Marketing strategy distribution method and device, electronic equipment and computer program product

Abstract

The invention discloses a marketing strategy distribution method and apparatus, an electronic device and a computer program product. The method comprises the steps of obtaining feature data of a user; according to the feature data of the user, utilizing a preset marketing strategy evaluation model to predict conversion rate gains of the user under different marketing strategies; according to the conversion rate gain of the user under different marketing strategies and a preset marketing budget, a multi-option backpack problem is constructed and solved, and a target marketing strategy corresponding to the user is obtained; and distributing the target marketing strategy corresponding to the user to the user. According to the marketing strategy recommendation method and device, the preset marketing strategy evaluation model is adopted to recommend the marketing strategy, the accuracy and individuation level of marketing strategy recommendation are remarkably improved, the user experience and satisfaction degree are enhanced, and the customer loyalty and the re-purchase rate are improved. Besides, in combination with budget constraints and marketing objectives of enterprises, recommendation problems are converted into multi-option knapsack problems to be solved, and maximization of customer values and marketing effects under limited resources is guaranteed.

Description

Technical Field

[0001] The present application relates to the technical field of financial product marketing, and in particular to a method, device, electronic device, and computer program product for distributing marketing strategies. Background Art

[0002] At the forefront of digital marketing in the banking industry, precision marketing relies on in-depth analysis of multi-dimensional portrait labels such as user personal characteristics, asset status, and behavioral trajectories. Faced with many marketing strategies that may be suitable for a single user, in order to avoid excessively disturbing customers and reduce marketing costs, how to distribute marketing strategies in a personalized manner to achieve the optimal customer purchase rate has become a challenge.

[0003] Existing methods comprehensively consider factors such as the past effects of marketing strategies and the importance of customer positioning, and give each strategy a comprehensive score through an expert review system to form a priority ranking mechanism. In the face of a user adapting to multiple strategies, the highest priority marketing strategy is automatically determined for the customer based on this priority sequence.

[0004] However, the above method has at least the following technical problems:

[0005] 1) Marketing strategies are distributed completely according to strategy priorities, without any adjustments based on the users themselves, making it difficult to meet the personalized needs of different users;

[0006] 2) Expert scoring is highly subjective, and each formulation requires research, data analysis and manual judgment by business personnel. It cannot be adjusted frequently and has poor timeliness;

[0007] 3) The current distribution plan includes naturally converted customers and marketing-averse customers, rather than marketing-sensitive customers, resulting in a waste of marketing resources and reduced customer satisfaction;

[0008] 4) With limited marketing budget, the issue of maximizing the ROI (Return on Investment) of marketing activities is not considered. Summary of the invention

[0009] In order to solve at least one of the technical problems mentioned above, the embodiments of the present application provide a marketing strategy distribution method, device, electronic device, and computer program product to improve the accuracy and personalization level of marketing strategy recommendations, reduce marketing costs and improve return on investment.

[0010] The present application embodiment adopts the following technical solutions:

[0011] In a first aspect, an embodiment of the present application provides a method for distributing a marketing strategy, the method for distributing a marketing strategy comprising:

[0012] Obtain user feature data;

[0013] According to the characteristic data of the user, using a preset marketing strategy evaluation model to predict the conversion rate gain of the user under different marketing strategies;

[0014] According to the conversion rate gain of the user under different marketing strategies and the preset marketing budget, a multi-option knapsack problem is constructed and solved to obtain the target marketing strategy corresponding to the user;

[0015] The target marketing strategy corresponding to the user is distributed to the user.

[0016] Optionally, the preset marketing strategy evaluation model is trained in the following manner:

[0017] Build a user's marketing strategy dataset;

[0018] The preset marketing strategy evaluation model is obtained by training the causal forest algorithm according to the marketing strategy data set of the user.

