a decrypted binary vector.

[0042] The first determining module is used to determine the directed graph based on each of the decrypted binary vectors;

[0043] The second determining module is used to determine the target sorting method of the nodes in the directed graph, and based on the target sorting method and the directed graph, determine the peer evaluation ranking result of the salesperson's business quality.

[0044] Thirdly, embodiments of the present invention provide an electronic device, including a processor and a memory storing a computer program, wherein the processor executes the program to implement the service quality mutual evaluation method described in the first aspect.

[0045] Fourthly, embodiments of the present invention provide a storage medium, which is an artificial intelligence readable storage medium, including an artificial intelligence program, which, when executed by a processor, implements the business quality mutual evaluation method described in the first aspect.

[0046] The business quality mutual evaluation method, apparatus, electronic device, and storage medium provided in this invention allow each salesperson to transmit encrypted binary vectors to a server after performing a business quality mutual evaluation task, thereby ranking the business quality of each salesperson. This approach can optimize the business quality evaluation system by using mutual evaluation tasks in pairs of salespersons while protecting privacy and security. Furthermore, by using a directed graph node sorting method to determine the business quality mutual evaluation ranking results, the application scope of the ranking results is effectively expanded, the accuracy of the business quality mutual evaluation is improved, and the efficiency of the business quality mutual evaluation is increased. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in this invention or 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 some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0048] Figure 1 This is a flowchart illustrating the business quality mutual evaluation method provided in an embodiment of the present invention;

[0049] Figure 2 This is a flowchart illustrating the overall scheme of the business quality mutual evaluation method provided in this embodiment of the invention;

[0050] Figure 3 This is a schematic diagram of the functional modules of an embodiment of the business quality mutual evaluation device of the present invention;

[0051] Figure 4 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. In the description of this specification, the reference to the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., means that the specific feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any suitable manner in one or more embodiments or examples. Furthermore, without contradiction, those skilled in the art can combine and combine the different embodiments or examples and the features of different embodiments or examples described in this specification. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] The following describes in detail the service quality mutual evaluation method, apparatus, electronic device and storage medium provided by the present invention with reference to embodiments.

[0054] Figure 1 This is a flowchart illustrating the business quality mutual evaluation method provided in an embodiment of the present invention. (Refer to...) Figure 1 This invention provides a method for mutual evaluation of business quality, which may include:

[0055] Step 100: Initiate a business quality mutual evaluation task for at least two sales representatives;

[0056] Step 200: Obtain the encrypted binary vector transmitted by each of the aforementioned salespersons after performing the business quality mutual evaluation task;

[0057] Step 300: Decrypt the encrypted binary vector to obtain the decrypted binary vector.

[0058] It should be noted that the service quality mutual evaluation method provided in this embodiment of the invention is implemented based on a service quality mutual evaluation device, and the service quality mutual evaluation method is applied to a server. Therefore, this embodiment of the invention uses a server as the execution subject as an example to describe the service quality mutual evaluation method.

[0059] It needs further explanation that the performance evaluation of sales personnel is an unquantifiable evaluation item. Peer review, or peer assessment, is a common solution, offering a fast, low-cost evaluation method that quickly identifies concealed defects, promoting overall quality improvement in the next cycle and helping the sales team enhance their understanding of the task. However, during peer review, it is difficult for sales personnel to rank all performance quality, while their ability to compare specific performance qualities is stronger. In other words, it is very difficult for a salesperson to evaluate the performance quality of a group of sales personnel, but it is relatively easy to compare the performance quality of two sales personnel. Therefore, this embodiment of the invention uses a server-based privacy-secure binary consensus transmission protocol to transmit binary discriminative consensus in performance quality peer review, that is, to judge the performance relationship between two sales personnel, for example, salesperson A is better than salesperson B (salesperson B is worse than salesperson A) or salesperson B is better than salesperson A (salesperson A is worse than salesperson B). It is assumed that the sales personnel are honest and curious, that is, they want to obtain the evaluation content made by other sales personnel, but will not forge requests, keys, or other information. Furthermore, after recording the salespersons' evaluations as private information, it is also possible to obtain the peer review ranking results of the salespersons' performance.

[0060] Specifically, when peer review of sales performance is required, the server generates at least two random number generators. Further, based on an asymmetric encryption algorithm, the server distributes these random number generators to at least two user terminals, i.e., to at least two sales personnel. It should be noted that each salesperson receives a random number generator through their user terminal. The random number generator generates two independent pseudo-random numbers based on the salesperson's execution of the server-initiated peer review task. These two pseudo-random numbers are then used as two components of a vector to generate a binary vector returned to the server.

[0061] It should be further noted that the asymmetric encryption algorithm uses different public keys and private keys. The public key is the key used to encrypt data, and the private key is the key used to decrypt data. In the asymmetric encryption algorithm, after the data is encrypted with the public key, it can only be decrypted with the corresponding private key. Therefore, the asymmetric encryption algorithm is widely used in scenarios such as digital signatures, data transmission, and key exchange. In the embodiment of the present invention, it is assumed that each user terminal of each salesman in the binary consensus transmission protocol has a public-private key pair provided by a trusted electronic certification service provider for signing evaluation information to the server, and there is a shared public-private key pair for the server to transmit information to each user.

[0062] Further, after the user terminal receives the random number generator, the server initiates a business quality mutual evaluation task to at least two user terminals, that is, initiates a business quality mutual evaluation task to at least two salesmen. Further, after each salesman receives the business quality mutual evaluation task initiated by the server through the user terminal, each salesman executes the business quality mutual evaluation task through the user terminal, that is, evaluates the business quality between every two salesmen. It should be noted that the business quality mutual evaluation task initiated by the server includes the business completion information of each salesman and the business quality mutual evaluation subtasks with every two salesmen as a group. The salesman who conducts the evaluation can evaluate the superiority and inferiority relationship between the two salesmen based on the business completion information.

