Marketing decision-making method and system based on intelligent agent and text technology

Abstract

The invention discloses a marketing decision-making method and system based on an intelligent agent and a text technology. The method comprises the following steps: acquiring a first financial service consultation text of a first customer and first financial service handling information of the first customer in a marketing area; analyzing the first financial service consultation text by using a financial service recommendation model to obtain a to-be-recommended first financial service, determining a first marketing grid to which the first customer belongs according to the first financial service handling information, and determining a first grid portrait of the first marketing grid; analyzing the first grid portrait and the business information of the first financial business by using a marketing strategy recommendation model to obtain a corresponding first marketing strategy; and recommending the first financial service to the first customer according to the first marketing strategy. According to the method and the device, the technical problems of low recommendation efficiency and relatively poor accuracy caused by the fact that financial businesses are recommended only through subjective judgment and historical experience of marketing management personnel in related technologies are solved.

Description

Technical Field

[0001] The present application relates to the field of financial management technology, and more specifically, to a marketing decision-making method and system based on intelligent agent and text technology. Background Art

[0002] As an important provider of rural financial services, township banks have long relied on traditional methods of manual judgment and experience accumulation to recommend financial services when serving different customer groups. The core of this model is that bank account managers use their personal experience to assess customer needs and risks based on their in-depth understanding of the customer's basic information, financial status and credit history, and thus decide which type of financial service is most suitable for a specific customer group. However, with the expansion of business scale and the diversification of customer needs, the traditional manual recommendation method has gradually exposed its limitations, which are specifically reflected in the following aspects:

[0003] (1) In the daily operations of township banks, the amount of customer data is huge, including basic customer information, financial statements, historical transaction records, credit assessment reports, etc. The ability to manually process this data is limited, and it is difficult to accurately analyze the characteristics and needs of customer groups in a short period of time. In particular, the lack of application of big data analysis and complex statistical models has led to insufficient accuracy in business matching and an inability to effectively tap into the value of potential customers.

[0004] (2) The personal experience and subjective judgment of the account manager dominate the traditional recommendation process. Although this can provide services based on the personalized needs of customers to a certain extent, it also brings inconsistency and uncertainty to the recommendation results.

[0005] (3) Manual processing usually involves multiple links, such as customer information collection, risk assessment, solution design and business recommendation. Each link requires the account manager to invest a lot of time and energy, resulting in slow service response and difficulty in adapting to the rapidly changing market environment and customer needs. Especially during peak hours, low service efficiency will seriously affect customer satisfaction and the bank's market competitiveness.

[0006] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0007] The embodiments of the present application provide a marketing decision-making method and system based on intelligent agents and text technology, so as to at least solve the technical problem that the related technology recommends financial services only through the subjective judgment and historical experience of marketing managers, resulting in low recommendation efficiency and poor accuracy.

[0008] According to one aspect of an embodiment of the present application, a marketing decision-making method based on intelligent agent and text technology is provided, including: obtaining a first financial business consultation text input by a first customer, and first financial business handling information of the first customer in a preset marketing area, wherein the first financial business consultation text at least includes: first customer information and first business demand information of the first customer, and the first financial business handling information is at least used to reflect the number of times the first customer has handled financial business at all financial business handling points in the marketing area; using a pre-trained financial business recommendation model to perform text analysis on the first financial business consultation text to obtain a first financial business to be recommended to the first customer, and determining a first marketing grid to which the first customer belongs based on the first financial business handling information, and determining a first grid portrait of the first marketing grid; using a pre-trained marketing strategy recommendation model to analyze the first grid portrait and the business information of the first financial business to obtain a first marketing strategy for recommending the first financial business to the first customer; and recommending the first financial business to the first customer based on the first marketing strategy.

[0009] Optionally, the training process of the financial service recommendation model includes: constructing a first neural network model; obtaining multiple groups of first training samples, wherein each group of first training samples includes: a second financial service consultation text input by the second customer and a second financial service selected by the second customer from multiple financial services selected from the second customer based on the second financial service consultation text; for each group of first training samples, inputting the second financial service consultation text in the first training sample into the first neural network model to obtain the predicted financial service output by the first neural network model; constructing a first objective loss function of the first neural network model based on the second financial service and the predicted financial service in each group of first training samples, and optimizing the first objective loss function until a first preset convergence condition is met, thereby obtaining a trained financial service recommendation model.

[0010] Optionally, obtaining multiple sets of first training samples includes: obtaining second financial business consultation texts respectively input by multiple second customers, wherein the second financial business consultation texts include at least: second customer information and second business demand information of the second customer, and the second customer information includes at least one of the following: account balance, deposit amount, income information, financial management information, consumption information, and credit information; obtaining from the financial institution a similar customer set of each second customer and a first financial business set associated with the second business demand information in a financial business set that has been handled by each similar customer in the similar customer set; for each second customer, determining, based on the similar customer set and the first financial business set, and according to the following formula, the inclination of the second customer to select each financial business in the first financial business set, including: , where represents the inclination of the second customer u to select financial service i in the first financial service set, represents the weight coefficient, represents the first inclination of the second customer u to choose financial service i, and The expression is as follows: , where represents the set of similar customers of the second customer u, It represents the average of the inclinations of all customers in a financial institution to choose to handle all financial services sold by the financial institution. represents the bias value of the second customer u, represents the bias value of financial business i, represents the inclination of similar customer v in the similar customer set to choose financial service i, It represents the average value of the tendency of similar customers v to choose financial institutions to sell all financial services. represents the similarity between the second customer u and the similar customer v in the similar customer set, and The expression is: , where I represents the first financial business set, represents the average value of the second customer u's tendency to choose a financial institution to sell all financial services, represents the inclination of the second customer u to choose financial service i, represents the second inclination of the second customer u to choose financial service i, and The expression is: , where represents the second financial business set that is most similar to financial business i, represents the deviation of the second customer u to financial service i, represents the similarity between financial service i and financial service j in the second financial service set, and The expression is: , where represents the propensity of similar customer v to choose financial service i, represents the inclination of similar customer v to choose financial service j, represents the inverse customer frequency of financial service j, and The expression is: ,middle, Represents the total number of customers who have a tendency to choose financial service j in the similar customer set; feeds back the financial services with a tendency higher than a preset threshold value to the second customer, and receives the second customer's financial service selection instruction, wherein the financial service selection instruction carries at least the second financial service selected by the second customer from the financial services with a tendency higher than the preset threshold value; the second financial service consultation texts input by each second customer and the second financial service selected by each second customer constitute multiple groups of first training samples.

[0011] Optionally, determining the first marketing grid to which the first customer belongs based on the first financial business processing information includes: determining the number of times the first customer has handled financial business at each financial business processing point within the marketing area based on the first financial business processing information, and deleting the financial business processing points whose number of processing times is lower than a first number threshold from all the financial business processing points within the marketing area to obtain a set of financial business processing points; taking the financial business processing point at which the first customer has handled the most times in the set of financial business processing points as the first financial business processing point, and taking the financial business processing points within the marketing area whose distance from the first financial business processing point is less than a preset distance threshold as the first clustering set; taking the financial business processing point at which the first customer has handled the most times among other financial business processing points within the marketing area outside the first clustering set as the second financial business processing point, and taking the financial business processing point within the marketing area whose distance from the first financial business processing point is less than the distance threshold and is within the first clustering set as the second cluster set; repeat the above steps until all financial business handling points in the financial business handling point set are clustered into each cluster set; determine the total number of times the first customer has handled business at all financial business handling points in each cluster set, and determine the target cluster set whose total number is less than the second number threshold, and use the position coordinates of the target financial business handling point in the target cluster set as the position coordinates of the target cluster set, wherein the target financial business handling point is the financial business handling point with the largest number of handling times in the target cluster set; determine the first customer's frequently handled business area according to the position coordinates of the target cluster set, and determine the marketing grid where the frequently handled business area is located according to a preset marketing area grid division table, and use the marketing grid as the first marketing grid to which the first customer belongs, wherein the marketing area grid division table includes multiple marketing grids obtained by dividing the marketing area according to a preset grid area division rule.

