Business management method and management decision-making platform based on intelligent knowledge graph

Abstract

The present application discloses a business management method and management decision platform based on intelligent knowledge graph. It includes: obtaining the first customer data of the first customer; using the grid management model to analyze the first customer data, obtaining the corresponding first push grid, and matching the first push grid with the grid management table to determine the first push strategy and the first customer manager corresponding to the first customer; when the first customer initiates a business processing request according to the first push strategy, obtaining the first business processing process data when the first customer manager handles the business processing request; using the risk classification model to analyze the first business processing process data, obtaining the corresponding risk category label, and executing the corresponding risk management decision on the first business processing process data according to the risk category label. The present application solves the technical problem that the related technology cannot quickly guide and assist customers in handling business, and cannot predict business handling risks in real time, resulting in low business handling efficiency and high risk.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and specifically, to a business management method and management decision-making platform based on intelligent knowledge graph. Background Art

[0002] At present, when handling business in the financial field, especially in rural and town scenarios, most of the account managers handle the corresponding business for customers offline, resulting in a long business processing cycle, low business processing efficiency, and poor customer experience, which fails to generate sufficient appeal to high-quality customers; in addition, the information stored back for offline business handling is mostly in paper version, and the risks in the business handling process are difficult to predict, which makes it difficult for account managers to quickly review the risks of business handling after handling the business.

[0003] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0004] The embodiments of the present application provide a business management method and management decision-making platform based on an intelligent knowledge graph, so as to at least solve the technical problems that related technologies are unable to quickly guide and assist customers in handling business, and are unable to predict business handling risks in real time, resulting in low business handling efficiency and high risks.

[0005] According to one aspect of an embodiment of the present application, a business management method based on an intelligent knowledge graph is provided, including: obtaining first customer data of a first customer within an enterprise business area; analyzing the first customer data using a pre-trained grid management model to determine a first push grid corresponding to the first customer, and matching the first push grid with a preset grid management table to determine a first push strategy and a first account manager corresponding to the first customer, wherein the grid management table records the correspondence between each push grid obtained by dividing the enterprise business area and the account manager and the push strategy; when the first customer initiates a business processing request according to the first push strategy, obtaining business processing process data of the first account manager in the process of handling the business processing request; analyzing the first business processing process data using a pre-trained risk classification model to obtain a corresponding risk category label, and executing corresponding risk management decisions on the first business processing process data according to the risk category label.

[0006] Optionally, the training process of the grid management model includes: obtaining multiple groups of first training samples, wherein each group of first training samples includes: second customer data of a second customer within the enterprise's business area and a second push grid corresponding to the second customer; constructing a first neural network model; for each group of training samples, inputting the second customer data in the training samples into the first neural network model to determine the predicted push grid corresponding to the second customer; determining a first target loss function based on the second push grids and predicted push grids in the multiple groups of first training samples, and adjusting the model parameters of the first neural network model through the first target loss function until the model converges to obtain a grid management model.

[0007] Optionally, obtaining multiple groups of first training samples includes: obtaining initial customer data of multiple second customers within the business area of ​​the enterprise, and preprocessing the initial customer data to obtain second customer data of each second customer, wherein the preprocessing includes at least one of the following: data cleaning, data organization, and data standardization, and the second customer data includes at least one of the following: basic data, financial data, consumer behavior data, and business processing data; using a clustering algorithm to extract features from the second customer data of each second customer to obtain at least one key feature related to the business needs of each second customer, wherein the type of the key feature includes at least one of the following: consumer behavior features, loan behavior features, and credit business processing features; according to at least one key feature of each second customer, using a grid management method to divide each second customer into a corresponding second push grid; the second customer data of multiple second customers and the second push grids corresponding to each second customer form multiple groups of first training samples.

[0008] Optionally, the process of constructing the grid management table includes: for each push grid within the enterprise business area, constructing a grid portrait of the push grid based on at least one key feature related to the business needs of each second customer in the push grid, and determining the propensity score of each second customer in the push grid for different push strategies based on the grid portrait, and using the push strategy with a propensity score not lower than a preset score threshold as the push strategy of the push grid; obtaining task data of multiple push tasks within the enterprise business area, wherein the task data at least includes: account information of the account manager assigned to the push task; determining the mapping relationship between the account information of each account manager and each push task, and the corresponding relationship between each second customer and each push task, respectively, based on the second customer data of each second customer and the task data of each push task; analyzing the mapping relationship and the corresponding relationship using a second neural network model to determine the customer allocation information corresponding to the account information of each account manager, and determining at least one push grid served by the account manager based on the customer allocation information; and constructing a grid management table based on the push strategies and account managers corresponding to multiple push grids.

[0009] Optionally, determining the propensity score of each second customer in the push grid for different push strategies based on the grid portrait includes: analyzing the grid portrait, dividing the plurality of second customers in the push grid into a plurality of customer groups, and calculating the propensity score of each customer group in the push grid for different push strategies according to the following formula:

[0010]

[0011] in, represents the propensity score of customer group x for the push strategy containing the yth push strategy feature, represents the evaluation score of customer group x on the zth push strategy, represents the push strategy containing the yth push strategy feature and the zth push strategy, represents the number of times the push strategy containing the rth push strategy feature is responded to by the customer group, Z represents the total number of push strategies, and R represents the total number of push strategy features.

[0012] Optionally, before determining at least one push grid served by the account manager based on the customer allocation information, the method also includes: obtaining historical business processing record information of each second customer in the account manager's service push grid, wherein the historical business processing record information includes at least one of the following: the second customer's feedback and evaluation on the account manager's business processing, and the timeliness of the business processing; determining the allocation influencing factors based on the historical business processing records, and adjusting the customer allocation information based on the allocation influencing factors.

[0013] Optionally, the training process of the risk classification model includes: obtaining multiple groups of second training samples and multiple groups of third training samples, wherein the second training samples include: second business processing process data and a first category label corresponding to the second business processing process data, and the third training samples include: similar business processing process data whose data similarity with the second business processing process data is higher than a preset threshold and a second category label corresponding to the similar business processing process data; constructing a third neural network model; for each group of second training samples, using the third neural network model to analyze the second business processing process data to obtain the corresponding first prediction category label; for each group of third training samples, using the third neural network model to analyze the similar business processing process data to obtain the corresponding second prediction category label; constructing a multi-layer adaptive loss function based on the first category labels and the corresponding first prediction category labels in the multiple groups of second training samples and the second category labels and the corresponding second prediction category labels in the multiple groups of third training samples, and constructing a classification loss function based on the first category labels and the corresponding first prediction category labels in each group of second training samples; determining the second target loss function based on the multi-layer adaptive loss function and the classification loss function, and updating the model parameters of the third neural network model based on the second target loss function until the model converges to obtain a risk classification model.

[0014] According to another aspect of the embodiment of the present application, a management decision platform is also provided, which includes: a push management system, a business management system, and a risk control management system, wherein the push management system is used to obtain first customer data of a first customer within the business area of ​​the enterprise; use a pre-trained grid management model to analyze the first customer data, determine a first push grid corresponding to the first customer, and match the first push grid with a preset grid management table to determine a first push strategy and a first account manager corresponding to the first customer, wherein the grid management table records the correspondence between each push grid obtained by dividing the business area of ​​the enterprise and the account manager and the push strategy; the business management system is used to obtain first business processing process data in the process of the first account manager handling the business processing request when the first customer initiates a business processing request according to the first push strategy; the risk control management system is used to analyze the first business processing process data using a pre-trained risk classification model to obtain a corresponding risk category label, and execute corresponding risk management decisions on the first business processing process data according to the risk category label.