[0063] Further, after each salesman executes the business quality mutual evaluation task through the user terminal, the random number generator generates two pseudo-random numbers, determines the two pseudo-random numbers as the components of the binary vector, and generates a binary vector. Among them, there is a size relationship between the two components in the binary vector. Therefore, the binary vector can be used to represent the superiority and inferiority relationship of the business quality between the two salesmen.

[0064] In one embodiment, salesman C determines through the user terminal that salesman A is superior to salesman B. The random number generator generates two pseudo-random numbers x and y, and x > y. Further, the random number generator generates a corresponding binary vector <x, y>, where the binary vector includes x > y, which can represent that salesman A is superior to salesman B or salesman B is inferior to salesman A.

[0065] In another embodiment, salesman C determines through the user terminal that salesman A is inferior to salesman B. The random number generator generates two pseudo-random numbers x and y, and x < y. Further, the random number generator generates a corresponding binary vector <y, x>, where the binary vector includes y > x, which can represent that salesman A is inferior to salesman B or salesman B is superior to salesman A.

[0066] Furthermore, the user terminal encrypts each generated binary vector based on an asymmetric encryption algorithm to obtain encrypted binary vectors. Then, the random number generator transmits each encrypted binary vector to the server.

[0067] Furthermore, the server obtains the encrypted binary vectors transmitted by each salesperson after performing the business quality mutual evaluation task. Furthermore, the server decrypts each encrypted binary vector based on an asymmetric encryption algorithm to obtain each decrypted binary vector.

[0068] Step 400: Determine the directed graph based on each of the decrypted binary vectors;

[0069] Step 500: Determine the target sorting method for the nodes in the directed graph, and based on the target sorting method and the directed graph, determine the peer evaluation ranking result of the salesperson's business quality.

[0070] Furthermore, the server determines the business quality ranking between each pair of salespersons based on the decrypted binary vectors, and constructs a directed graph based on these rankings. It should be noted that a directed graph is a network structure composed of nodes and directed edges. Directed edges represent the relationships between nodes. In this embodiment, each node represents a salesperson, and each directed edge represents the business quality ranking between two salespersons. Furthermore, each directed edge points from the inferior salesperson to the superior salesperson. In one embodiment, salesperson A is superior to salesperson B. Therefore, in the directed graph, there is a directed edge between the node representing salesperson B and the node representing salesperson A, pointing from the node representing salesperson B to the node representing salesperson A. Thus, the business quality ranking between each pair of salespersons can determine each node and the pointing relationships between nodes, thereby constructing a directed graph.

[0071] It's important to note that directed graphs include directed acyclic graphs (DAGs) and directed cyclic graphs (DCRs). A DAG contains no loops or cycles, meaning there are no paths originating from a node, traversing several edges, and returning to that node. DAGs can be used to represent structures with a specific order. A DCR, on the other hand, contains at least one loop or cycle, meaning there are paths originating from a node, traversing several edges, and returning to that node. DCRs can represent structures with cyclicity. Ideally, the constructed directed graph is a DAG, and the order of nodes can be determined based on its structural characteristics, thus determining the peer review ranking of sales personnel's business quality. However, in reality, the generated directed graph may contain cycles, i.e., DCRs. For DCRs, a complex network ranking algorithm can be proposed based on the connectivity of nodes, where node connectivity refers to the number of paths in the directed graph that pass through the same node.

[0072] Furthermore, the server determines the target sorting method for nodes in the directed graph, and based on the target sorting method and the directed graph, determines the peer review ranking results for the salesperson's business quality.

[0073] The business quality mutual evaluation method provided by this invention involves each salesperson transmitting an encrypted binary vector to a server after performing a business quality mutual evaluation task, thereby ranking the business quality of each salesperson. This method can achieve business quality evaluation by using mutual evaluation tasks in pairs of salespersons while protecting privacy and security, thus optimizing the business quality evaluation system for salespersons. At the same time, by using a directed graph node sorting method to determine the business quality mutual evaluation ranking results, the application scope of the ranking results is effectively expanded, the accuracy of business quality mutual evaluation is improved, and the efficiency of business quality mutual evaluation is increased.

[0074] Further, based on the decryption binary vectors described in step 400, determining the directed graph includes:

[0075] Obtain the business completion information of each salesperson in the business quality mutual evaluation task;

[0076] The business completion information is input into a binary discriminant model to obtain a binary discriminant vector output by the binary discriminant model; the binary discriminant model is used to determine the business quality relationship between two salespersons.

[0077] The loss function is calculated based on the binary discriminant vector and the decryption binary vector to obtain the loss value of the binary discriminant model;

[0078] If the loss value is less than a preset threshold, a random number is generated;

[0079] Based on the random number, a directed graph is determined.

[0080] It should be noted that while a binary consensus transmission protocol can be used to evaluate salespersons, the evaluation process is quite lengthy. For the evaluation of N salespersons, the salesperson conducting the evaluation needs to perform n(n-1) / 2 comparisons, and the server needs to send a comparison request to each of the N salespersons for each comparison. To reduce the burden of the evaluation process, a binary discriminant model that is basically consistent with the salesperson evaluation can be used.

[0081] Specifically, the server obtains the business completion information of each salesperson in the business quality mutual evaluation task. Further, the server inputs the business completion information into the binary discriminant model to obtain the binary discriminant vector output by the binary discriminant model. The binary discriminant model is used to determine the business quality superiority or inferiority relationship between two salespersons. Therefore, the binary discriminant vector output by the binary discriminant model can determine the business quality superiority or inferiority relationship between two salespersons.

[0082] It should be noted that the binary discriminant model is an online training model, which is a process of training and testing simultaneously. It can be combined with the binary discriminant vector predicted by the binary discriminant model and the binary vector obtained by the salesperson after performing the business quality mutual evaluation task to obtain the business quality mutual evaluation task results, thereby improving the accuracy of the binary discriminant model.

[0083] Furthermore, the server calculates the loss function based on the binary discriminant vector and the decrypted binary vector to obtain the loss value of the binary discriminant model. The formula for the loss function is as follows:

[0084] Where m is the number of decrypted binary vectors or binary discriminant vectors, and the number of decrypted binary vectors and binary discriminant vectors is the same, y pred y is the binary discriminant vector output by the model. true To decrypt a binary vector, loss is the loss value.