[0012] Optionally, the marketing grid in which the business area is usually handled is determined according to a preset marketing area grid division table, including: determining the regional location information of each marketing grid in the marketing area; using a spatial index algorithm to analyze the regional location information of each marketing grid and the regional location information of the business area usually handled, to determine the marketing grid in which the business area is usually handled.

[0013] Optionally, determining a first grid portrait of the first marketing grid includes: obtaining business demand information and customer information of all third customers in the first marketing grid, and preprocessing the business demand information and customer information of each third customer; for each third customer in the first marketing grid, extracting features of the third customer's customer information to obtain all key features related to the business demand information of the third customer, wherein the types of key features include at least one of the following: consumption behavior features, loan behavior features, and credit business processing features; constructing a first grid portrait of the first marketing grid based on all the key features of each third customer.

[0014] Optionally, the training process of the marketing strategy recommendation model includes: constructing a second neural network model; obtaining multiple groups of second training samples, wherein each group of second training samples includes: a second grid portrait of the second marketing grid to which the second customer belongs and business information of the second amount financial business to be recommended to the second customer, a second marketing strategy for recommending the second financial business to the second customer, and the second marketing strategy is a marketing strategy with the highest customer conversion rate, the lowest strategy execution cost, and the highest customer satisfaction; for each group of second training samples, the second grid portrait and the business information of the second amount financial business in the second training sample are input into the second neural network model to obtain a predicted marketing strategy for recommending the second financial business to the second customer output by the second neural network model; a second objective loss function of the second neural network model is constructed based on the second marketing strategies and predicted marketing strategies in each group of second training samples, and the second objective loss function is optimized until a preset second convergence condition is met to obtain a marketing strategy recommendation model that has completed training.

[0015] Optionally, recommending a first financial service to a first customer based on a first marketing strategy includes: determining a first marketing manager corresponding to the first marketing strategy based on a correspondence between a preset marketing strategy management table and the first marketing strategy, wherein the marketing strategy management table contains correspondence between multiple marketing strategies and marketing managers; and instructing the first marketing manager to recommend the first financial service to the first customer according to the first marketing strategy.

[0016] According to another aspect of an embodiment of the present application, a marketing decision system based on intelligent agent and text technology is also provided, including: an acquisition module, used to acquire a first financial business consultation text input by a first customer, and first financial business handling information of the first customer in a preset marketing area, wherein the first financial business consultation text at least includes: first customer information and first business demand information of the first customer, and the first financial business handling information is at least used to reflect the number of times the first customer has handled financial business at all financial business handling points in the marketing area; a first determination module, used to perform text analysis on the first financial business consultation text using a pre-trained financial business recommendation model, obtain the first financial business to be recommended to the first customer, and determine the first marketing grid to which the first customer belongs based on the first financial business handling information, and determine the first grid portrait of the first marketing grid; a second determination module, used to analyze the first grid portrait and the business information of the first financial business using a pre-trained marketing strategy recommendation model, and obtain a first marketing strategy for recommending the first financial business to the first customer; a recommendation module, used to recommend the first financial business to the first customer based on the first marketing strategy.

[0017] According to another aspect of an embodiment of the present application, a computer program product is also provided, the computer program product comprising: a computer program, wherein when the computer program is executed by a processor, the above-mentioned marketing decision-making method based on intelligent agent and text technology is implemented.

[0018] In the embodiment of the present application, the artificial intelligence system uses deep learning technology to perform text analysis on the first financial business consultation text input by the customer, accurately identifies the customer's needs and recommends corresponding financial products, which greatly improves the efficiency and personalization of the service compared to manual consultation and recommendation methods; at the same time, the artificial intelligence system also uses big data analysis technology to analyze the first financial business handling information to determine the first marketing grid to which the first customer belongs, determine the first grid portrait of the first marketing grid, and combine the business information of the first financial business to be recommended to the first customer to generate a first marketing strategy that matches the customer and the grid portrait. This strategy not only takes into account the customer's needs and preferences, but also fully combines the resources within the grid, achieving the technical effect of intelligent financial business and marketing strategy recommendation, and achieving the purpose of improving the customer service efficiency and precision marketing of village banks. This solves the technical problem that the relevant technology recommends financial business only through the subjective judgment and historical experience of marketing managers, resulting in low recommendation efficiency and poor accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] 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:

[0020] Figure 1 is a flowchart of an optional marketing decision-making method based on intelligent agent and text technology according to an embodiment of the present application;

[0021] Figure 2 is a schematic diagram of the structure of an optional marketing decision-making system based on agent and text technology according to an embodiment of the present application;

[0022] Figure 3 is a schematic diagram of the structure of an optional marketing decision-making system according to an embodiment of the present application;

[0023] Figure 4 is a schematic diagram of the structure of an optional data storage module according to an embodiment of the present application;

[0024] Figure 5 It is a schematic diagram of the structure of an optional electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0025] 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.

[0026] It should be noted that the terms "first", "second", etc. in the specification, claims and drawings of the present application 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, service 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, services or devices.

[0027] Example 1

[0028] According to an embodiment of the present application, a marketing decision-making method based on intelligent agent and text technology is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0029] Figure 1 is a flowchart of a marketing decision-making method based on an agent and text technology provided according to an embodiment of the present application, such as Figure 1 As shown, the method includes the following steps S102-S108:

[0030] Step S102, obtaining a first financial service consultation text input by the first customer and first financial service handling information of the first customer in a preset marketing area.

[0031] In the technical solution provided in the above step S102, the above first financial business consultation text refers to the text information submitted by the first customer when consulting financial business using the online service platform of the financial institution (such as a village bank), such as a mobile banking application, WeChat applet or the official website of the bank. This includes but is not limited to the first customer's specific questions about a certain type of loan, deposit or financial management business, or consultation on details such as the financial business handling process, required documents, interest rates, and terms. Therefore, the above first financial business consultation text at least includes: the first customer information and the first business demand information of the first customer.

[0032] In addition, the first financial business handling information of the first customer in the marketing area refers to the data records generated during the application, handling and subsequent management of financial business through online channels or offline outlets within the specific marketing area designated by the financial institution (such as village banks), that is, the first financial business handling information is at least used to reflect the number of times the first customer handles financial business at all financial business handling points in the marketing area. This includes but is not limited to the customer's basic information, financial status, credit score, business application progress, contract signing details, loan records, repayment behavior, and interaction records with bank staff. These data and information are of great value to financial institutions in assessing customer risks, providing personalized services, monitoring business processes and conducting post-loan management, and are the key data source for financial institutions to achieve precision marketing and efficient business management.