[0015] Optionally, the management decision platform also includes: a resource database, wherein the resource database is used to periodically obtain multiple types of resource information and store the resource information according to the type of each type of resource information, wherein the types of resource information include but are not limited to: product introduction information, training and learning information, operation service information, and corporate culture information.

[0016] According to another aspect of an embodiment of the present application, a computer program product is also provided, which includes a stored computer program, wherein the computer program implements the above-mentioned business management method based on intelligent knowledge graph when executed by a processor.

[0017] In the embodiment of the present application, the management decision platform analyzes the first customer data of the first customer by acquiring the grid management model to quickly and intelligently determine the first push grid corresponding to the first customer, and at the same time determines the first push strategy and the first customer manager corresponding to the first customer according to the pre-established grid management table, so that when the first customer initiates a business processing request through the push channel corresponding to the corresponding first push strategy, the first customer manager can quickly and targetedly provide services to the first customer. At the same time, in order to ensure the security of the first customer manager in handling business for the first customer, the management decision platform can also obtain the business processing process data of the first customer manager in the process of handling the business processing request, and use the pre-trained risk classification model to analyze the business processing process data, obtain the risk category label corresponding to the business processing process data, and execute the corresponding risk management decision on the business processing process data according to the risk category label, wherein the risk management decision includes intercepting, verifying or releasing the data. Thereby realizing the online rapid processing of customer business, and real-time monitoring of customer business processing process data to reduce business processing risks, thereby solving the technical problems that the relevant technology cannot quickly guide and assist customers in handling business, and cannot predict business processing risks in real time, resulting in low business processing efficiency and high risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0019] Figure 1 is a schematic diagram of the structure of an optional management decision platform according to an embodiment of the present application;

[0020] Figure 2 is a schematic diagram of the structure of another optional management decision platform according to an embodiment of the present application;

[0021] Figure 3 is a schematic diagram of an optional architecture of a resource database 11 according to an embodiment of the present application;

[0022] Figure 4 is a schematic diagram of the architecture of an optional push management system 12 according to an embodiment of the present application;

[0023] Figure 5is a schematic diagram of the architecture of an optional business management system 13 according to an embodiment of the present application;

[0024] Figure 6 is a schematic diagram of the architecture of an optional risk control management system 14 according to an embodiment of the present application;

[0025] Figure 7 It is a hardware structure block diagram of an optional computer terminal (or mobile device) for implementing business management based on intelligent knowledge graph according to an embodiment of the present application;

[0026] Figure 8 It is a flow chart of an optional business management based on intelligent knowledge graph according to an embodiment of the present application. DETAILED DESCRIPTION

[0027] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0028] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0029] In addition, the relevant information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for display, data for analysis, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. For example, an interface is set up between this system and the relevant users or organizations. Before obtaining relevant information, it is necessary to send an acquisition request to the aforementioned user or organization through the interface, and obtain relevant information after receiving the consent information fed back by the aforementioned user or organization.

[0030] In order to better understand the embodiments of the present application, the technical terms involved in the embodiments of the present application are explained as follows:

[0031] Smoke index: An indicator used to measure the density of smoke produced during a fire or combustion process. It is often used to assess the severity of the fire and its possible impact on the environment, human health and property.

[0032] Example 1

[0033] According to an embodiment of the present application, a management decision platform is provided, and the architecture diagram of the management decision platform is as follows: Figure 1 As shown, it includes: resource database 11, push management system 12, business management system 13, risk control management system 14, among which,

[0034] The resource database 11 is used to periodically obtain multiple types of resource information and store the resource information according to the types of each type of resource information, wherein the types of resource information include but are not limited to: product introduction information, training and learning information, operation service information, and corporate culture information.

[0035] The push management system 12 is used to obtain the first customer data of the first customer within the business area of ​​the enterprise; use the pre-trained grid management model to analyze the first customer data, determine the first push grid corresponding to the first customer, and match the first push grid with the preset grid management table to determine the first push strategy and the first customer manager corresponding to the first customer, wherein the grid management table records the correspondence between each push grid obtained by dividing the business area of ​​the enterprise and the customer manager and push strategy.

[0036] The business management system 13 is used to obtain the business processing process data of the first account manager in the process of processing the business processing request when the first customer initiates the business processing request according to the first push strategy.

[0037] The risk control management system 14 is used to analyze the business processing process data using a pre-trained risk classification model, obtain risk category labels corresponding to the business processing process data, and execute corresponding risk management decisions on the business processing process data according to the risk category labels.

[0038] Figure 2 This is a schematic diagram of the architecture of another more complete management decision-making platform according to an embodiment of the present application. Figure 2 The specific structure of each system in the management decision-making platform is explained.

[0039] Specifically, the resource database 11 includes a plurality of storage units, and each storage unit is used to store different types of resource information.

[0040] Figure 3 is a schematic diagram of the structure of an optional resource database 11 according to an embodiment of the present application, such as Figure 3 As shown. The resource database 11 includes four major sections: basic knowledge, training and learning, financial products, and operational services, and the relevant resource content of each section is stored in the corresponding storage unit. Each section covers a number of learning materials such as organizational establishment, corporate culture, system construction, on-the-job training, financial policies, examination question banks, service etiquette, warning education, employee behavior norms, personal financial products, corporate financial products, and promotional materials, so as to meet the learning needs of employees in different positions and departments of the enterprise, so that each employee can easily find learning content related to their work.

[0041] Specifically, the push management system 12 is mainly used to dynamically associate the comprehensively collected customer information (i.e., the customer data of newly acquired target customers) with the customer stock information (i.e., the customer data acquired historically), and uses the push grid as a carrier to unify the management of customer information, business information, customer consumption behavior information, business opportunity information, push information, push tasks and performance monitoring, so as to realize the establishment of customer files and cards, and effectively support grid-based push and precise push.

[0042] The push management system 12 at least includes: a customer management module 121, a grid division module 122, and a grid management module 123, wherein:

[0043] The customer management module 121 may obtain initial customer data of a first customer within the business area of ​​the enterprise, and pre-process the initial customer data to obtain first customer data of the first customer.

[0044] Specifically, the customer management module 121 can obtain the initial customer data of any customer (i.e., the first customer) within the business area of ​​the enterprise through multiple channels, such as mobile phone APPs with positioning functions, WeChat applets, official accounts, etc., and perform pre-processing operations on the acquired initial customer data. The specific steps include but are not limited to: checking data consistency, processing invalid values, processing missing values, deleting duplicate information and correcting erroneous information to obtain high-quality first customer data, which includes but is not limited to: basic data, financial data, consumer behavior data, and business processing data.

[0045] The grid division module 122 may analyze the first customer data using a grid management model to determine a first push grid corresponding to the first customer.

[0046] Optionally, the grid division module 122 may train the grid management model according to the following method, including:

[0047] Step S11, obtaining multiple groups of first training samples, wherein each group of first training samples includes: second customer data of a second customer in the business area of ​​the enterprise and a second push grid corresponding to the second customer;

[0048] Step S12, constructing a first neural network model;

[0049] Step S13, iteratively training the first neural network model using multiple groups of training samples to obtain a grid management model.