[0019] Optionally, constructing a user's marketing strategy dataset includes:

[0020] Obtain a user list, the user list including a historical marketing user list and a historical non-marketing user list for each marketing strategy in a marketing strategy list within a preset time period;

[0021] Collecting feature data of each user in the user list, and constructing a user feature vector according to the feature data of each user;

[0022] Calculate the marketing evaluation score corresponding to each user in the user list;

[0023] A marketing strategy data set for the user is constructed according to the user feature vector, the marketing strategy list, and the marketing evaluation score.

[0024] Optionally, calculating the marketing evaluation score corresponding to each user in the user list includes:

[0025] Calculate the marketing evaluation scores corresponding to the users in the historical marketing user list according to the product purchase rate and asset management scale increase rate of the users in the historical marketing user list within a preset time after the marketing;

[0026] The marketing evaluation scores corresponding to the users in the historical non-marketing user list are calculated based on the product purchase rate and asset management scale improvement rate of the users in the historical non-marketing user list within a preset time without marketing.

[0027] Optionally, the step of training the user's marketing strategy dataset using a causal forest algorithm to obtain the preset marketing strategy evaluation model comprises:

[0028] constructing a plurality of data subsets according to the marketing strategy dataset of the user;

[0029] Each subset is split according to a preset splitting criterion, wherein the preset splitting criterion includes calculating the difference between the maximum value of the marketing evaluation scores corresponding to the two nodes after the splitting and the maximum value of the marketing evaluation scores corresponding to the nodes before the splitting;

[0030] When the splitting stop condition is met, the splitting is stopped and multiple causal trees are obtained as the preset marketing strategy evaluation model.

[0031] Optionally, constructing and solving a multi-option knapsack problem according to the conversion rate gain of the user under different marketing strategies and a preset marketing budget to obtain a target marketing strategy corresponding to the user includes:

[0032] Construct multiple constraints for the multi-option knapsack problem based on the user's conversion rate gain under different marketing strategies and the preset marketing budget;

[0033] Based on multiple constraints of the multi-option knapsack problem, the multi-option knapsack problem is solved using a solving algorithm to obtain a target marketing strategy corresponding to the user.

[0034] Optionally, the multiple constraints for constructing a multi-option knapsack problem based on the conversion rate gain of the user under different marketing strategies and a preset marketing budget include:

[0035] Constructing a constraint condition for maximizing the conversion rate gain according to the conversion rate gain of the user under different marketing strategies and a marketing strategy distribution matrix;

[0036] According to the single-user marketing cost of each marketing strategy, the marketing strategy distribution matrix and the preset marketing budget, construct the constraint conditions of the preset marketing budget;

[0037] A constraint condition on the number of marketing strategies distributed to a single user is constructed according to the marketing strategy distribution matrix.

[0038] In a second aspect, an embodiment of the present application further provides a marketing strategy distribution device, the marketing strategy distribution device comprising:

[0039] An acquisition unit, used for acquiring characteristic data of a user;

[0040] A prediction unit, configured to predict the conversion rate gain of the user under different marketing strategies using a preset marketing strategy evaluation model according to the characteristic data of the user;

[0041] A construction and solution unit, used to construct and solve a multi-option knapsack problem according to the conversion rate gain of the user under different marketing strategies and a preset marketing budget, so as to obtain a target marketing strategy corresponding to the user;

[0042] The distribution unit is used to distribute the target marketing strategy corresponding to the user to the user.

[0043] In a third aspect, an embodiment of the present application further provides an electronic device, including:

[0044] Processor; and

[0045] A memory arranged to store computer executable instructions, which, when executed, cause the processor to perform any of the aforementioned methods for distributing marketing strategies.

[0046] In a fourth aspect, an embodiment of the present application also provides a computer-readable storage medium, which stores one or more programs. When the one or more programs are executed by an electronic device including multiple applications, the electronic device executes any of the aforementioned methods for distributing marketing strategies.