[0033] Step S104, using the pre-trained financial service recommendation model to perform text analysis on the first financial service consultation text, obtain the first financial service to be recommended to the first customer, and determine the first marketing grid to which the first customer belongs based on the first financial service processing information, and determine the first grid portrait of the first marketing grid.

[0034] In the technical solution provided in the above step S104, the above financial service recommendation model is trained based on learning a large amount of historical customer data and financial service data, and can identify the correlation between customer characteristics and financial services. The artificial intelligence system calls the pre-trained financial service recommendation model to understand and extract key information in the first financial service consultation text, such as the customer's needs, preferences and financial status, so as to recommend the first financial service that best meets the first customer's current needs and his basic information.

[0035] At the same time, the artificial intelligence system determines the first marketing grid to which the first customer belongs by analyzing the first financial business handling information. Among them, the artificial intelligence system can divide the marketing area according to the geographical location of each financial business handling point (such as multiple branch business halls of a village bank in area A), obtain multiple sub-areas, and divide each customer into different sub-areas according to the financial business handling information of each customer, and obtain multiple marketing grids. Therefore, the system can determine the first marketing grid to which the first customer belongs based on the first financial business handling information of the first customer, and determine the first grid portrait of the first marketing grid, wherein the first grid portrait can be comprehensively determined by the customer information characteristics and business demand conditions of all customers in the marketing grid.

[0036] That is to say, the technical solution provided in the above step S104 realizes the identification of customer personalized needs and accurate business recommendations through text analysis and data mining technology, combined with pre-trained models and marketing grid division strategies, while ensuring that the marketing activities of financial institutions can effectively cover the target customer groups, thereby improving marketing efficiency and customer satisfaction.

[0037] Step S106, using the pre-trained marketing strategy recommendation model to analyze the first grid portrait and the business information of the first financial business, to obtain a first marketing strategy for recommending the first financial business to the first customer.

[0038] In the technical solution provided in the above step S106, the above marketing strategy recommendation model learns and understands the correlation between different marketing grid portraits, various financial product business information and different marketing strategies through the combination of deep learning model and marketing strategy algorithm. This correlation may be based on multi-dimensional data such as historical customer behavior, financial product preferences, market trends and the effects of past marketing activities.

[0039] That is to say, when determining the overall marketing strategy, the marketing strategy recommendation model fully considers the comprehensive characteristics of the first grid portrait, such as customer age distribution, average income level, credit score distribution, and business information of the first financial business to be recommended to the first customer. By analyzing this information, the marketing strategy recommendation model can formulate a first marketing strategy that not only meets the financial business marketing needs of the customer groups of the marketing grid, but also meets the personalized needs of a single customer, so that it can take into account both personalized needs and group needs, thereby maximizing the acceptance and conversion rate of business recommendations, while improving customer satisfaction and the business efficiency of financial institutions.

[0040] Step S108: recommending a first financial service to a first customer according to a first marketing strategy.

[0041] In the technical solution provided in the above step S108, after the artificial intelligence system obtains the first financial service and the first marketing strategy to be recommended to the first customer respectively through the above steps S104 and S106, it can recommend the first financial service to the first customer based on the first marketing strategy, thereby meeting the customer's personalized needs and ensuring customer experience and satisfaction.

[0042] Based on the scheme defined by the above steps S102 to S108, it can be known that in the embodiment of the present application, the artificial intelligence system uses deep learning technology to perform text analysis on the first financial business consultation text input by the customer, accurately identifies the customer's needs and recommends corresponding financial products. Compared with manual consultation and recommendation methods, this greatly improves the efficiency and personalization level of the service; at the same time, the artificial intelligence system also uses big data analysis technology to analyze the first financial business handling information to determine the first marketing grid to which the first customer belongs, determine the first grid portrait of the first marketing grid, and combine the business information of the first financial business to be recommended to the first customer to generate a first marketing strategy that matches the customer and the grid portrait. This strategy not only takes into account the customer's needs and preferences, but also fully combines the resources within the grid, realizes the technical effect of intelligent financial business and marketing strategy recommendation, and achieves the purpose of improving the customer service efficiency and precision marketing of village banks.

[0043] The following explains the various steps of the marketing decision-making method based on intelligent agents and text technology in combination with the specific implementation process.

[0044] Optionally, the training process of the financial service recommendation model may include the following steps S11-S14, wherein:

[0045] Step S11, constructing a first neural network model.

[0046] Among them, the model architecture of the first neural network model can be a convolutional neural network, Transformer, etc., or it can be a fusion model composed of multiple models.

[0047] Step S12, obtaining multiple groups of first training samples.

[0048] Among them, multiple groups of first training samples are all from historical data, and each group of first training samples includes: the second financial business consultation text input by the second customer and the second financial business selected by the second customer from among multiple financial businesses selected by the second customer based on the second financial business consultation text. These sample data provide a basis for model learning, enabling it to understand the relationship between business demand information and customer information in the consultation text and the recommended financial business.

[0049] Step S13, for each group of first training samples, the second financial business consultation text in the first training sample is input into the first neural network model to obtain the predicted financial business output by the first neural network model.

[0050] Step S14, constructing the first target loss function of the first neural network model according to the second financial business and the predicted financial business in each group of first training samples (that is, the first target loss function quantifies the difference between the predicted result and the actual result), and optimizing the first target loss function (preset optimization algorithm, such as gradient descent algorithm and its variants, etc.) until the first preset convergence condition is met, to obtain a trained financial business recommendation model, wherein the above-mentioned first preset convergence condition may be that the value of the loss function is lower than a certain threshold or the change after continuous iterations is less than a certain threshold, so as to indicate that the model training has reached a stable state.

[0051] Specifically, the above-mentioned financial service recommendation model establishes a mapping relationship between text features and service recommendations by learning a large number of historical customers' consultation texts and service selection behaviors, so that when new customers consult, the most suitable financial services can be recommended quickly and accurately, greatly improving customer experience and the business efficiency of financial institutions.

[0052] Optionally, in the technical solution provided in the above step S12, multiple groups of training samples may be obtained in the following manners, including:

[0053] Step S121, obtaining second financial business consultation texts respectively input by multiple second customers, wherein the second financial business consultation texts at least include: second customer information and second business demand information of the second customer, and the second customer information includes at least one of the following: account balance, deposit amount, income information, financial information, consumption information, and credit information.

[0054] Step S122, obtaining from the financial service processing point a similar customer set of each second customer and a first financial service set associated with the second service demand information in a financial service set that has been processed by each similar customer in the similar customer set. Among them, the similar customer set of each second customer can be determined by a similarity calculation method to determine a customer set with similar needs, preferences and financial status to the second customer. The first financial service set can be understood as a service set related to the demand that is screened out from a financial service set that has been processed by each similar customer in its similar customer set based on the second service demand information of the second customer.