[0050] Furthermore, in the technical solution provided in the above step S11, the grid division module 122 can obtain multiple groups of first training samples through the following steps:

[0051] Step S111, obtaining second customer data of multiple second customers within the enterprise business area.

[0052] That is to say, the customer management module 121 can first obtain the initial customer data of multiple second customers within the enterprise's business area, and pre-process the initial customer data to obtain the second customer data of each second customer, wherein the pre-processing includes at least one of the following: data cleaning, data organization, and data standardization, and the second customer data includes at least one of the following: basic data, financial data, consumer behavior data, and business processing data; then, the grid division module 122 can directly obtain the second customer data of multiple second customers within the enterprise's business area from the customer management module 121.

[0053] Step S112: extracting features from the second customer data of each second customer using a clustering algorithm to obtain at least one key feature related to the business needs of each second customer.

[0054] That is, the grid division module 122 uses a clustering algorithm (such as a k-means clustering algorithm, a hierarchical clustering algorithm, etc.) or a naive Bayes algorithm to group the second customer data of multiple second customers to obtain corresponding clusters; after clustering is completed, the characteristics of the second customer data in each cluster are analyzed, and the common characteristics or representative characteristics of the second customer data in the cluster are extracted, and at least one key feature related to the business needs of the second customer is screened from the extracted common characteristics or representative characteristics, such as consumption behavior characteristics, loan behavior characteristics, credit business processing characteristics, etc. It should be noted that in order to quickly extract at least one key feature related to the business needs of the second customer, before clustering, the principal component analysis method can be used for dimensionality reduction to reduce the feature dimension.

[0055] Step S113: According to at least one key feature of each second customer, each second customer is divided into a corresponding second push grid using a grid management method.

[0056] Specifically, the grid division module 122 may calculate the similarity between each key feature of each second customer, and divide multiple second customers whose similarity is greater than a preset similarity threshold into the same push grid to obtain a second push grid corresponding to each second customer.

[0057] Step S114 , forming a plurality of groups of first training samples from the second customer data of the plurality of second customers and the second push grids corresponding to the respective second customers.

[0058] In addition, in the technical solution provided in the above step S2, the model type of the first neural network model constructed includes but is not limited to: convolutional neural network model, recurrent neural network model, clustering model, naive Bayes model, etc.

[0059] In the technical solution provided in the above step S13, the grid division module 122 can be trained to obtain a grid management model through the following steps:

[0060] Step S131, for each group of first training samples, inputting the second customer data of the second customer in the first training sample into the first neural network model to determine the prediction grid corresponding to the second customer;

[0061] Step S132, determining a first target loss function based on the second push grids and predicted push grids in the multiple groups of first training samples, and adjusting the model parameters of the first neural network model by the first target loss function until the model converges to obtain a grid management model.

[0062] It should be noted that after the grid division module 122 divides the first customer into grids using the above-mentioned pre-trained grid management model, it can adjust and optimize the grid management model according to the accuracy of the final division result, so that the model performance of the grid management model is continuously improved. Compared with the traditional manual division method, the present application scheme does not rely on the historical experience and subjective judgment of the account manager, so the division efficiency and accuracy are higher.

[0063] Furthermore, the grid management module 123 may first construct a grid management table; match the first push grid with the preset grid management table to determine the first push strategy and the first account manager corresponding to the first customer. Therefore, the grid management module 123 may be subdivided into a grid management table formulation unit 1231 and a matching unit 1232 according to the function, wherein the grid management table formulation unit 1231 is mainly used to construct a grid management table for the business area of ​​the enterprise; and the matching unit 1232 may be used to match the first push grid with the preset grid management table to determine the first push strategy and the first account manager corresponding to the first customer.

[0064] Furthermore, the grid management table formulation unit 1231 can be further subdivided into a strategy formulation subunit 12311 and a customer manager determination subunit 12312, wherein:

[0065] For each push grid in the enterprise business area, the strategy formulation subunit 12311 may determine its corresponding push strategy according to the following method, including:

[0066] Step 1: Build a grid profile of the push grid based on at least one key feature related to the business needs of each second customer in the push grid.

[0067] Step 2: Determine the propensity score of each second customer in the push grid for different push strategies based on the grid portrait.

[0068] The second step above can be implemented as follows: First, analyze the grid portrait and divide the multiple second customers in the push grid into multiple customer groups. Among them, the customers in each customer group have higher similarity and identity; then, calculate the propensity score of each customer group in the push grid for different push strategies according to the following formula:

[0069]

[0070] in, represents the propensity score of customer group x for the push strategy containing the yth push strategy feature, represents the evaluation score of customer group x on the zth push strategy, represents the push strategy containing the yth push strategy feature and the zth push strategy, represents the number of times the push strategy containing the rth push strategy feature is responded to by the customer group, Z represents the total number of push strategies, and R represents the total number of push strategy features

[0071] Step 3: Use the push strategy with a propensity score not lower than the preset score threshold as the push strategy for the push grid.

[0072] Among them, the above third step is to obtain the propensity score of each customer group through the above propensity score calculation formula, and use the push strategy with the highest propensity score as the push strategy of the customer group, summarize the push strategies of each customer group in the push grid, and finally obtain the push strategy set of the push grid. This method can achieve accurate task push for each customer in the push grid and improve customer experience.

[0073] In addition, the strategy formulation subunit 12311 can also periodically obtain the implementation performance of each push network in executing the corresponding push strategy, and adjust the push strategy of the push network according to the implementation performance and preset performance indicators, so as to continuously optimize and improve the push strategy to enhance customer satisfaction.

[0074] The account manager determination subunit 12312 may determine the account manager providing services to the customers in the push grid in the following manner:

[0075] Step 1: Obtain task data of multiple push tasks within the business area of ​​the enterprise, wherein the task data at least includes: account information (such as account ID) of the account manager assigned to the push task;

[0076] Step 2: According to the second customer data of each second customer and the task data of each push task, respectively determine the mapping relationship between the account information of each customer manager and each push task, and the corresponding relationship between each second customer and each push task.

[0077] Step 3: Analyze the mapping relationship and the corresponding relationship using the second neural network model to determine the customer allocation information corresponding to the account information of each account manager, and determine at least one push grid served by the account manager based on the customer allocation information.

[0078] That is to say, the account manager determination subunit 12312 considers the customer's historical participation information for a certain push task into the considerations for assigning customers to account managers, such as reassigning a customer who has participated in a certain push task to the task. Therefore, the account manager determination subunit 12312 can input the mapping relationship between push tasks and account managers and the corresponding relationship between push tasks and customers into the second neural network model. The second neural network model can quickly and accurately match and classify different data by pre-learning the relationship between the data, thereby obtaining customer allocation information assigned to each account manager.

[0079] Furthermore, with the grid as the key, the push strategy corresponding to each push grid and the account manager of each customer in the service push grid as the value, multiple key-value pairs are formed, and the grid management table of the enterprise business area is constructed by these key-value pairs.