[0085] Furthermore, the server compares the loss value with a preset threshold, which is set according to the actual situation.

[0086] Furthermore, if the loss value is greater than or equal to a preset threshold, the server will re-initiate the business quality mutual evaluation task to at least two salespersons. It should be noted that if the loss value is greater than the preset threshold, it indicates a significant discrepancy between the model's predicted data and the actual data, meaning the model's prediction accuracy is low and its usability is limited. Therefore, the model's predicted data will not be used to determine the directed graph.

[0087] Furthermore, after each salesperson receives the re-initiated business quality mutual evaluation task from the server through their user terminal, they execute the re-initiated task, which means evaluating the business quality between each pair of salespersons.

[0088] Furthermore, after each salesperson executes the re-initiated business quality mutual evaluation task through the user terminal, the random number generator regenerates two new pseudo-random numbers, determines the two new pseudo-random numbers as the components of a new binary vector, and generates a new binary vector.

[0089] Furthermore, the user terminal encrypts each of the generated new binary vectors based on an asymmetric encryption algorithm to obtain new encrypted binary vectors. The user terminal then transmits each of the new encrypted binary vectors to the server.

[0090] Furthermore, the server obtains the new encrypted binary vectors transmitted by each salesperson after performing the business quality mutual evaluation task. Further, the server decrypts each new encrypted binary vector based on an asymmetric encryption algorithm to obtain each new decrypted binary vector.

[0091] Furthermore, based on the new decryption binary vectors, the server redetermines the business quality superiority / inferiority relationship between each pair of salespersons, and constructs a directed graph based on the redetermined business quality superiority / inferiority relationship between each pair of salespersons.

[0092] Furthermore, if the loss value is less than a preset threshold, the server generates a random number using a random number generator.

[0093] Furthermore, the server determines the directed graph based on random numbers.

[0094] This invention obtains the business completion information of each salesperson in a business quality mutual evaluation task, inputs the business completion information into a binary discriminant model, obtains the binary discriminant vector output by the binary discriminant model, calculates the loss function based on the binary discriminant vector and the decrypted binary vector, and obtains the loss value of the binary discriminant model. If the loss value is less than a preset threshold, a random number is generated, and then a directed graph is determined based on the random number, realizing the online training process of the binary discriminant model. Furthermore, the binary discriminant model can reduce the burden of the evaluation process and efficiently determine the business quality superiority or inferiority relationship between pairs of salespersons. Thus, the construction of a directed graph can be accelerated based on the business quality superiority or inferiority relationship between pairs of salespersons, thereby improving the efficiency of business quality mutual evaluation.

[0095] Further, determining the directed graph based on the random number includes:

[0096] If the random number is greater than or equal to a preset random value, then a directed graph is determined based on the binary discriminant vector.

[0097] Specifically, the server compares the random number with a preset random value, where the preset random value is set according to the actual situation.

[0098] Furthermore, if the random number is greater than or equal to the preset random value, the server determines the directed graph based on the binary discriminant vector.

[0099] This invention sets a random number and compares it with a preset random value. By comparing the random number, it selects whether to trust the judgment of the binary discriminant model. After determining that the random value is greater than or equal to a preset threshold, a directed graph is determined based on the binary discriminant vector. The binary discriminant vector output by the binary discriminant model can improve the accuracy and effectiveness of the directed graph, thereby helping to improve the credibility of mutual evaluation of business quality. Furthermore, by using a collaborative binary consensus transmission protocol, the efficiency of mutual evaluation of business quality is guaranteed while ensuring the accuracy of the binary discriminant model. The efficiency improvement brought by the binary discriminant model allows for the simultaneous evaluation of business personnel using multiple indicators, so that the business evaluation problem can ultimately be reduced to a node ranking problem on a complex network.

[0100] Furthermore, determining the directed graph based on the random number further includes:

[0101] If the random number is less than the preset random value, then the business quality mutual evaluation task will be re-initiated to at least two salespersons;

[0102] Obtain the new encrypted binary vector transmitted by each of the aforementioned salespersons after they have executed the re-initiated business quality mutual evaluation task;

[0103] The new encrypted binary vector is decrypted to obtain a new decrypted binary vector;

[0104] Based on the new decryption binary vector, a directed graph is determined.

[0105] Specifically, if the random number is less than the preset random value, the server will re-initiate the business quality mutual evaluation task to at least two salespersons.

[0106] Furthermore, after each salesperson receives the re-initiated business quality mutual evaluation task from the server through their user terminal, they execute the re-initiated task, which means evaluating the business quality between each pair of salespersons.

[0107] Furthermore, after each salesperson executes the re-initiated business quality mutual evaluation task through the user terminal, the random number generator regenerates two new pseudo-random numbers, determines the two new pseudo-random numbers as the components of a new binary vector, and generates a new binary vector.

[0108] Furthermore, the user terminal encrypts each of the generated new binary vectors based on an asymmetric encryption algorithm to obtain new encrypted binary vectors. The user terminal then transmits each of the new encrypted binary vectors to the server.

[0109] Furthermore, the server obtains the new encrypted binary vectors transmitted by each salesperson after performing the business quality mutual evaluation task. Further, the server decrypts each new encrypted binary vector based on an asymmetric encryption algorithm to obtain each new decrypted binary vector.

[0110] Furthermore, based on the new decryption binary vectors, the server redetermines the business quality superiority / inferiority relationship between each pair of salespersons, and constructs a directed graph based on the redetermined business quality superiority / inferiority relationship between each pair of salespersons.

[0111] This invention implements an online training process for the binary discriminant model by setting a random number and comparing it with a preset random value. This comparison allows for the selection of whether to trust the binary discriminant model's judgment. During training, the binary discriminant model maintains a certain update rate to balance accuracy and utilization. Furthermore, by using a collaborative binary consensus transmission protocol, the efficiency of mutual evaluation of business quality is ensured while maintaining the accuracy of the binary discriminant model. Utilizing the binary discriminant model improves efficiency, enabling the simultaneous evaluation of sales personnel using multiple indicators. Ultimately, this reduces the business evaluation problem to a node ranking problem on a complex network.