[0055] Step S123, for each second customer, the inclination of the second customer to select each financial service in the first financial service set is determined according to the similar customer set and the first financial service set and the following formula, including:

[0056]

[0057] in, represents the inclination of the second customer u to select financial service i in the first financial service set, represents the weight coefficient, represents the first inclination of the second customer u to choose financial service i, and The expression is as follows:

[0058] ,

[0059] In the formula, represents the set of similar customers of the second customer u, It represents the average propensity of all customers in a financial institution to choose to handle all financial services sold by the financial institution (i.e. the global average of all propensities). represents the bias value of the second customer u (referring to the average value of the tendency of the second customer u to choose to handle all financial services sold by financial institutions and The difference between ,in, , indicating that the second customer tends to give a higher tendency; , indicating that the second customer tends to give a lower tendency; , indicating that the inclination given by the second customer is comparable to the average level). Indicates the bias value of financial business i (referring to the average tendency of financial business i and The difference between ,in, , indicating that financial business i tends to obtain a higher tendency; , indicating that financial business i tends to obtain a lower tendency; , indicating that the propensity obtained by financial business i is equivalent to the average level). represents the inclination of similar customer v in the similar customer set to choose financial service i, It represents the average value of the tendency of similar customers v to choose financial institutions to sell all financial services. represents the similarity between the second customer u and the similar customer v in the similar customer set, and The expression is:

[0060] ,

[0061] Where I represents the first financial business set, represents the average value of the second customer u's tendency to choose a financial institution to sell all financial services, Represents the inclination of the second customer u to choose financial service i.

[0062] in addition, represents the second inclination of the second customer u to choose financial service i, and The expression is:

[0063] ,

[0064] In the formula, represents the second financial business set that is most similar to financial business i, represents the deviation of the second customer u towards financial service i (referring to the difference between the second customer u's inclination towards financial service i and the second customer u's overall rating behavior or the overall rating behavior of the service, that is, ,in, , indicating that the second customer u has a higher tendency to choose to handle financial service i; , indicating that the second customer u has a lower tendency to choose to handle financial service i; , indicating that the second customer u's inclination to choose financial service i is consistent with the inclination of other financial services). represents the similarity between financial service i and financial service j in the second financial service set, and The expression is:

[0065] ,

[0066] In the formula, represents the propensity of similar customer v to choose financial service i, represents the inclination of similar customer v to choose financial service j, represents the inverse customer frequency of financial business j, which is used to perform “soft punishment” on active customers (i.e. customers who have a tendency to use a large number of financial products) to reduce the impact of these data on The dominant influence of the calculation results is used to improve the diversity and fairness of the recommendation results, and the influence of customers with less inclination is amplified to ensure that the preferences of these customers can be captured to improve the accuracy of the recommendation results. The expression is:

[0067] ,

[0068] In the formula, Represents the total number of customers in the similar customer set who have a tendency to choose financial service j.

[0069] It should be noted that in the above calculation of similarity and , tendency and When considering the unknown customer preference for certain financial services, or Replace to fill the missing values. In the above formula of the embodiment of the present application, Fill missing values. In addition, when calculating the propensity By introducing customer bias and business bias , avoiding direct reliance on the customer's ratings of other services and solving the circular dependency problem. This ensures that even if some propensities are missing, the propensity of the second customer to each financial service in the first financial service set can be reasonably calculated.

[0070] Step S124, feeding back the financial services with a tendency higher than the preset threshold value to the second customer, and receiving the second customer's financial service selection instruction, wherein the financial service selection instruction carries at least a second financial service selected by the second customer from the financial services with a tendency higher than the preset threshold value.

[0071] In the technical solution provided in the above step S124, the artificial intelligence system filters financial services with a tendency lower than a predefined threshold value, thereby ensuring that the services fed back to the second customer are highly matched and have a high tendency; then, the system presents the filtered financial service information to the second customer in a user-friendly manner, and the customer can select the service of interest according to his or her own needs and issue a financial service selection instruction through the system. After receiving the customer's selection instruction, the system can quickly determine the second financial service selected by the customer and provide the customer with a personalized service experience, thereby realizing intelligent service recommendations and seamless customer experience.

[0072] Step S125, the second financial service consultation texts input by each second customer and the second financial service selected by each second customer form a plurality of groups of first training samples.

[0073] After collecting the first financial business consultation text submitted by the first customer when conducting financial business consultation, the artificial intelligence system can first call the pre-trained financial business recommendation model, understand and extract key information in the text through deep learning technology and natural language processing algorithms, such as customer needs, preferences and financial status, and analyze the key information to recommend the first financial business that best meets the first customer's current needs and his basic information.

[0074] At the same time, the artificial intelligence system can also use the first financial business handling information of the first customer to determine the first customer's financial business handling through data analysis, and then locate it in the corresponding marketing grid.

[0075] As an optional implementation, in the technical solution provided in the above step S104, the system may determine the first marketing grid to which the first customer belongs according to the following method, including:

[0076] Step 1: Determine the number of times the first customer has handled financial business at each financial business handling point within the marketing area based on the first financial business handling information, and delete the financial business handling points whose number of handling times is lower than the first number threshold from all financial business handling points within the marketing area to obtain a set of financial business handling points (i.e., including all active financial business handling points).

[0077] Step 2: The financial business processing point where the first customer has handled the most transactions in the set of financial business processing points is taken as the first financial business processing point, and the financial business processing points within the marketing area whose distance from the first financial business processing point is less than a preset distance threshold are taken as the first cluster set; the financial business processing point where the first customer has handled the most transactions among other financial business processing points outside the first cluster set within the marketing area is taken as the second financial business processing point, and the financial business processing points within the marketing area whose distance from the first financial business processing point is less than the distance threshold and outside the first cluster set are taken as the second cluster set; repeat the above steps until all financial business processing points in the set of financial business processing points are clustered into respective cluster sets.

[0078] Step 3: Determine the total number of times the first customer has handled business at all financial business handling points in each cluster set, and determine the target cluster set whose total number is less than the second number threshold, and use the location coordinates of the target financial business handling point in the target cluster set as the location coordinates of the target cluster set, where the target financial business handling point is the financial business handling point with the most handling times in the target cluster set.

[0079] Step 4: Determine the first customer's frequently handled business area based on the location coordinates of the target cluster set, and determine the marketing grid where the frequently handled business area is located according to the preset marketing area grid division table, and use the marketing grid as the first marketing grid to which the first customer belongs. The marketing area grid division table includes multiple marketing grids obtained by dividing the marketing area according to the preset grid area division rules. This step makes full use of geographic space information, ensures the accuracy of marketing grid division and the rationality of customer positioning, and provides strong support for subsequent personalized marketing and services.

[0080] Through the first to fourth steps mentioned above, intelligent geographic positioning and marketing grid division based on customer behavior data (i.e., financial business processing information) are realized, which not only helps financial institutions understand customer business preferences and processing habits, but also promotes the rational allocation of resources and the effective implementation of marketing strategies, and improves the accuracy and efficiency of customer service.

[0081] Optionally, in the technical solution provided in the fourth step above, the marketing grid where the business area is usually handled can be determined according to the following steps, including: first determining the regional location information (such as regional longitude and latitude coordinate information) of each marketing grid in the marketing area; then using a spatial index algorithm to analyze the regional location information of each marketing grid and the regional location information of the business area usually handled, and determining the marketing grid where the business area is usually handled. The spatial index algorithm can be an R-tree or a quadtree, which is not specifically limited here.

[0082] In addition to the above-listed implementation schemes for determining spatial relationships based on a spatial index algorithm, based on the basic concept of the present invention, those skilled in the art may also determine which marketing grid a frequently handled business area belongs to through other technical schemes. For example, those skilled in the art may make changes to the above-mentioned implementation schemes, which should also be within the scope of protection of the present invention.