[0080] In addition, in order to more accurately assign information about the customers served by the account managers, the account manager determination subunit 12312 can first obtain the historical business processing record information of each second customer in the account manager service push grid after determining the customer allocation information corresponding to the account information of each account manager, wherein the historical business processing record information includes at least one of the following: the second customer's feedback and evaluation on the account manager's business processing, and the timeliness of the business processing; determine the allocation influencing factors based on the historical business processing records, and adjust the customer allocation information based on the allocation influencing factors.

[0081] Figure 4 is a schematic diagram of an optional push management system 12 according to an embodiment of the present application, such as Figure 4 As shown. The push management system can include four major functional modules: customer management, grid management, push management, and operation analysis. Among them, customer management refers to collecting customer data of multiple customers in the business area of ​​the enterprise through multiple data sources, and establishing files and cards for each customer. Based on the collected customer data, a complete profile, credit rating and customer relationship analysis are performed on the customer, so that the enterprise can more accurately grasp customer needs, thereby providing customers with more personalized customized services and products, and at the same time fully record the service records of account managers to customers, making customer management more refined; grid management refers to the key characteristics of each customer in the business area of ​​the enterprise, and the business area of ​​the enterprise is divided into Multiple customers within the company are divided into different grids, and based on the relationship between push task data and customer data, different push strategies are formulated for each grid and at least one account manager is assigned to serve each grid to ensure that the entire enterprise's business area management is seamless; push management refers to being driven by push tasks within the grid, completely recording the account manager's push work, adjusting the push strategy corresponding to the grid according to the final push result of the push strategy, and forming a closed loop of push management work; operational analysis refers to analyzing customer distribution, product penetration, account manager's performance, and push results to adjust the company's performance appraisal indicators for account managers, business strategies, etc.

[0082] Optionally, the business management system 13 may obtain business processing process data during the process of the first customer manager handling the business processing request when the first customer initiates the business processing request according to the first push strategy.

[0083] Specifically, the business management system 13 can be divided into a mobile application system and a background management system according to different user objects. Figure 5 A schematic diagram of a processing flow of an optional business management system 13 on a mobile terminal according to an embodiment of the present application is shown as follows: Figure 5As shown. For account managers, when processing business processing requests initiated by customers in the corresponding push grid, the mobile business management system can realize the following functions: customer information management, new query authorization, pre-loan due diligence, business application, electronic contract signing, loan issuance, post-loan inspection, etc.; and for business approval personnel, the mobile business management system can support mobile approval for business initiated by account managers through mobile terminals. Therefore, the business processing process data includes but is not limited to: contract signing, electronic signature, face recognition, processing of image data, etc. The above-mentioned business management system 13 can realize paperless management of business processing materials, thereby greatly improving the efficiency of business processing; at the same time, the electronic credentials and image information materials in the business processing process are stored and called, thereby greatly improving the security of business processing and improving the chain of evidence of business processing materials.

[0084] It should be noted that the main functions of the above-mentioned business management system 13 at the backend include: basic parameter configuration of the business management system 13, business parameter configuration and operation analysis function.

[0085] Specifically, the risk control management system 14 can use the pre-trained risk classification model to analyze the business processing process data, obtain the risk category labels corresponding to the business processing process data, and execute corresponding risk management decisions on the business processing process data according to the risk category labels.

[0086] Optionally, the risk control management system 14 may train the risk classification model according to the following method, including:

[0087] Step S21, obtaining multiple groups of second training samples and multiple groups of third training samples.

[0088] Specifically, the second training sample includes: second business processing process data and a first category label corresponding to the second business processing process data; the third training sample includes: similar business processing process data whose data similarity with the second business processing process data is higher than a preset threshold and a second category label corresponding to the similar business processing process data.

[0089] Step S22, constructing a third neural network model.

[0090] Similarly, the model types of the third neural network model include but are not limited to: convolutional neural network model, recurrent neural network model, etc.

[0091] Step S23: for each group of second training samples, the second business handling process data is analyzed using the third neural network model to obtain the corresponding first prediction category label.

[0092] Step S24: for each group of third training samples, similar business handling process data is analyzed using the third neural network model to obtain a corresponding second prediction category label.

[0093] Step S25, constructing a multi-layer adaptive loss function based on the first category labels and the corresponding first predicted category labels in multiple groups of second training samples and the second category labels and the corresponding second predicted category labels in multiple groups of third training samples, and constructing a classification loss function based on the first category labels and the corresponding first predicted category labels in each group of second training samples.

[0094] Among them, the expression of the above multi-layer adaptive loss function can be written as:

[0095]

[0096] in, Indicates that the second business handling process data is The result obtained after layer-by-layer network processing, Indicates that similar business processing data through the second business processing data through the first The result obtained after layer-by-layer network processing, represents the network parameters of the third neural network, The second business processing process data indicating similar business processing process data and the preset category label is in the The maximum mean difference (MMD) of the network layer reflects the difference between these two types of data in the first Layer network distribution distance, A hyperparameter that represents the degree of penalty for distance from the distribution.

[0097] The expression of the above classification loss function can be written as:

[0098]

[0099] Wherein, the above N represents the total number of second training samples, represents the j-th second business handling process data, represents the network parameters of the third neural network, represents the first predicted category label obtained by processing the j-th second business handling process data through the third neural network, represents the first category label of the j-th second business handling process data, and J represents the cross entropy loss function.

[0100] Step S26, determining a second target loss function based on the multi-layer adaptive loss function and the classification loss function, and updating the model parameters of the third neural network model based on the second target loss function until the model converges to obtain a risk classification model.

[0101] Furthermore, when the risk control management system 14 uses the risk classification model trained in the above steps S21-S26 to analyze the first business handling request process data (types of which include: data files, image data, etc. in the business handling process) initiated by the first account manager to handle the target customer, the risk category label corresponding to the first business handling process data is obtained, wherein the above risk category label may include: high risk, medium risk, low risk. Then, the risk control management system 14 can also perform corresponding risk management decisions on the first business handling process data according to the above determined risk category label, including:

[0102] First, determine the risk value corresponding to the risk category label of the first business handling process data;

[0103] When the risk value is greater than the first preset risk threshold, it indicates that the risk of the first business processing process data is high, and the management decision platform can send a corresponding first warning message to the account manager's account. The first warning message is used to prompt the interception of the business processing process data.

[0104] When the risk value is not greater than the first preset risk threshold and greater than the second preset risk threshold, it indicates that the first business handling process data has certain risks. At this time, the management decision platform can send a second warning message to the account manager's account, wherein the second warning message is used to prompt the verification of the first business handling process data;

[0105] When the risk value is not greater than the second preset risk threshold, it indicates that the risk of the first business processing process data is low and can be ignored. At this time, the target customer manager continues to handle the target customer's business processing request.

[0106] Specifically, Figure 6 is a schematic diagram of a processing flow of an optional risk control management system 14 according to an embodiment of the present application, such as Figure 6As shown. The risk control management system includes several major functional modules such as risk model management, process management, early warning management, and business risk classification. Among them, risk model management is to select model elements, model rules, and matching dimensions according to the business scope of the enterprise to build a risk classification model; process management is to formulate a risk approval process based on matching elements and process models; early warning management configures the frequency and time configuration of early warning products to provide corresponding early warning prompts for each business process data; business risk classification is to conduct initial classification, certification, and review of various business process data based on the characteristics of various business data and the above-mentioned risk classification model and risk approval process, and finally generate a risk classification report. Therefore, through the above-mentioned risk management platform, it can be connected with the associated business system to obtain the corresponding business process data, and then the system selects different risk classification models for various businesses to identify, measure, evaluate, and analyze the risks of the business process data, and finally generate risk classification results for the associated system to use the risk classification results.