[0112] Further, based on the target sorting method of the nodes in the directed graph determined in step 500, and based on the target sorting method and the directed graph, the peer evaluation ranking result of the salesperson's business quality is determined, including:

[0113] If the directed graph is a directed acyclic graph, then delete the nodes with an out-degree of zero in the directed graph and determine the nodes with an out-degree of zero as the first and second nodes;

[0114] Iteratively execute the step of deleting nodes with an out-degree of zero in the directed graph, and determine the sorting information of all nodes in the directed graph except for the first and second nodes;

[0115] Based on the sorting information of each node in the directed graph, the peer evaluation ranking results of the salesperson's business quality are determined.

[0116] Specifically, the server judges the network structure of the directed graph. If the server determines that the directed graph is a directed acyclic graph, the server determines and deletes the node with an out-degree of zero in the directed graph, and currently determines the node with an out-degree of zero as the first node. In other words, the current node with an out-degree of zero is ranked first, which means that the business quality peer evaluation ranking of the business person corresponding to the current node with an out-degree of zero is ranked first.

[0117] It should be noted that if a directed graph does not contain self-loops and there is only one directed edge between two nodes, the directed graph is a directed acyclic graph. Therefore, when the number of nodes is less than or equal to 2, the directed graph is a directed acyclic graph.

[0118] Furthermore, the server continues to identify and delete nodes with an out-degree of zero in the directed graph, and determines the current node with an out-degree of zero as the second node. In other words, the ranking information of the current node with an out-degree of zero is second, which means that the business quality peer evaluation ranking of the salesperson corresponding to the current node with an out-degree of zero is second.

[0119] Furthermore, the server iteratively executes the steps described above to identify and delete nodes with an out-degree of zero in the directed graph, and determines the sorting information of all nodes in the directed graph except for the first and second nodes.

[0120] It should be noted that the server can sort the nodes in a directed graph (i.e., a directed cyclic graph) based on its structural characteristics. Since each node corresponds to a salesperson, the server can determine the peer review ranking of the salespersons' business quality based on the sorting information of each node in the directed graph.

[0121] Furthermore, the server determines the peer review ranking results of the salespersons' business quality based on the sorting information of each node in the directed graph.

[0122] This invention, after determining that the directed graph is a directed acyclic graph (DAG), identifies and deletes nodes with an out-degree of zero, and designates these nodes as the first node. The process iteratively continues, deleting nodes with zero out-degree and determining the ranking of all nodes except the first node. Based on this ranking, the ranking of salespersons' performance in peer review is determined. Leveraging the structural characteristics of DAGs, this invention solves the node ranking problem on complex networks, standardizes the ranking process, improves the fairness of performance evaluation, and offers efficient and concise calculations with low computational costs, saving debugging and runtime costs. Furthermore, determining the ranking through node ranking effectively expands the scope of application of the ranking results and improves the accuracy and efficiency of peer review.

[0123] Furthermore, the step of determining the target sorting method for nodes in the directed graph, and determining the peer review ranking result of the salesperson's business quality based on the target sorting method and the directed graph, further includes:

[0124] If the directed graph is a directed cyclic graph, then the connectivity of each node in the directed graph is calculated by traversing it; the connectivity of a node refers to the number of times the same node is passed through all paths in the directed graph.

[0125] Based on the connectivity of each node in the directed graph, determine the node in the directed graph with the largest decrease in connectivity;

[0126] Based on the node with the largest decrease in connectivity, the directed graph is divided into at least two subgraphs; each subgraph includes a superior subgraph and a inferior subgraph; the superior subgraph is the subgraph pointed to by the node with the largest decrease in connectivity; the inferior subgraph is the subgraph that points to the node with the largest decrease in connectivity.

[0127] Based on each of the subgraphs, the peer review ranking results of the salesperson's business quality are determined.

[0128] Specifically, the server determines the network structure of the directed graph. If the server determines that the directed graph is a directed cyclic graph, the server traverses and calculates the connectivity of each node in the directed graph. The connectivity of a node refers to the number of times the same node is passed through all paths in the directed graph.

[0129] Furthermore, the server compares the connectivity of each node in the directed graph to obtain the connectivity comparison result.

[0130] Furthermore, based on the comparison results of connectivity, the server determines the node with the smallest connectivity in the directed graph as the node with the largest decrease in connectivity in the directed graph.

[0131] Furthermore, the server divides the directed graph into two subgraphs based on the node with the largest decrease in connectivity. This can also be understood as the server dividing the directed graph into two subgraphs by deleting the node with the largest decrease in connectivity, and determining the ranking information between the nodes in the two subgraphs and the node with the largest decrease in connectivity in the directed graph. The subgraphs include superior subgraphs and inferior subgraphs. The superior subgraph is the subgraph pointed to by the node with the largest decrease in connectivity, and the inferior subgraph is the subgraph that points to the node with the largest decrease in connectivity. Therefore, the nodes in the superior subgraph are ranked higher than the node with the largest decrease in connectivity, and the nodes in the inferior subgraph are ranked lower than the node with the largest decrease in connectivity. The node with the largest decrease in connectivity is between the nodes in the two subgraphs.

[0132] Furthermore, the server iteratively performs the step of partitioning the directed graph into at least two subgraphs. Further, based on each subgraph, the server determines the peer evaluation ranking results of the salespersons' business quality.

[0133] This invention, based on the connectivity of each node in a directed graph, identifies the node with the largest decrease in connectivity. Based on this node, the directed graph is divided into at least two subgraphs. Then, based on each subgraph, the peer review ranking of sales personnel is determined. By utilizing the structural characteristics of directed cyclic graphs to determine the maximum connectivity decrease method, this invention solves the node ranking problem on complex networks. It standardizes the ranking process, improves the fairness of sales personnel performance evaluation, and offers efficient and concise calculations with low computational costs, saving debugging and running time. Furthermore, determining the peer review ranking results through node ranking effectively expands the scope of application of the ranking results and improves the accuracy and efficiency of peer review of business quality.