[0083] Furthermore, after determining the first marketing grid to which the first customer belongs, the system may also determine the first grid profile of the first marketing grid according to the following method. Including:

[0084] First, the business demand information and customer information of all third customers in the first marketing grid are obtained, and the business demand information and customer information of each third customer are pre-processed;

[0085] Then, for each third customer in the first marketing grid, feature extraction is performed on the customer information of the third customer to obtain all key features related to the business demand information of the third customer, wherein the types of key features include at least one of the following: consumption behavior features, loan behavior features, and credit business handling features;

[0086] Finally, the first grid portrait of the first marketing grid is constructed based on all the key characteristics of each third customer.

[0087] In the above scheme, the system first pre-processes the business demand information and customer information of all third customers in the first marketing grid, wherein the pre-processing includes but is not limited to: data cleaning, data sorting, data standardization, etc.; then, the common features or representative features of all third customers are extracted, and at least one key feature related to the business demand information of each third customer is screened from the extracted common features or representative features, such as consumption behavior features, loan behavior features, credit business handling features, etc. It should be noted that in order to quickly extract each key feature, before clustering, the principal component analysis method can be used to reduce the dimension of the information to reduce the feature dimension.

[0088] Furthermore, the system may analyze the first grid portrait of the first marketing grid determined above and the business information of the first financial business to be recommended to the first customer through a marketing strategy recommendation model to determine a first marketing strategy for recommending the first financial business to the first customer.

[0089] Optionally, the training process of the marketing strategy recommendation model includes the following steps S21-S24, wherein:

[0090] Step S21, constructing a second neural network model. The model architecture of the second neural network model can also be a convolutional neural network, Transformer, etc., or a fusion model composed of multiple models.

[0091] Step S22, obtaining multiple groups of second training samples.

[0092] Each group of second training samples includes: a second grid portrait of the second marketing grid to which the second customer belongs, business information of the second amount financial business to be recommended to the second customer, and a second marketing strategy for recommending the second financial business to the second customer. It should be noted that the second marketing strategy can be a marketing strategy with the highest customer conversion rate, the lowest strategy execution cost, and the highest customer satisfaction determined based on historical experience.

[0093] Step S23, for each group of second training samples, the second grid portrait and the business information of the second amount financial business in the second training sample are input into the second neural network model to obtain the predictive marketing strategy for recommending the second financial business to the second customer output by the second neural network model.

[0094] Step S24, constructing the second objective loss function of the second neural network model based on the second marketing strategy and the predicted marketing strategy in each group of second training samples, and optimizing the second objective loss function (using a preset optimization algorithm, such as a gradient descent algorithm and its variants, etc.) until the preset second convergence condition is met, thereby obtaining a marketing strategy recommendation model that has completed training.

[0095] During the above model training process, through the learning of a large number of samples, it is possible to establish the relationship between marketing grid portraits, financial business information and effective marketing strategies, and ultimately achieve intelligent and personalized recommendations on marketing strategies for specific customers in the new grid, thereby improving the marketing efficiency and customer satisfaction of financial institutions.

[0096] Finally, the system can recommend the first financial service to the first customer according to the first marketing strategy determined above, thereby achieving personalized product recommendation.

[0097] In addition, in order to determine the professionalism and pertinence of the strategy execution, the system can also determine the first marketing manager corresponding to the first marketing strategy based on the correspondence between the preset marketing strategy management table and the first marketing strategy; and then instruct the first marketing manager to recommend the first financial service to the first customer in accordance with the first marketing strategy.

[0098] Among them, the above-mentioned marketing strategy management table contains the correspondence between multiple marketing strategies and marketing managers. And the above-mentioned correspondence can be identified by a pre-trained matching model. Specifically, the above-mentioned matching model can extract the characteristics of the association between marketing strategies and managers from historical marketing data through machine learning algorithms (such as decision trees, support vector machines, neural networks, etc.), including but not limited to the type of marketing strategy (such as deposit product promotion, loan product recommendation, etc.), the effect of strategy execution (such as conversion rate, customer satisfaction, etc.), and the personal attributes of managers (such as professional fields, past performance, customer group preferences, etc.) and behavioral characteristics (such as the frequency, method, effect, etc. of executing strategies). In other words, the above-mentioned matching model is trained with a large amount of historical data to identify and learn the effective association between marketing strategies and managers. During the training process, the model will continuously optimize parameters to improve the accuracy and effect of strategy matching. Usually, an objective function (such as maximizing the expected benefits of strategy execution) is set, and the model is adjusted based on this objective function until a satisfactory matching effect is achieved.

[0099] In summary, the above-mentioned marketing decision-making method based on intelligent agent and text technology has the following technical advantages over the existing marketing decision-making methods:

[0100] (1) Through deep learning technology, it is possible to automatically extract and integrate customer information and business demand information from financial business consultation texts, rather than relying solely on manual analysis, thereby improving the accuracy and efficiency of information processing. At the same time, by using pre-trained financial business recommendation models, intelligent analysis based on customer consultation texts can accurately identify customer needs and recommend appropriate financial products. Compared with existing technologies, this greatly improves the personalization and accuracy of recommendation results.

[0101] (2) Determine the marketing grid to which the customer belongs based on financial business information, and conduct in-depth customer positioning based on grid portraits. This is more dynamic and accurate than traditional static customer classification, and can better match customers with grid resources.

[0102] (3) Through the marketing strategy recommendation model, marketing strategies are automatically generated based on grid profiles and financial business information. Compared with existing manual or rule-driven strategies, this solution can respond to market changes more quickly and intelligently.

[0103] (4) This solution can intelligently match marketing strategies with the professional fields and past performance of marketing managers to ensure the effectiveness of marketing strategy execution, surpassing the traditional one-to-one fixed allocation model.

[0104] Example 2

[0105] According to the embodiment of the present application, a marketing decision system based on agent and text technology for implementing the method in embodiment 1 is also provided. Figure 2 As shown, the marketing decision system based on agent and text technology at least includes: an acquisition module 21, a first determination module 22, a second determination module 23 and a recommendation module 24, wherein:

[0106] The acquisition module 21 is used to acquire a first financial business consultation text input by the first customer, and first financial business handling information of the first customer in a preset marketing area, wherein the first financial business consultation text at least includes: first customer information and first business demand information of the first customer, and the first financial business handling information at least reflects the number of times the first customer has handled financial business at all financial business handling points in the marketing area;

[0107] A first determination module 22 is used to perform text analysis on the first financial service consultation text using a pre-trained financial service recommendation model to obtain a first financial service to be recommended to the first customer, and determine a first marketing grid to which the first customer belongs based on the first financial service handling information, and determine a first grid profile of the first marketing grid;

[0108] A second determination module 23 is used to analyze the first grid portrait and the business information of the first financial business by using a pre-trained marketing strategy recommendation model to obtain a first marketing strategy for recommending the first financial business to the first customer;

[0109] The recommendation module 24 is configured to recommend a first financial service to a first customer according to a first marketing strategy.

[0110] The following describes the functions of each module of the marketing decision-making system based on intelligent agents and text technology in conjunction with the specific implementation process.