[0107] In the embodiment of the present application, the management decision platform analyzes the first customer data of the first customer by acquiring the grid management model to quickly and intelligently determine the first push grid corresponding to the first customer, and at the same time determines the first push strategy and the first customer manager corresponding to the first customer according to the pre-established grid management table, so that when the first customer initiates a business processing request through the push channel corresponding to the corresponding first push strategy, the first customer manager can quickly and targetedly provide services to the first customer. At the same time, in order to ensure the security of the first customer manager in handling business for the first customer, the management decision platform can also obtain the first business processing process data in the process of the first customer manager handling the business processing request, and use the pre-trained risk classification model to analyze the first business processing process data, obtain the risk category label corresponding to the first business processing process data, and execute the corresponding risk management decision on the first business processing process data according to the risk category label, wherein the risk management decision includes intercepting, verifying or releasing the data. Thereby realizing the online rapid processing of customer business, and real-time monitoring of the first business processing process data to reduce the business processing risk, thereby solving the technical problems that the related technology cannot quickly guide and assist customers in handling business, and cannot predict the business processing risk in real time, resulting in low business processing efficiency and high risk.

[0108] Example 2

[0109] In addition, the above-mentioned management decision platform can execute the business management method based on the intelligent knowledge graph provided in the embodiment of the present application in a mobile terminal, a computer terminal or a similar computing device. Figure 7 The hardware structure block diagram of a computer terminal (or mobile device) for implementing a business management method based on an intelligent knowledge graph is shown. Figure 7 As shown, the computer terminal 70 (or mobile device 70) may include one or more (shown as 702a, 702b, ..., 702n in the figure) processors 702 (the processor 702 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 704 for storing data, and a transmission device 706 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply and / or a camera. It can be understood by those skilled in the art that Figure 7 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 7 More or fewer components as shown, or with Figure 7 Different configurations shown.

[0110] It should be noted that the one or more processors 702 and / or other data processing circuits described above may generally be referred to herein as "data processing circuits". The data processing circuits may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuit may be a single independent processing module, or may be incorporated in whole or in part into any of the other components in the computer terminal 70 (or mobile device). As involved in the embodiments of the present application, the data processing circuit acts as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0111] The memory 704 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the business management method based on the intelligent knowledge graph in the embodiment of the present application. The processor 702 executes various functional applications and data processing by running the software programs and modules stored in the memory 704, that is, the business management method based on the intelligent knowledge graph of the above-mentioned application is realized. The memory 704 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 704 may further include a memory remotely arranged relative to the processor 702, and these remote memories may be connected to the computer terminal 70 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0112] The transmission device 706 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the computer terminal 70. In one example, the transmission device 706 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 706 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0113] The display may be, for example, a touch screen liquid crystal display (LCD) that enables a user to interact with a user interface of the computer terminal 70 (or mobile device).

[0114] In the above operating environment, Figure 8 is a flow chart of an optional business management method based on intelligent knowledge graph according to an embodiment of the present application, such as Figure 8 As shown, the method at least includes steps S802-S808, wherein:

[0115] Step S802, obtaining the first customer data of the first customer in the business area of ​​the enterprise.

[0116] In the technical solution provided in step S802, the management decision-making platform can obtain the first customer data of the first customer in the business area of ​​the enterprise, wherein the first customer data includes at least a plurality of customer attribute data, such as basic data, financial data, consumer behavior data, business handling data, etc. Among them, the basic data includes but is not limited to: age, gender, education, occupation, income, etc.; the financial data includes at least one of the following: deposit status, loan status, preference for purchasing financial products, etc.; the consumer behavior data includes at least one of the following: transaction frequency, transaction amount, transaction channel, etc.; the business handling data includes at least one of the following: loan business default rate, business transaction risk level, etc.

[0117] Step S804, using the pre-trained grid management model to analyze the first customer data, determine the first push grid corresponding to the first customer, and match the first push grid with the preset grid management table to determine the first push strategy and the first customer manager corresponding to the first customer.

[0118] In the technical solution provided in step S804, when the related technology adopts the grid management method for each customer in the enterprise business area, it usually only divides the enterprise business area into multiple grids (also called push units) based on geographical areas or markets, but this method cannot perform targeted push for different types of customers, resulting in poor push efficiency. For this reason, the management decision platform in the embodiment of the present application can use the grid management model to analyze the first customer data to quickly and intelligently determine the first push grid corresponding to the first customer. Further, in order to better push business to the first customer and provide business handling services, the management decision platform can also match the target push grid with the preset grid management table to obtain the first customer manager and the first push strategy for business push to the first customer, wherein the grid management table is the management decision platform pre-analyzing the customer data, customer manager data, and task data in the enterprise business area, and obtaining the corresponding relationship between each push grid in the enterprise business area and the customer manager of the customer in the service push grid, and the corresponding relationship between each push grid and the push strategy.

[0119] Step S806, when the first customer initiates a business processing request according to the first push strategy, obtain the first business processing process data in the process of the first customer manager processing the business processing request.

[0120] In the technical solution provided in step S806, when the first customer initiates a business processing request through the push channel corresponding to the first push strategy, the management decision platform can obtain the first business processing process data when the first customer manager handles the business processing request, wherein the first business processing process data can be data at any processing stage of the business processing process, such as the customer information management stage (including customer creation, customer information synchronization, customer information maintenance, etc.), the credit verification stage (including the signing of the credit authorization letter, the review of the credit report, etc.), the business application stage (including loan application, due diligence, signing of electronic materials, etc.), the mobile approval stage (including business information review, business processing image material review, department leader signing of approval opinions, etc.), the loan issuance stage (including loan issuance entry, loan note signing, collection of loan issuance images, etc.), and the post-loan management stage (including post-loan overdue review, post-loan inspection entry, signing of electronic reports, etc.).

[0121] Step S805, using the pre-trained risk classification model to analyze the first business processing process data, obtain corresponding risk category labels, and execute corresponding risk management decisions on the first business processing process data according to the risk category labels.

[0122] In the technical solution provided in step S808, the management decision platform can use the pre-trained risk classification model to analyze the first business processing process data, obtain the risk category label corresponding to the first business processing process data, and execute corresponding risk management decisions on the first business processing process data according to the risk category label to greatly reduce the risk of business processing.

[0123] The above method of this embodiment is further introduced below.

[0124] As an optional implementation, in the technical solution provided in the above step S804, the management decision platform may train the grid management model by a method including:

[0125] Step S11, obtaining multiple groups of first training samples, wherein each group of first training samples includes: second customer data of a second customer in the business area of ​​the enterprise and a second push grid corresponding to the second customer;

[0126] Step S12, constructing a first neural network model;

[0127] Step S13, iteratively training the first neural network model using multiple groups of training samples to obtain a grid management model.

[0128] Furthermore, in the technical solution provided in the above step S11, the management decision platform can obtain multiple groups of first training samples through the following steps:

[0129] Step S111, obtaining second customer data of multiple second customers within the enterprise business area.