[0134] Furthermore, when determining the peer review ranking results of the salesperson's business quality based on each of the subgraphs, the following steps are performed for each subgraph:

[0135] Iterate through and calculate the connectivity of each node in the current subgraph;

[0136] Based on the connectivity of each node in the current subgraph, determine the node in the current subgraph with the largest decrease in connectivity;

[0137] Based on the node with the largest decrease in connectivity in the current subgraph, a target subgraph is determined; the target subgraph is a directed acyclic graph.

[0138] Based on the target subgraph of the current subgraph and the target subgraphs of other subgraphs, the salesperson's business quality peer review ranking result is determined.

[0139] Specifically, the server iterates through and calculates the connectivity of each node in the current subgraph. Furthermore, the server compares the connectivity of each node in the current subgraph to obtain the connectivity comparison result.

[0140] Furthermore, based on the comparison results of connectivity, the server determines the node with the smallest connectivity in the current subgraph as the node with the largest decrease in connectivity in the current subgraph.

[0141] Furthermore, the server determines the target subgraph based on the node with the largest decrease in connectivity in the current subgraph. The target subgraph is a directed acyclic graph (DAG). It's important to note that determining the target subgraph based on the node with the largest decrease in connectivity in the current subgraph involves the server iteratively searching for this node until the resulting subgraph is a DAG. If the subgraph is indeed a DAG, nodes with an out-degree of zero are deleted, and the nodes in the subgraph are ranked.

[0142] Furthermore, the server determines the ranking information of nodes in the current subgraph based on the target subgraph and the node with the largest decrease in connectivity in the current subgraph. It should be noted that the target subgraph is a directed acyclic graph. As explained above, the ranking information of a directed acyclic graph can be determined by iteratively deleting nodes with an out-degree of zero, which will not be repeated here.

[0143] Furthermore, the server determines the sorting information of nodes in other subgraphs based on the target subgraph and the nodes with the largest decrease in connectivity in other subgraphs.

[0144] Furthermore, based on the sorting information of nodes in the current subgraph, the sorting information of nodes in other subgraphs, and the sorting information between nodes in the two subgraphs of the aforementioned directed graph and the node in the directed graph with the largest decrease in connectivity, the server determines the sorting information of all nodes in the directed graph.

[0145] Furthermore, the server determines the peer review ranking of salespersons' business quality based on the sorting information of all nodes in the directed graph.

[0146] This invention, through traversing and calculating the connectivity of each node in the current subgraph, identifies the node with the largest decrease in connectivity based on the connectivity of each node in the current subgraph. It then determines the target subgraph based on this node. Furthermore, based on the target subgraph of the current subgraph and the target subgraphs of other subgraphs, it determines the peer evaluation ranking results for sales personnel. By leveraging the structural characteristics of directed acyclic graphs (DAGs) and directed cyclic graphs, it solves the node sorting problem on complex networks. This standardizes the sorting process, improves the fairness of sales personnel performance evaluation, and the calculation process is efficient, concise, and cost-effective, saving debugging and running time. Simultaneously, determining the peer evaluation ranking results through node sorting effectively expands the scope of application of the ranking results and improves the accuracy and efficiency of peer evaluation of business quality.

[0147] Further, refer to Figure 2 , Figure 2 This is a flowchart of the overall scheme of the business quality mutual evaluation method provided in the embodiments of the present invention. Therefore, the overall process of the business quality mutual evaluation method provided by the present invention can be understood as follows:

[0148] When it is necessary to conduct mutual evaluation of the sales quality of salespersons, the server generates at least two random number generators. Furthermore, based on an asymmetric encryption algorithm, the server distributes the random number generators to at least two user terminals, that is, to at least two salespersons.

[0149] Furthermore, after the user terminal receives the random number generator, the server initiates a business quality mutual evaluation task to at least two user terminals, which means initiating a business quality mutual evaluation task to at least two salespersons. Further, after each salesperson receives the business quality mutual evaluation task initiated by the server through their user terminal, each salesperson executes the business quality mutual evaluation task through their user terminal, which means evaluating the business quality between each pair of salespersons.

[0150] Furthermore, after each salesperson performs a business quality mutual evaluation task through the user terminal, the random number generator generates two pseudo-random numbers, which are then used as components of a binary vector to generate the binary vector.

[0151] Furthermore, the user terminal encrypts each generated binary vector based on an asymmetric encryption algorithm to obtain encrypted binary vectors. Then, the random number generator transmits each encrypted binary vector to the server.

[0152] Furthermore, the server obtains the encrypted binary vectors transmitted by each salesperson after performing the business quality mutual evaluation task. Furthermore, the server decrypts each encrypted binary vector based on an asymmetric encryption algorithm to obtain each decrypted binary vector.

[0153] Furthermore, the server obtains the business completion information of each salesperson in the business quality mutual evaluation task. Further, the server inputs the business completion information into the binary discriminant model to obtain the binary discriminant vector output by the binary discriminant model.

[0154] Furthermore, the server calculates the loss function based on the binary discriminant vector and the decrypted binary vector to obtain the loss value of the binary discriminant model.

[0155] Furthermore, the server compares the loss value with a preset threshold.

[0156] Furthermore, if the loss value is greater than or equal to a preset threshold, the server will re-initiate the business quality mutual evaluation task to at least two salespersons.

[0157] Furthermore, after each salesperson receives the re-initiated business quality mutual evaluation task from the server through their user terminal, they execute the re-initiated task, which means evaluating the business quality between each pair of salespersons.

[0158] Furthermore, after each salesperson executes the re-initiated business quality mutual evaluation task through the user terminal, the random number generator regenerates two new pseudo-random numbers, determines the two new pseudo-random numbers as the components of a new binary vector, and generates a new binary vector.