[0111] Optionally, the first determining module 22 may also train the financial service recommendation model according to the following process, including:

[0112] Step 1: Build the first neural network model;

[0113] Step 2: obtaining multiple groups of first training samples, wherein each group of first training samples includes: a second financial service consultation text input by the second customer and a second financial service selected by the second customer from among multiple financial services selected from the second customer based on the second financial service consultation text;

[0114] Step 3: For each group of first training samples, input the second financial business consultation text in the first training sample into the first neural network model to obtain the predicted financial business output by the first neural network model;

[0115] Step 4: Construct the first objective loss function of the first neural network model based on the second financial business and the predicted financial business in each group of first training samples, and optimize the first objective loss function until the first preset convergence condition is met to obtain a trained financial business recommendation model.

[0116] Specifically, in the technical solution provided in the second step above, the first determining module 22 may obtain multiple groups of first training samples according to the following method, including:

[0117] First, obtain second financial business consultation texts respectively input by multiple second customers, wherein the second financial business consultation texts at least include: second customer information and second business demand information of the second customer, and the second customer information includes at least one of the following: account balance, deposit amount, income information, financial management information, consumption information, and credit information.

[0118] Next, a similar customer set of each second customer and a first financial service set associated with the second service demand information in a financial service set handled by each similar customer in the similar customer set are obtained from within the financial institution.

[0119] For each second customer, the inclination of the second customer to select each financial service in the first financial service set is determined according to the similar customer set and the first financial service set in accordance with the following formula, including:

[0120] ,

[0121] In the formula, represents the inclination of the second customer u to select financial service i in the first financial service set, represents the weight coefficient, represents the first inclination of the second customer u to choose financial service i, and The expression is as follows:

[0122] ,

[0123] In the formula, represents the set of similar customers of the second customer u, It represents the average of the inclinations of all customers in a financial institution to choose to handle all financial services sold by the financial institution. represents the bias value of the second customer u, represents the bias value of financial business i, represents the inclination of similar customer v in the similar customer set to choose financial service i, It represents the average value of the tendency of similar customers v to choose financial institutions to sell all financial services. represents the similarity between the second customer u and the similar customer v in the similar customer set, and The expression is:

[0124] ,

[0125] Where I represents the first financial business set, represents the average value of the second customer u's tendency to choose a financial institution to sell all financial services, represents the inclination of the second customer u to choose financial service i, represents the second inclination of the second customer u to choose financial service i, and The expression is:

[0126] ,

[0127] In the formula, represents the second financial business set that is most similar to financial business i, represents the deviation of the second customer u to financial service i, represents the similarity between financial service i and financial service j in the second financial service set, and The expression is:

[0128] ,

[0129] In the formula, represents the propensity of similar customer v to choose financial service i, represents the inclination of similar customer v to choose financial service j, represents the inverse customer frequency of financial service j, and The expression is:

[0130] ,

[0131] In the formula, Represents the total number of customers in the similar customer set who have a tendency to choose financial service j.

[0132] Then, the financial services with a tendency higher than the preset threshold value are fed back to the second customer, and a financial service selection instruction of the second customer is received, wherein the financial service selection instruction carries at least a second financial service selected by the second customer from the financial services with a tendency higher than the preset threshold value.

[0133] Finally, the second financial service consultation texts input by each second customer and the second financial service selected by each second customer constitute multiple groups of first training samples.

[0134] Furthermore, the first determining module 22 may determine the first marketing grid to which the first customer belongs according to the following method, including:

[0135] Step 1: Determine the number of times the first customer has handled financial services at each financial service point in the marketing area based on the first financial service handling information, and delete the financial service points whose number of handling times is lower than the first number threshold from all financial service points in the marketing area to obtain a set of financial service points;

[0136] Step 2: The financial business handling point where the first customer has handled the most transactions in the financial business handling point set is taken as the first financial business handling point, and the financial business handling points within the marketing area whose distance from the first financial business handling point is less than a preset distance threshold are taken as the first cluster set; the financial business handling point where the first customer has handled the most transactions in other financial business handling points outside the first cluster set within the marketing area is taken as the second financial business handling point, and the financial business handling points within the marketing area whose distance from the first financial business handling point is less than the distance threshold and outside the first cluster set are taken as the second cluster set; the above steps are repeated until all financial business handling points in the financial business handling point set are clustered into respective cluster sets;

[0137] Step 3: Determine the total number of times the first customer has handled financial business at all financial business handling points in each cluster set, and determine a target cluster set whose total number is less than the second number threshold, and use the position coordinates of the target financial business handling point in the target cluster set as the position coordinates of the target cluster set, wherein the target financial business handling point is the financial business handling point with the largest number of handling times in the target cluster set;

[0138] Step 4: Determine the first customer's frequently handled business area based on the location coordinates of the target cluster set, and determine the marketing grid where the frequently handled business area is located according to the preset marketing area grid division table, and use the marketing grid as the first marketing grid to which the first customer belongs, wherein the marketing area grid division table includes multiple marketing grids obtained by dividing the marketing area according to the preset grid area division rules.

[0139] Specifically, in the fourth step above, the first determination module 22 may first determine the regional location information of each marketing grid within the marketing area; and then use the spatial index algorithm to analyze the regional location information of each marketing grid and the regional location information of the frequently handled business area to determine the marketing grid where the frequently handled business area is located.

[0140] Furthermore, the first determining module 22 may determine the first grid portrait of the first marketing grid by the following method, including:

[0141] First, the business demand information and customer information of all third customers in the first marketing grid are obtained, and the business demand information and customer information of each third customer are pre-processed;

[0142] Next, for each third customer in the first marketing grid, feature extraction is performed on the customer information of the third customer to obtain all key features related to the business demand information of the third customer, wherein the types of key features include at least one of the following: consumption behavior features, loan behavior features, and credit business handling features;

[0143] Then, the first grid portrait of the first marketing grid is constructed based on all the key characteristics of each third customer.

[0144] Optionally, the second determination module 23 may also train the marketing strategy recommendation model according to the following method, including:

[0145] Step 1: Build the second neural network model;

[0146] Step 2: Acquire multiple groups of second training samples, wherein each group of second training samples includes: a second grid portrait of the second marketing grid to which the second customer belongs, business information of the second amount financial business to be recommended to the second customer, and a second marketing strategy for recommending the second financial business to the second customer, and the second marketing strategy is a marketing strategy with the highest customer conversion rate, the lowest strategy execution cost, and the highest customer satisfaction;

[0147] Step 3: For each group of second training samples, the second grid portrait and the business information of the second amount financial business in the second training sample are input into the second neural network model to obtain the predictive marketing strategy for recommending the second financial business to the second customer output by the second neural network model;

[0148] Step 4: Construct the second objective loss function of the second neural network model according to the second marketing strategy and the predicted marketing strategy in each group of second training samples, and optimize the second objective loss function until the preset second convergence condition is met to obtain a marketing strategy recommendation model that has completed training.

[0149] Optionally, the recommendation module 24 can recommend the first financial service to the first customer according to the following method, including: determining the first marketing manager corresponding to the first marketing strategy based on the correspondence between a preset marketing strategy management table and the first marketing strategy, wherein the marketing strategy management table contains the correspondence between multiple marketing strategies and marketing managers; instructing the first marketing manager to recommend the first financial service to the first customer according to the first marketing strategy.