[0130] That is to say, the management decision-making platform can first obtain the initial customer data of multiple second customers within the enterprise's business area, and pre-process these initial customer data to obtain the second customer data of each second customer, wherein the pre-processing includes at least one of the following: data cleaning, data organization, and data standardization, and the second customer data includes at least one of the following: basic data, financial data, consumer behavior data, and business processing data.

[0131] Step S112: extracting features from the second customer data of each second customer using a clustering algorithm to obtain at least one key feature related to the business needs of each second customer.

[0132] Specifically, the management decision platform uses a clustering algorithm (such as a k-means clustering algorithm, a hierarchical clustering algorithm, etc.) or a naive Bayes algorithm to group the second customer data of multiple second customers to obtain corresponding clusters; after clustering is completed, the characteristics of the second customer data in each cluster are analyzed, and the common characteristics or representative characteristics of the second customer data in the cluster are extracted, and at least one key feature related to the business needs of the second customer is screened from the extracted common characteristics or representative characteristics, such as consumption behavior characteristics, loan behavior characteristics, credit business processing characteristics, etc. It should be noted that in order to quickly extract at least one key feature related to the business needs of the second customer, before clustering, the principal component analysis method can be used for dimensionality reduction to reduce the feature dimension.

[0133] Step S113: According to at least one key feature of each second customer, each second customer is divided into a corresponding second push grid using a grid management method.

[0134] Specifically, the management decision platform may calculate the similarity between each key feature of each second customer, and classify multiple second customers whose similarity is greater than a preset similarity threshold into the same push grid to obtain a second push grid corresponding to each second customer.

[0135] Step S114 , forming a plurality of groups of first training samples from the second customer data of the plurality of second customers and the second push grids corresponding to the respective second customers.

[0136] In addition, in the technical solution provided in the above step S2, the model type of the first neural network model constructed includes but is not limited to: convolutional neural network model, recurrent neural network model, clustering model, naive Bayes model, etc.

[0137] In the technical solution provided in the above step S13, the management decision platform can train the grid management model through the following steps:

[0138] Step S131, for each group of first training samples, inputting the second customer data of the second customer in the first training sample into the first neural network model to determine the prediction grid corresponding to the second customer;

[0139] Step S132, determining a first target loss function based on the second push grids and predicted push grids in the multiple groups of first training samples, and adjusting the model parameters of the first neural network model by the first target loss function until the model converges to obtain a grid management model.

[0140] It should be noted that after the management decision platform performs grid division on the first customer through the above-mentioned pre-trained grid management model, it can adjust and optimize the grid management model according to the accuracy of the final division result, so that the model performance of the grid management model is continuously improved. Compared with the traditional manual division method, this application scheme does not rely on the historical experience and subjective judgment of the account manager, so the division efficiency and accuracy are higher.

[0141] As an optional implementation, in the technical solution provided in the above step S804, the management decision platform can construct the grid management table by the following method, including:

[0142] For each push grid in the enterprise's business area, first determine the push strategy for each push grid according to the following method, including:

[0143] Step 1: Build a grid profile of the push grid based on at least one key feature related to the business needs of each second customer in the push grid.

[0144] Step 2: Determine the propensity score of each second customer in the push grid for different push strategies based on the grid portrait.

[0145] The second step above can be implemented as follows: First, analyze the grid portrait and divide the multiple second customers in the push grid into multiple customer groups. Among them, the customers in each customer group have higher similarity and identity; then, calculate the propensity score of each customer group in the push grid for different push strategies according to the following formula:

[0146]

[0147] in, represents the propensity score of customer group x for the push strategy containing the yth push strategy feature, represents the evaluation score of customer group x on the zth push strategy, represents the push strategy containing the yth push strategy feature and the zth push strategy, represents the number of times the push strategy containing the rth push strategy feature is responded to by the customer group, Z represents the total number of push strategies, and R represents the total number of push strategy features

[0148] Step 3: Use the push strategy with a propensity score not lower than the preset score threshold as the push strategy for the push grid.

[0149] Among them, the above third step is to obtain the propensity score of each customer group through the above propensity score calculation formula, and use the push strategy with the highest propensity score as the push strategy of the customer group, summarize the push strategies of each customer group in the push grid, and finally obtain the push strategy set of the push grid. This method can achieve accurate task push for each customer in the push grid and improve customer experience.

[0150] Optionally, the management decision platform can also periodically obtain the implementation performance of each push network in executing the corresponding push strategy, adjust the push strategy of the push network according to the implementation performance and preset performance indicators, and continuously optimize and improve the push strategy to enhance customer satisfaction.

[0151] In addition, the management decision platform can also determine the account manager who provides services to customers in the push grid in the following way:

[0152] Step 1: Obtain task data of multiple push tasks within the business area of ​​the enterprise, wherein the task data at least includes: account information (such as account ID) of the account manager assigned to the push task;

[0153] Step 2: According to the second customer data of each second customer and the task data of each push task, respectively determine the mapping relationship between the account information of each customer manager and each push task, and the corresponding relationship between each second customer and each push task.

[0154] Step 3: Analyze the mapping relationship and the corresponding relationship using the second neural network model to determine the customer allocation information corresponding to the account information of each account manager, and determine at least one push grid served by the account manager based on the customer allocation information.

[0155] That is to say, the management decision platform considers the customer's historical participation information in a certain push task into the considerations for assigning customers to account managers, such as reassigning a customer who has participated in a certain push task to the task. Therefore, the management decision platform can input the mapping relationship between push tasks and account managers and the corresponding relationship between push tasks and customers into the second neural network model. The second neural network model can quickly and accurately match and classify different data by pre-learning the relationship between the data, thereby obtaining customer allocation information assigned to each account manager.

[0156] Furthermore, with the grid as the key, the push strategy corresponding to each push grid and the account manager of each customer in the service push grid as the value, multiple key-value pairs are formed, and the grid management table of the enterprise business area is constructed by these key-value pairs.

[0157] In addition, in order to more accurately assign customer allocation information to account managers, the management decision platform can first obtain the historical business processing record information of each second customer in the account manager service push grid after determining the customer allocation information corresponding to the account information of each account manager. The historical business processing record information includes at least one of the following: the second customer's feedback and evaluation on the account manager's business processing, and the timeliness of the business processing; determine the allocation influencing factors based on the historical business processing records, and adjust the customer allocation information based on the allocation influencing factors.

[0158] Finally, each push grid is used as a key, and the push strategy corresponding to the push grid and the customer manager serving the push grid are used as values ​​to form multiple key-value pairs, and the grid management table is constructed by these key-value pairs.

[0159] As an optional implementation, in the technical solution provided in the above step S808, the management decision platform may train the risk classification model by the following method, including:

[0160] Step S21, obtaining multiple groups of first training samples and multiple groups of second training samples.

[0161] Specifically, the second training sample includes: second business processing process data and a first category label corresponding to the second business processing process data; the third training sample includes: similar business processing process data whose data similarity with the second business processing process data is higher than a preset threshold and a second category label corresponding to the similar business processing process data.

[0162] Step S22, constructing a third neural network model.

[0163] Similarly, the model types of the third neural network model include but are not limited to: convolutional neural network model, recurrent neural network model, etc.

[0164] Step S23: for each group of second training samples, the second business handling process data is analyzed using the third neural network model to obtain the corresponding first prediction category label.