[0159] Furthermore, the user terminal encrypts each of the generated new binary vectors based on an asymmetric encryption algorithm to obtain new encrypted binary vectors. The user terminal then transmits each of the new encrypted binary vectors to the server.

[0160] Furthermore, the server obtains the new encrypted binary vectors transmitted by each salesperson after performing the business quality mutual evaluation task. Further, the server decrypts each new encrypted binary vector based on an asymmetric encryption algorithm to obtain each new decrypted binary vector.

[0161] Furthermore, based on the new decryption binary vectors, the server redetermines the business quality superiority / inferiority relationship between each pair of salespersons, and constructs a directed graph based on the redetermined business quality superiority / inferiority relationship between each pair of salespersons.

[0162] Furthermore, if the loss value is less than a preset threshold, the server generates a random number using a random number generator.

[0163] Furthermore, the server compares the random number with a preset random value.

[0164] Furthermore, if the random number is greater than or equal to the preset random value, the server determines the directed graph based on the binary discriminant vector.

[0165] Furthermore, if the random number is less than the preset random value, the server will re-initiate the business quality mutual evaluation task to at least two salespersons.

[0166] Furthermore, after each salesperson receives the re-initiated business quality mutual evaluation task from the server through their user terminal, they execute the re-initiated task, which means evaluating the business quality between each pair of salespersons.

[0167] Furthermore, after each salesperson executes the re-initiated business quality mutual evaluation task through the user terminal, the random number generator regenerates two new pseudo-random numbers, determines the two new pseudo-random numbers as the components of a new binary vector, and generates a new binary vector.

[0168] Furthermore, the user terminal encrypts each of the generated new binary vectors based on an asymmetric encryption algorithm to obtain new encrypted binary vectors. The user terminal then transmits each of the new encrypted binary vectors to the server.

[0169] Furthermore, the server obtains the new encrypted binary vectors transmitted by each salesperson after performing the business quality mutual evaluation task. Further, the server decrypts each new encrypted binary vector based on an asymmetric encryption algorithm to obtain each new decrypted binary vector.

[0170] Furthermore, based on the new decryption binary vectors, the server redetermines the business quality superiority / inferiority relationship between each pair of salespersons, and constructs a directed graph based on the redetermined business quality superiority / inferiority relationship between each pair of salespersons.

[0171] Furthermore, the server judges the network structure of the directed graph. If the server determines that the directed graph is a directed acyclic graph, the server determines and deletes the node with an out-degree of zero in the directed graph, and currently determines the node with an out-degree of zero as the first node. That is to say, the current node with an out-degree of zero is ranked first, which means that the business quality peer evaluation ranking of the business person corresponding to the current node with an out-degree of zero is ranked first.

[0172] Furthermore, the server iteratively executes the steps described above to identify and delete nodes with an out-degree of zero in the directed graph, and determines the sorting information of all nodes in the directed graph except for the first and second nodes.

[0173] Furthermore, the server determines the peer review ranking results of the salespersons' business quality based on the sorting information of each node in the directed graph.

[0174] Furthermore, if the server determines that the directed graph is a directed cyclic graph, the server will traverse and calculate the connectivity of each node in the directed graph.

[0175] Furthermore, the server compares the connectivity of each node in the directed graph to obtain the connectivity comparison result.

[0176] Furthermore, based on the comparison results of connectivity, the server determines the node with the smallest connectivity in the directed graph as the node with the largest decrease in connectivity in the directed graph.

[0177] Furthermore, the server divides the directed graph into two subgraphs based on the node with the largest decrease in connectivity. This can also be understood as the server dividing the directed graph into two subgraphs by deleting the node with the largest decrease in connectivity, and determining the ordering information between the nodes in the two subgraphs and the node with the largest decrease in connectivity in the directed graph.

[0178] Furthermore, the server iterates through and calculates the connectivity of each node in the current subgraph. Then, the server compares the connectivity of each node in the current subgraph to obtain the connectivity comparison result.

[0179] Furthermore, based on the comparison results of connectivity, the server determines the node with the smallest connectivity in the current subgraph as the node with the largest decrease in connectivity in the current subgraph.

[0180] Furthermore, the server determines the target subgraph of the current subgraph based on the node with the largest decrease in connectivity in the current subgraph, where the target subgraph is a directed acyclic graph.

[0181] Furthermore, the server determines the sorting information of nodes in the current subgraph based on the target subgraph of the current subgraph and the node in the current subgraph with the largest decrease in connectivity.

[0182] Furthermore, the server determines the sorting information of nodes in other subgraphs based on the target subgraph and the nodes with the largest decrease in connectivity in other subgraphs.

[0183] Furthermore, based on the sorting information of nodes in the current subgraph, the sorting information of nodes in other subgraphs, and the sorting information between nodes in the two subgraphs of the aforementioned directed graph and the node in the directed graph with the largest decrease in connectivity, the server determines the sorting information of all nodes in the directed graph.

[0184] Furthermore, the server determines the peer review ranking of salespersons' business quality based on the sorting information of all nodes in the directed graph.

[0185] Furthermore, the present invention also provides a business quality mutual evaluation device.

[0186] Reference Figure 3 , Figure 3 This is a schematic diagram of the functional modules of an embodiment of the business quality mutual evaluation device of the present invention.

[0187] The business quality mutual evaluation device includes:

[0188] Task initiation module 310 is used to initiate a business quality mutual evaluation task to at least two salespersons.

[0189] The acquisition module 320 is used to acquire the encrypted binary vector transmitted by each of the salespersons after performing the business quality mutual evaluation task; the binary vector is used to represent the business quality superiority or inferiority relationship between two salespersons.

[0190] Decryption module 330 is used to decrypt the encrypted binary vector to obtain a decrypted binary vector;

[0191] The first determining module 340 is used to determine a directed graph based on each of the decrypted binary vectors;

[0192] The second determining module 350 is used to determine the target sorting method of the nodes in the directed graph, and based on the target sorting method and the directed graph, determine the peer evaluation ranking result of the salesperson's business quality.