[0150] Specifically, Figure 3 is a schematic diagram of an optional marketing decision system architecture according to an embodiment of the present application, such as Figure 3 The marketing decision system can include four major functional modules: customer management, grid management, push management, and operation analysis, among which:

[0151] 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 customer portrait, credit rating and customer relationship analysis are carried out, so that financial institutions can more accurately grasp customer needs, thereby providing customers with more personalized customized services and products. At the same time, the service records of marketing managers to customers are fully recorded, making customer management work more refined;

[0152] Grid management refers to dividing multiple customers in the business area of ​​an enterprise into different grids according to the financial business handling characteristics of each customer in the business area of ​​the enterprise, and determining the marketing strategy for personalized recommendations to individual customers in the grid based on the grid portrait and the business information for personalized recommendations to individual customers in the grid, and determining the marketing managers who will implement the corresponding marketing strategy based on the matching relationship between the marketing strategy and the marketing managers, to ensure that there are no dead ends or blind spots in the management of the entire business area of ​​the enterprise;

[0153] Push management refers to the push tasks within the grid as the driving force, completely recording the marketing work of marketing managers, adjusting the marketing strategies and marketing managers corresponding to the grid according to the final marketing strategy execution results, and forming a closed loop of marketing management work;

[0154] Operational analysis refers to the analysis of customer distribution, product penetration, performance of marketing managers, and results of marketing strategy implementation in order to adjust financial institutions' performance evaluation indicators and operating strategies for marketing managers.

[0155] In addition, the marketing decision system also includes a data storage module, and the data storage module includes a plurality of storage units, and each storage unit is used to store different types of resource information.

[0156] For example, Figure 4 is a schematic diagram of the structure of an optional data storage module according to an embodiment of the present application, such as Figure 4 As shown. The data storage module includes five major sections: basic knowledge, system classification, business introduction, training and learning, and operation management, and the relevant resource content of each section is stored in the corresponding storage unit. Among them, the basic knowledge section includes institutional situation, corporate culture, organizational structure and other contents; the system classification section includes corporate governance system, business system, management system, risk system and other contents; the business introduction section includes deposit products, loan products, financial products and other contents; training and learning includes system construction, on-the-job training materials, examination question bank and other contents; operation management includes account management, cash management, counter management, electronic banking management and other contents. The above modules cover the learning needs and business handling needs of employees in different positions and departments of financial institutions, so that every employee can easily find learning content related to their work.

[0157] It should be noted that each module in the marketing decision-making system based on intelligent agent and text technology in the embodiment of the present application corresponds one by one to each implementation step of the marketing decision-making method based on intelligent agent and text technology in Example 1. Since a detailed description has been given in Example 1, some details not reflected in this embodiment can be referred to Example 1, and will not be repeated here.

[0158] Example 3

[0159] According to an embodiment of the present application, a computer program product is also provided, which includes a computer program, wherein when the computer program is executed by a processor, the marketing decision-making method based on intelligent agent and text technology in Example 1 is implemented.

[0160] According to an embodiment of the present application, a non-volatile storage medium is also provided, which includes a stored computer program, wherein the device where the non-volatile storage medium is located executes the marketing decision-making method based on intelligent agent and text technology in Example 1 by running the computer program.

[0161] According to an embodiment of the present application, a processor is also provided, which is used to run a computer program, wherein the marketing decision-making method based on intelligent agent and text technology in Example 1 is executed when the computer program is running.

[0162] According to an embodiment of the present application, an electronic device is also provided, which includes: a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to execute the marketing decision-making method based on intelligent agent and text technology in Example 1 through the computer program.

[0163] Optionally, when the computer program is running, the following steps are executed: obtaining a first financial business consultation text input by a first customer, and first financial business handling information of the first customer in a preset marketing area, wherein the first financial business consultation text at least includes: first customer information and first business demand information of the first customer, and the first financial business handling information is at least used to reflect the number of times the first customer has handled financial business at all financial business handling points in the marketing area; using a pre-trained financial business recommendation model to perform text analysis on the first financial business consultation text to obtain a first financial business to be recommended to the first customer, and determining a first marketing grid to which the first customer belongs based on the first financial business handling information, and determining a first grid portrait of the first marketing grid; using a pre-trained marketing strategy recommendation model to analyze the first grid portrait and the business information of the first financial business to obtain a first marketing strategy for recommending the first financial business to the first customer; and recommending the first financial business to the first customer based on the first marketing strategy.

[0164] As an optional implementation, the electronic device may be in the form of a mobile terminal, a computer terminal or a similar computing device. Figure 5 FIG. 1 shows a hardware structure block diagram of an electronic device for implementing a marketing decision method based on an agent and text technology. Figure 5As shown, the electronic device 50 may include one or more (502a, 502b, ..., 502n are used to illustrate) processors 502 (the processor 502 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 504 for storing data, and a transmission device 506 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 5 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 5 More or fewer components as shown, or with Figure 5 Different configurations shown.

[0165] It should be noted that the one or more processors 502 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 electronic device 50. As described 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).

[0166] The memory 504 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the marketing decision-making method based on agent and text technology in the embodiment of the present application. The processor 502 executes various functional applications and data processing by running the software programs and modules stored in the memory 504, that is, realizing the vulnerability detection method of the above-mentioned application. The memory 504 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 examples, the memory 504 may further include a memory remotely arranged relative to the processor 502, and these remote memories may be connected to the electronic device 50 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 a combination thereof.

[0167] The transmission device 506 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 electronic device 50. In one example, the transmission device 506 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 506 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0168] The display may be, for example, a touch screen liquid crystal display (LCD) that may enable a user to interact with a user interface of the electronic device 50 .

[0169] The serial numbers of the above embodiments are only for description and do not represent the advantages or disadvantages of the embodiments.

[0170] 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.

[0171] 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 system 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.

[0172] 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.

[0173] 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.

[0174] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent business, 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 prior art, or all or part of the technical solution can be embodied in the form of a software business. The computer software business is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of each embodiment method 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.

[0175] 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 marketing decision-making method based on intelligent agent and text technology, characterized in that: include: Acquire a first financial business consultation text input by a first customer, and first financial business handling information of the first customer in a preset marketing area, wherein the first financial business consultation text at least includes: first customer information and first business demand information of the first customer, and the first financial business handling information is at least used to reflect the number of times the first customer has handled financial business at all financial business handling points in the marketing area; Using a pre-trained financial service recommendation model to perform text analysis on the first financial service consultation text, obtain a first financial service to be recommended to the first customer, and determine a first marketing grid to which the first customer belongs based on the first financial service handling information, and determine a first grid profile of the first marketing grid; Analyze the first grid portrait and the business information of the first financial business by using a pre-trained marketing strategy recommendation model to obtain a first marketing strategy for recommending the first financial business to the first customer; The first financial service is recommended to the first customer according to the first marketing strategy.

2. The method according to claim 1, characterized in that The training process of the financial service recommendation model includes: Constructing the first neural network model; Acquire multiple groups of first training samples, wherein each group of the first training samples includes: a second financial service consultation text input by a second customer and a second financial service selected by the second customer from among multiple financial services selected from the second customer based on the second financial service consultation text; For each group of the first training samples, inputting the second financial business consulting text in the first training samples into the first neural network model to obtain the predicted financial business output by the first neural network model; A first objective loss function of the first neural network model is constructed based on the second financial business in each group of the first training samples and the predicted financial business, and the first objective loss function is optimized until a first preset convergence condition is met, so as to obtain the trained financial business recommendation model.