[0165] Step S24: for each group of third training samples, similar business handling process data is analyzed using the third neural network model to obtain a corresponding second prediction category label.

[0166] Step S25, constructing a multi-layer adaptive loss function based on the first category labels and the corresponding first predicted category labels in multiple groups of second training samples and the second category labels and the corresponding second predicted category labels in multiple groups of third training samples, and constructing a classification loss function based on the first category labels and the corresponding first predicted category labels in each group of second training samples.

[0167] Among them, the expression of the above multi-layer adaptive loss function can be written as:

[0168]

[0169] in, Indicates that the second business handling process data is The result obtained after layer-by-layer network processing, The similar business processing process data of the second business processing process data is represented by the first The result obtained after layer-by-layer network processing, represents the network parameters of the third neural network, The second business processing process data indicating similar business processing process data and the preset category label is in the The maximum mean difference (MMD) of the network layer reflects the difference between these two types of data in the first Layer network distribution distance, A hyperparameter that represents the degree of penalty for distance from the distribution.

[0170] The expression of the above classification loss function can be written as:

[0171]

[0172] Wherein, the above N represents the total number of second training samples, represents the j-th second business handling process data, represents the network parameters of the third neural network, represents the first predicted category label obtained by processing the j-th second business handling process data through the third neural network, represents the first category label of the j-th second business handling process data, and J represents the cross entropy loss function.

[0173] Step S26, determining a second target loss function based on the multi-layer adaptive loss function and the classification loss function, and updating the model parameters of the third neural network model based on the second target loss function until the model converges to obtain a risk classification model.

[0174] Furthermore, when the risk management system in the management decision platform uses the risk classification model trained in the above steps S21-S26 to analyze the first business handling request process data (types of which include: data files, image data, etc. in the business handling process) initiated by the first account manager to handle the target customer, the risk category label corresponding to the first business handling process data is obtained, wherein the above risk category label may include: high risk, medium risk, low risk. Then the risk control management system 14 can also perform corresponding risk management decisions on the first business handling process data according to the above determined risk category label, including:

[0175] First, determine the risk value corresponding to the risk category label of the first business handling process data;

[0176] When the risk value is greater than the first preset risk threshold, it indicates that the risk of the first business processing process data is high, and the management decision platform can send a corresponding first warning message to the account manager's account. The first warning message is used to prompt the interception of the business processing process data.

[0177] When the risk value is not greater than the first preset risk threshold and greater than the second preset risk threshold, it indicates that the first business handling process data has certain risks. At this time, the management decision platform can send a second warning message to the account manager's account, wherein the second warning message is used to prompt the verification of the first business handling process data;

[0178] When the risk value is not greater than the second preset risk threshold, it indicates that the risk of the first business processing process data is low and can be ignored. At this time, the target customer manager continues to handle the target customer's business processing request.

[0179] It should be noted that in the above steps, the neural network models used (including the first neural network model, the second neural network model, and the third neural network model) can be loaded into the memory. For example, the original data of the neural network model can be loaded from the non-volatile memory to the volatile memory so that the processor can run the neural network model, wherein the original data of the neural network model refers to unprocessed data, usually including parameters and structural data of the neural network model, and the structural data can be a calculation relationship based on the parameters, such as the forward propagation calculation relationship between intermediate layers and neurons. Specifically, the structural data can include code related to the structure of the first neural network, such as code for performing related calculations between intermediate layers and neurons.

[0180] In one embodiment, an area for loading a neural network model can be divided in the memory, which may include a structure data storage area and a parameter storage area. The structure data storage area is used to store structure-related codes, and the parameters referenced by it can point to the address of a specific parameter in the parameter storage area through a pointer. During the training process of the grid management model, it may be necessary to frequently update the parameters, and the parameter values ​​in the parameter storage area can be updated.

[0181] In the embodiment of the present application, the management decision platform analyzes the first customer data of the first customer by acquiring the grid management model to quickly and intelligently determine the first push grid corresponding to the first customer, and at the same time determines the first push strategy and the first customer manager corresponding to the first customer according to the pre-established grid management table, so that when the first customer initiates a business processing request through the push channel corresponding to the corresponding first push strategy, the first customer manager can quickly and targetedly provide services to the first customer. At the same time, in order to ensure the security of the first customer manager in handling business for the first customer, the management decision platform can also obtain the first business processing process data in the process of the first customer manager handling the business processing request, and use the pre-trained risk classification model to analyze the first business processing process data, obtain the risk category label corresponding to the first business processing process data, and execute the corresponding risk management decision on the first business processing process data according to the risk category label, wherein the risk management decision includes intercepting, verifying or releasing the data. Thereby realizing the online rapid processing of customer business, and real-time monitoring of the first business processing process data to reduce the business processing risk, thereby solving the technical problems that the related technology cannot quickly guide and assist customers in handling business, and cannot predict the business processing risk in real time, resulting in low business processing efficiency and high risk.

[0182] Example 3

[0183] According to an embodiment of the present application, a non-volatile storage medium is also provided, in which a program is stored, wherein when the program is running, the device where the non-volatile storage medium is located is controlled to execute the business management method based on the intelligent knowledge graph in Example 2.

[0184] Optionally, the device where the non-volatile storage medium is located implements the following steps by running the program: obtaining first customer data of a first customer within the business area of ​​the enterprise; analyzing the first customer data using a pre-trained grid management model to determine a first push grid corresponding to the first customer, and matching the first push grid with a preset grid management table to determine a first push strategy and a first account manager corresponding to the first customer, wherein the grid management table records the correspondence between each push grid obtained by dividing the business area of ​​the enterprise and the account manager and the push strategy; when the first customer initiates a business processing request according to the first push strategy, obtaining first business processing process data in the process of the first account manager handling the business processing request; analyzing the first business processing process data using a pre-trained risk classification model to obtain a corresponding risk category label, and executing corresponding risk management decisions on the first business processing process data according to the risk category label.

[0185] According to an embodiment of the present application, a computer program product is also provided, which includes a stored computer program, wherein when the computer program is executed by a processor, the business management method based on the intelligent knowledge graph in Example 2 is implemented.

[0186] Optionally, the computer program executes the following steps: obtaining first customer data of a first customer within the business area of ​​the enterprise; analyzing the first customer data using a pre-trained grid management model to determine a first push grid corresponding to the first customer, and matching the first push grid with a preset grid management table to determine a first push strategy and a first account manager corresponding to the first customer, wherein the grid management table records the correspondence between each push grid obtained by dividing the business area of ​​the enterprise and the account manager and the push strategy; when the first customer initiates a business processing request according to the first push strategy, obtaining first business processing process data in the process of the first account manager handling the business processing request; analyzing the first business processing process data using a pre-trained risk classification model to obtain a corresponding risk category label, and executing corresponding risk management decisions on the first business processing process data according to the risk category label.

[0187] According to an embodiment of the present application, a processor is also provided, which is used to run a program, wherein the business management method based on the intelligent knowledge graph in Example 2 is executed when the program is running.