[0193] The business quality peer review device provided by this invention, after each salesperson performs a business quality peer review task, transmits an encrypted binary vector to a server to rank the business quality of each salesperson. It can realize the business quality evaluation method by using peer review tasks in pairs of salespersons while protecting privacy and security, thus optimizing the business quality evaluation system for salespersons. At the same time, by using a directed graph node sorting method to determine the business quality peer review ranking results of salespersons, it effectively expands the scope of application of the ranking results, improves the accuracy of business quality peer review, and improves the efficiency of business quality peer review.

[0194] In one embodiment, the first determining module 340 is further configured to:

[0195] Obtain the business completion information of each salesperson in the business quality mutual evaluation task;

[0196] The business completion information is input into a binary discriminant model to obtain a binary discriminant vector output by the binary discriminant model; the binary discriminant model is used to determine the business quality relationship between two salespersons.

[0197] The loss function is calculated based on the binary discriminant vector and the decryption binary vector to obtain the loss value of the binary discriminant model;

[0198] If the loss value is less than a preset threshold, a random number is generated;

[0199] Based on the random number, a directed graph is determined.

[0200] In one embodiment, the first determining module 340 is further configured to:

[0201] If the random number is greater than or equal to a preset random value, then a directed graph is determined based on the binary discriminant vector.

[0202] In one embodiment, the first determining module 340 is further configured to:

[0203] If the random number is less than the preset random value, then the business quality mutual evaluation task will be re-initiated to at least two salespersons;

[0204] Obtain the new encrypted binary vector transmitted by each of the aforementioned salespersons after they have executed the re-initiated business quality mutual evaluation task;

[0205] The new encrypted binary vector is decrypted to obtain a new decrypted binary vector;

[0206] Based on the new decryption binary vector, a directed graph is determined.

[0207] In one embodiment, the second determining module 350 is further configured to:

[0208] If the directed graph is a directed acyclic graph, then delete the nodes with an out-degree of zero in the directed graph and determine the nodes with an out-degree of zero as the first and second nodes;

[0209] Iteratively execute the step of deleting nodes with an out-degree of zero in the directed graph, and determine the sorting information of all nodes in the directed graph except for the first and second nodes;

[0210] Based on the sorting information of each node in the directed graph, the peer evaluation ranking results of the salesperson's business quality are determined.

[0211] In one embodiment, the second determining module 350 is further configured to:

[0212] If the directed graph is a directed cyclic graph, then the connectivity of each node in the directed graph is calculated by traversing it; the connectivity of a node refers to the number of times the same node is passed through all paths in the directed graph.

[0213] Based on the connectivity of each node in the directed graph, determine the node in the directed graph with the largest decrease in connectivity;

[0214] Based on the node with the largest decrease in connectivity, the directed graph is divided into at least two subgraphs; each subgraph includes a superior subgraph and a inferior subgraph; the superior subgraph is the subgraph pointed to by the node with the largest decrease in connectivity; the inferior subgraph is the subgraph that points to the node with the largest decrease in connectivity.

[0215] Based on each of the subgraphs, the peer review ranking results of the salesperson's business quality are determined.

[0216] In one embodiment, the second determining module 350 is further configured to:

[0217] Iterate through and calculate the connectivity of each node in the current subgraph;

[0218] Based on the connectivity of each node in the current subgraph, determine the node in the current subgraph with the largest decrease in connectivity;

[0219] Based on the node with the largest decrease in connectivity in the current subgraph, a target subgraph is determined; the target subgraph is a directed acyclic graph.

[0220] Based on the target subgraph of the current subgraph and the target subgraphs of other subgraphs, the salesperson's business quality peer review ranking result is determined.

[0221] Figure 4 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 4 As shown, the electronic device may include: a processor 410, a communication interface 420, a memory 430, and a communication bus 440, wherein the processor 410, the communication interface 420, and the memory 430 communicate with each other via the communication bus 440. The processor 410 can call an artificial intelligence program in the memory 430 to execute the steps of the business quality mutual evaluation method, such as including:

[0222] Initiate a peer review task on business quality for at least two sales representatives;

[0223] Obtain the encrypted binary vector transmitted by each of the aforementioned salespersons after performing the mutual evaluation task of business quality; the binary vector is used to characterize the relationship of business quality superiority or inferiority between two salespersons.

[0224] The encrypted binary vector is decrypted to obtain the decrypted binary vector;

[0225] Based on the aforementioned decryption binary vectors, a directed graph is determined;

[0226] Determine the target sorting method for the nodes in the directed graph, and based on the target sorting method and the directed graph, determine the peer evaluation ranking result of the salesperson's business quality.

[0227] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in an AI-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This AI software product is stored in a storage medium and includes several instructions to cause an AI device (which may be a personal AI, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this 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.

[0228] On the other hand, embodiments of the present invention also provide a medium, which is an artificial intelligence readable storage medium storing an artificial intelligence program. The artificial intelligence program is used to cause a processor to execute the steps of the methods provided in the above embodiments, including, for example:

[0229] Initiate a peer review task on business quality for at least two sales representatives;

[0230] Obtain the encrypted binary vector transmitted by each of the aforementioned salespersons after performing the mutual evaluation task of business quality; the binary vector is used to characterize the relationship of business quality superiority or inferiority between two salespersons.

[0231] The encrypted binary vector is decrypted to obtain the decrypted binary vector;

[0232] Based on the aforementioned decryption binary vectors, a directed graph is determined;

[0233] Determine the target sorting method for the nodes in the directed graph, and based on the target sorting method and the directed graph, determine the peer evaluation ranking result of the salesperson's business quality.

[0234] The AI-readable storage medium can be any available medium or data storage device that the processor can access, including but not limited to magnetic storage (e.g., floppy disks, hard disks, magnetic tapes, magneto-optical disks (MO), etc.), optical storage (e.g., CDs, DVDs, BDs, HVDs, etc.), and semiconductor storage (e.g., ROMs, EPROMs, EEPROMs, non-volatile memory (NAND flash), solid-state drives (SSDs)).