3. The method according to claim 2, characterized in that Get the first training sample of multiple training sets, including: Acquire second financial business consultation texts respectively input by a plurality of second customers, wherein the second financial business consultation texts at least include: second customer information and second business demand information of the second customer, and the second customer information includes at least one of the following: account balance, deposit amount, income information, financial management information, consumption information, and credit information; Acquire from the financial institution a similar customer set of each second customer and a first financial service set associated with the second service demand information in a financial service set handled by each similar customer in the similar customer set; For each second customer, the propensity of the second customer to select each financial service in the first financial service set is determined according to the similar customer set and the first financial service set in accordance with the following formula, including: ， In the formula, represents the inclination of the second customer u to select the financial service i in the first financial service set, represents the weight coefficient, represents the first inclination of the second customer u to select the financial service i, and The expression is as follows: ， In the formula, represents the set of similar customers of the second customer u, It represents the average of the inclinations of all customers in a financial institution to choose to handle all financial services sold by the financial institution. represents the bias value of the second customer u, represents the bias value of the financial service i, represents the inclination of similar customer v in the similar customer set to select the financial service i, represents the average value of the propensity of the similar customer v to choose a financial institution to sell all financial services, represents the similarity between the second customer u and the similar customer v in the similar customer set, and The expression is: ， Wherein, I represents the first financial service set, represents the average value of the propensity of the second customer u to choose to handle all financial services offered by the financial institution, represents the inclination of the second customer u to select the financial service i, represents the second inclination of the second customer u to select the financial service i, and The expression is: ， In the formula, represents the second financial service set that is most similar to the financial service i, represents the deviation of the second customer u to the financial service i, represents the similarity between the financial service i and the financial service j in the second financial service set, and The expression is: ， In the formula, represents the inclination of the similar customer v to choose the financial service i, represents the inclination of the similar customer v to choose the financial service j, represents the inverse customer frequency of the financial service j, and The expression is: ， In the formula, represents the total number of customers in the similar customer set who have a tendency to choose the financial service j; Feedback the financial service whose tendency is higher than the preset threshold value to the second customer, and receive a financial service selection instruction from the second customer, wherein the financial service selection instruction carries at least a second financial service selected by the second customer from the financial services whose tendency is higher than the preset threshold value; The plurality of groups of first training samples are composed of the second financial service consultation texts input by each of the second customers and the second financial services selected by each of the second customers.

4. The method according to claim 1, characterized in that: Determining a first marketing grid to which the first customer belongs according to the first financial business handling information includes: Determine the number of times the first customer has handled financial services at each financial service handling point in the marketing area according to the first financial service handling information, and delete the financial service handling points whose number of handling times is lower than the first number threshold from all financial service handling points in the marketing area to obtain a set of financial service handling points; The financial business handling point at which the first customer has handled the most transactions in the set of financial business handling points is taken as the first financial business handling point, and the financial business handling points within the marketing area whose distance from the first financial business handling point is less than a preset distance threshold are taken as the first cluster set; the financial business handling point at which the first customer has handled the most transactions in other financial business handling points outside the first cluster set within the marketing area is taken as the second financial business handling point, and the financial business handling points within the marketing area whose distance from the first financial business handling point is less than the distance threshold and outside the first cluster set are taken as the second cluster set; the above steps are repeated until all financial business handling points within the set of financial business handling points are clustered into respective cluster sets; Determine the total number of times the first customer has handled financial business at all financial business handling points in each of the cluster sets, and determine a target cluster set whose total number is less than a second number threshold, and use the position coordinates of the target financial business handling point in the target cluster set as the position coordinates of the target cluster set, wherein the target financial business handling point is the financial business handling point with the largest number of handling times in the target cluster set; The frequently-handled business area of ​​the first customer is determined according to the location coordinates of the target cluster set, and the marketing grid in which the frequently-handled business area is located is determined according to a preset marketing area grid division table, and the marketing grid is used as the first marketing grid to which the first customer belongs, wherein the marketing area grid division table includes multiple marketing grids obtained by dividing the marketing area according to preset grid area division rules.

5. The method according to claim 4, characterized in that Determine the marketing grid where the frequently handled business area is located according to the preset marketing area grid division table, including: Determine the regional location information of each marketing grid in the marketing area; The regional location information of each of the marketing grids and the regional location information of the frequently handled business areas are analyzed by using a spatial index algorithm to determine the marketing grid in which the frequently handled business areas are located.

6. The method according to claim 1, characterized in that Determining a first grid profile of the first marketing grid includes: Acquire the business demand information and customer information of all third customers in the first marketing grid, and pre-process the business demand information and customer information of each third customer; For each third customer in the first marketing grid, feature extraction is performed on the customer information of the third customer to obtain all key features related to the business demand information of the third customer, wherein the type of the key feature includes at least one of the following: consumption behavior feature, loan behavior feature, and credit business handling feature; A first grid portrait of the first marketing grid is constructed based on all key features of each of the third customers.

7. The method according to claim 1, characterized in that The training process of the marketing strategy recommendation model includes: Constructing a second neural network model; Acquire multiple groups of second training samples, wherein each group of the second training samples includes: a second grid portrait of the second marketing grid to which the second customer belongs, business information of a second amount financial business to be recommended to the second customer, and a second marketing strategy for recommending the second financial business to the second customer, and the second marketing strategy is a marketing strategy with the highest customer conversion rate, the lowest strategy execution cost, and the highest customer satisfaction; For each group of the second training samples, the second grid portrait and the business information of the second amount financial business in the second training samples are input into the second neural network model to obtain a predictive marketing strategy output by the second neural network model for recommending the second financial business to the second customer; The second objective loss function of the second neural network model is constructed according to the second marketing strategy and the predicted marketing strategy in each group of the second training samples, and the second objective loss function is optimized until the preset second convergence condition is met, so as to obtain the marketing strategy recommendation model that has completed training.

8. The method according to claim 1, characterized in that Recommending the first financial service to the first customer according to the first marketing strategy includes: Determining a first marketing manager corresponding to the first marketing strategy according to a correspondence between a preset marketing strategy management table and the first marketing strategy, wherein the marketing strategy management table includes correspondences between a plurality of marketing strategies and marketing managers; Instruct the first marketing manager to recommend the first financial service to the first customer according to the first marketing strategy.

9. A marketing decision-making system based on intelligent agent and text technology, characterized in that: include: an acquisition module, used to acquire a first financial business consultation text input by a first customer, and first financial business handling information of the first customer in a preset marketing area, wherein the first financial business consultation text at least includes: first customer information and first business demand information of the first customer, and the first financial business handling information at least reflects the number of times the first customer has handled financial business at all financial business handling points in the marketing area; a first determination module, configured to perform text analysis on the first financial service consultation text using a pre-trained financial service recommendation model to obtain a first financial service to be recommended to the first customer, and determine a first marketing grid to which the first customer belongs based on the first financial service handling information, and determine a first grid profile of the first marketing grid; A second determination module is used to analyze the first grid portrait and the business information of the first financial business by using a pre-trained marketing strategy recommendation model to obtain a first marketing strategy for recommending the first financial business to the first customer; A recommendation module is used to recommend the first financial service to the first customer according to the first marketing strategy.

10. A computer program product, characterized in that include: A computer program, wherein when the computer program is executed by a processor, it implements the marketing decision-making method based on intelligent agent and text technology as described in any one of claims 1 to 8.

Images