[0188] Optionally, when the program is running, the following steps are executed: obtaining first customer data of a first customer within the business area of ​​the enterprise; analyzing the first customer data using a pre-trained grid management model to determine a first push grid corresponding to the first customer, and matching the first push grid with a preset grid management table to determine a first push strategy and a first account manager corresponding to the first customer, wherein the grid management table records the correspondence between each push grid obtained by dividing the business area of ​​the enterprise and the account manager and the push strategy; when the first customer initiates a business processing request according to the first push strategy, obtaining first business processing process data in the process of the first account manager handling the business processing request; analyzing the first business processing process data using a pre-trained risk classification model to obtain a corresponding risk category label, and executing corresponding risk management decisions on the first business processing process data according to the risk category label.

[0189] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0190] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0191] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed over multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0192] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0193] If the integrated unit is implemented in the form of 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 application, or the part that contributes to the relevant technology or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk and other media that can store program codes.

[0194] The above are only preferred implementations of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A business management method based on intelligent knowledge graph, characterized in that: include: Obtain the first customer data of the first customer in the enterprise's operating area; Analyze the first customer data using a pre-trained grid management model to determine a first push grid corresponding to the first customer, and match the first push grid with a preset grid management table to determine a first push strategy and a first account manager corresponding to the first customer, wherein the grid management table records the correspondence between each push grid obtained by dividing the business area of ​​the enterprise, the account manager, and the push strategy; When the first customer initiates a business processing request according to the first push strategy, obtaining first business processing process data during the process of the first customer manager processing the business processing request; Analyze the first business handling process data using a pre-trained risk classification model to obtain a corresponding risk category label, and execute a corresponding risk management decision on the first business handling process data according to the risk category label; The construction process of the grid management table includes: For each push grid within the business area of ​​the enterprise, a grid portrait of the push grid is constructed according to at least one key feature related to the business needs of each second customer in the push grid, and the grid portrait is analyzed to divide the multiple second customers in the push grid into multiple customer groups, and the propensity scores of each customer group in the push grid for different push strategies are calculated according to the following formula, and the push strategy with the propensity score not less than a preset score threshold is used as the push strategy of the push grid: ; in, represents the propensity score of customer group x for the push strategy containing the yth push strategy feature, represents the evaluation score of customer group x on the zth push strategy, represents the push strategy containing the yth push strategy feature and the zth push strategy, represents the number of times the push strategy containing the rth push strategy feature is responded to by the customer group, Z represents the total number of push strategies, and R represents the total number of push strategy features; Acquire task data of a plurality of push tasks within the business area of ​​the enterprise, wherein the task data at least includes: account information of a customer manager assigned to the push task; According to the second customer data of each second customer and the task data of each push task, respectively determine the mapping relationship between the account information of each customer manager and each push task, and the corresponding relationship between each second customer and each push task; Analyzing the mapping relationship and the corresponding relationship using a second neural network model to determine the customer allocation information corresponding to the account information of each of the customer managers, and determining at least one push grid served by the customer manager according to the customer allocation information; The grid management table is constructed by the push strategies corresponding to the plurality of push grids and the customer managers.

2. The method according to claim 1, characterized in that The training process of the grid management model includes: Acquire multiple groups of first training samples, wherein each group of the first training samples includes: second customer data of a second customer in the business area of ​​the enterprise and a second push grid corresponding to the second customer; Constructing the first neural network model; For each group of the training samples, inputting the second customer data in the training samples into the first neural network model to determine the prediction push grid corresponding to the second customer; A first target loss function is determined based on the second push grids and the predicted push grids in multiple groups of the first training samples, and the model parameters of the first neural network model are adjusted by the first target loss function until the model converges to obtain the grid management model.

3. The method according to claim 2, characterized in that Obtain multiple sets of first training samples, including: Acquire initial customer data of multiple second customers in the business area of ​​the enterprise, and pre-process the initial customer data to obtain second customer data of each second customer, wherein the pre-processing includes at least one of the following: data cleaning, data sorting, and data standardization, and the second customer data includes at least one of the following: basic data, financial data, consumer behavior data, and business processing data; Using a clustering algorithm to extract features from the second customer data of each of the second customers, obtaining at least one key feature related to the business needs of each of the second customers, wherein the type of the key feature includes at least one of the following: consumption behavior features, loan behavior features, and credit business handling features; According to at least one key feature of each of the second customers, each of the second customers is divided into a corresponding second push grid using a grid management method; A plurality of groups of the first training samples are formed by the second customer data of a plurality of the second customers and the second push grids corresponding to the respective second customers.

4. The method according to claim 1, characterized in that: Before determining at least one push grid served by the account manager according to the account allocation information, the method further includes: Obtaining historical business handling record information of each second customer in the push grid served by the customer manager, wherein the historical business handling record information includes at least one of the following: feedback and evaluation of the second customer on the business handled by the customer manager, and the timeliness of the business handling; The allocation influencing factors are determined according to the historical business handling records, and the customer allocation information is adjusted according to the allocation influencing factors.

5. The method according to claim 1, characterized in that The training process of the risk classification model includes: Acquire multiple groups of second training samples and multiple groups of third training samples, wherein the second training samples include: second business handling process data and first category labels corresponding to the second business handling process data, and the third training samples include: similar business handling process data having a data similarity with the second business handling process data higher than a preset threshold and second category labels corresponding to the similar business handling process data; Constructing a third neural network model; For each group of the second training samples, analyzing the second business handling process data using the third neural network model to obtain a corresponding first prediction category label; For each group of the third training samples, the similar business handling process data is analyzed using the third neural network model to obtain a corresponding second prediction category label; Constructing a multi-layer adaptive loss function based on multiple groups of first category labels in the second training samples and the corresponding first predicted category labels and multiple groups of second category labels in the third training samples and the corresponding second predicted category labels, and constructing a classification loss function based on the first category labels in each group of the second training samples and the corresponding first predicted category labels; A second objective loss function is determined based on the multi-layer adaptive loss function and the classification loss function, and model parameters of the third neural network model are updated based on the second objective loss function until the model converges to obtain the risk classification model.

6. A management decision-making platform, characterized in that: The management decision platform is used to execute the business management method based on intelligent knowledge graph described in any one of claims 1 to 5 above, and the management decision platform includes: a push management system, a business management system, and a risk control management system, wherein: The push management system is used to obtain first customer data of a first customer in the business area of ​​the enterprise; analyze the first customer data using a pre-trained grid management model to determine a first push grid corresponding to the first customer, and match the first push grid with a preset grid management table to determine a first push strategy and a first account manager corresponding to the first customer, wherein the grid management table records the correspondence between each push grid obtained by dividing the business area of ​​the enterprise, the account manager, and the push strategy; The business management system is used to obtain first business handling process data during the process of the first account manager processing the business handling request when the first customer initiates the business handling request according to the first push strategy; The risk control management system is used to analyze the first business processing process data using a pre-trained risk classification model to obtain corresponding risk category labels, and execute corresponding risk management decisions on the first business processing process data according to the risk category labels.

7. The management decision platform according to claim 6, characterized in that: The management decision platform also includes: a resource database, wherein: The resource database is used to periodically obtain multiple types of resource information and store the resource information according to the types of each type of resource information, wherein the types of resource information include but are not limited to: product introduction information, training and learning information, operation service information, and corporate culture information.

8. A computer program product comprising computer instructions, characterized in that: When the computer instructions are executed by the processor, the business management method based on the intelligent knowledge graph described in any one of claims 1 to 5 is implemented.

Images