[0235] The device embodiments described above are merely illustrative. 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 modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0236] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This artificial intelligence software product can be stored in an artificial intelligence readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions to cause an artificial intelligence device (which may be a personal artificial intelligence, a server, or a network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments.

[0237] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; 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; and these 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 mutual evaluation of business quality, characterized in that, Applied to servers, the service quality mutual evaluation method includes: Initiate a peer review task on business quality for at least two sales representatives; Obtain the encrypted binary vector transmitted by each of the aforementioned salespersons after performing the mutual evaluation task of business quality; the binary vector is used to characterize the relationship of business quality superiority or inferiority between two salespersons. The encrypted binary vector is decrypted to obtain the decrypted binary vector; Based on the aforementioned decryption binary vectors, a directed graph is determined; Determine the target sorting method for the nodes in the directed graph, and based on the target sorting method and the directed graph, determine the peer evaluation ranking result of the salesperson's business quality. The determination of the directed graph based on each of the decrypted binary vectors includes: Obtain the business completion information of each salesperson in the business quality mutual evaluation task; The business completion information is input into a binary discriminant model to obtain a binary discriminant vector output by the binary discriminant model; the binary discriminant model is used to determine the business quality relationship between two salespersons. The loss function is calculated based on the binary discriminant vector and the decryption binary vector to obtain the loss value of the binary discriminant model; If the loss value is less than a preset threshold, a random number is generated; Based on the random number, a directed graph is determined.

2. The business quality mutual evaluation method according to claim 1, characterized in that, The process of determining the target sorting method for nodes in the directed graph, and determining the peer review ranking result of the salesperson's business quality based on the target sorting method and the directed graph, includes: If the directed graph is a directed acyclic graph, then delete the nodes with an out-degree of zero in the directed graph and determine the nodes with an out-degree of zero as the first and second nodes; Iteratively execute the step of deleting nodes with an out-degree of zero in the directed graph, and determine the sorting information of all nodes in the directed graph except for the first and second nodes; Based on the sorting information of each node in the directed graph, the peer evaluation ranking results of the salesperson's business quality are determined.

3. The business quality mutual evaluation method according to claim 2, characterized in that, The step of determining the target sorting method for nodes in the directed graph, and determining the peer review ranking result of the salesperson's business quality based on the target sorting method and the directed graph, further includes: If the directed graph is a directed cyclic graph, then the connectivity of each node in the directed graph is calculated by traversing it; the connectivity of a node refers to the number of times the same node is passed through all paths in the directed graph. Based on the connectivity of each node in the directed graph, determine the node in the directed graph with the largest decrease in connectivity; Based on the node with the largest decrease in connectivity, the directed graph is divided into at least two subgraphs; each subgraph includes a superior subgraph and a inferior subgraph; the superior subgraph is the subgraph pointed to by the node with the largest decrease in connectivity; the inferior subgraph is the subgraph that points to the node with the largest decrease in connectivity. Based on each of the subgraphs, the peer review ranking results of the salesperson's business quality are determined.

4. The business quality mutual evaluation method according to claim 3, characterized in that, When determining the peer review ranking of the salesperson's business quality based on each of the subgraphs, the following steps are performed for each subgraph: Iterate through and calculate the connectivity of each node in the current subgraph; Based on the connectivity of each node in the current subgraph, determine the node in the current subgraph with the largest decrease in connectivity; Based on the node with the largest decrease in connectivity in the current subgraph, determine the target subgraph of the current subgraph; The target subgraph is a directed acyclic graph; Based on the target subgraph of the current subgraph and the target subgraphs of other subgraphs, the salesperson's business quality peer review ranking result is determined.

5. The business quality mutual evaluation method according to claim 1, characterized in that, The process of determining a directed graph based on the random number includes: If the random number is greater than or equal to a preset random value, then a directed graph is determined based on the binary discriminant vector.

6. The business quality mutual evaluation method according to claim 5, characterized in that, The process of determining a directed graph based on the random number further includes: If the random number is less than the preset random value, then the business quality mutual evaluation task will be re-initiated to at least two salespersons; Obtain the new encrypted binary vector transmitted by each of the aforementioned salespersons after they have executed the re-initiated business quality mutual evaluation task; The new encrypted binary vector is decrypted to obtain a new decrypted binary vector; Based on the new decryption binary vector, a directed graph is determined.

7. A business quality mutual evaluation device, characterized in that, include: The task initiation module is used to initiate business quality mutual evaluation tasks to at least two salespersons; The acquisition module is used to acquire the encrypted binary vector transmitted by each of the salespersons after performing the business quality mutual evaluation task; the binary vector is used to represent the business quality superiority or inferiority relationship between two salespersons. The decryption module is used to decrypt the encrypted binary vector to obtain a decrypted binary vector. The first determining module is used to determine the directed graph based on each of the decrypted binary vectors; The second determining module is used to determine the target sorting method of the nodes in the directed graph, and based on the target sorting method and the directed graph, determine the peer evaluation ranking result of the salesperson's business quality. The determination of the directed graph based on each of the decrypted binary vectors includes: Obtain the business completion information of each salesperson in the business quality mutual evaluation task; The business completion information is input into a binary discriminant model to obtain a binary discriminant vector output by the binary discriminant model; the binary discriminant model is used to determine the business quality relationship between two salespersons. The loss function is calculated based on the binary discriminant vector and the decryption binary vector to obtain the loss value of the binary discriminant model; If the loss value is less than a preset threshold, a random number is generated; Based on the random number, a directed graph is determined.

8. An electronic device comprising a processor and a memory storing an artificial intelligence program, characterized in that, When the processor executes the artificial intelligence program, it implements the business quality mutual evaluation method according to any one of claims 1 to 6.

9. A storage medium, said storage medium being an artificial intelligence readable storage medium, comprising an artificial intelligence program, characterized in that, When the artificial intelligence program is executed by the processor, it implements the business quality mutual evaluation method according to any one of claims 1 to 6.

Images