[0047] At least one of the above technical solutions adopted in the embodiments of the present application can achieve the following beneficial effects: the distribution method of the marketing strategy in the embodiments of the present application first obtains the characteristic data of the user; then, based on the characteristic data of the user, a preset marketing strategy evaluation model is used to predict the conversion rate gain of the user under different marketing strategies; then, a multi-option knapsack problem is constructed and solved according to the conversion rate gain of the user under different marketing strategies and the preset marketing budget, and the target marketing strategy corresponding to the user is obtained; finally, the target marketing strategy corresponding to the user is distributed to the user. The distribution method of the marketing strategy in the embodiments of the present application adopts a preset marketing strategy evaluation model to recommend marketing strategies, which significantly improves the accuracy and personalization level of marketing strategy recommendations, enhances user experience and satisfaction, and improves customer loyalty and repurchase rate. In addition, combined with the budget constraints and marketing goals of the enterprise, the recommendation problem is converted into a multi-option knapsack problem solution, which ensures the maximization of customer value and marketing effect under limited resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] 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:

[0049] Figure 1 A schematic diagram of a method for distributing a marketing strategy in an embodiment of the present application;

[0050] Figure 2A schematic diagram of the structure of a distribution device for a marketing strategy in an embodiment of the present application;

[0051] Figure 3 This is a schematic diagram of the structure of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION

[0052] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in combination with the specific embodiments of the present application and the corresponding drawings. 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 making creative work are within the scope of protection of the present application.

[0053] The technical solutions provided by various embodiments of the present application are described in detail below in conjunction with the accompanying drawings.

[0054] The present application embodiment provides a method for distributing a marketing strategy, such as Figure 1 As shown, a flow chart of a method for distributing a marketing strategy in an embodiment of the present application is provided, and the method for distributing a marketing strategy at least includes the following steps S110 to S140:

[0055] Step S110, obtaining characteristic data of the user.

[0056] When distributing marketing strategies to users, the embodiments of the present application need to first obtain the user's characteristic data. The characteristic data here may include data of multiple dimensions, for example, it may include the user's basic attribute data such as age, gender, region, etc., and may also include data related to product transactions such as contract status, asset level, medium- and long-term transactions, product holding status, etc., and may also include the user's online behavior data, etc. The more comprehensive and rich the characteristic data obtained, the higher the accuracy of the subsequent recommended marketing strategy, and the more it can meet the personalized needs of users. Of course, the technical personnel in this field can flexibly set which dimensions of user characteristic data to obtain according to actual needs, and no specific limitation is made here.

[0057] Step S120: predicting the conversion rate gain of the user under different marketing strategies using a preset marketing strategy evaluation model based on the characteristic data of the user.

[0058] The above acquired user feature data is used as input to a pre-trained marketing strategy evaluation model. The model can predict the user's conversion rate gain under the intervention of different marketing strategies. The conversion rate gain can be understood as the change in the conversion rate under the intervention of a certain marketing strategy relative to the conversion rate without marketing. The greater the conversion rate gain corresponding to the user under the intervention of a certain marketing strategy, the more likely the user is to purchase the marketing product under the intervention of the marketing strategy. The conversion rate of the embodiment of the present application can be characterized by a marketing evaluation score, which can be further obtained by comprehensive calculation of the product purchase rate and the asset management scale improvement rate.

[0059] The preset marketing strategy evaluation model in the embodiment of the present application can be obtained through training based on the causal forest algorithm, etc., and can be regularly updated and optimized. The specific training method can be flexibly set by technicians in this field according to actual needs, and is not specifically limited here.

[0060] Step S130, constructing and solving a multi-option knapsack problem according to the conversion rate gain of the user under different marketing strategies and the preset marketing budget, and obtaining the target marketing strategy corresponding to the user.

[0061] Since the existing distribution scheme does not consider the maximization of marketing activity ROI under a limited budget, the embodiment of the present application further considers the marketing budget constraints of the enterprise based on the conversion rate gains of users under different marketing strategies obtained in the above steps, and combines the budget constraints and marketing goals of the enterprise to transform the marketing strategy recommendation problem into a multi-option knapsack problem and solve it, thereby obtaining the best marketing strategy for each user.

[0062] Step S140: Distribute the target marketing strategy corresponding to the user to the user.

[0063] The target marketing strategy corresponding to each user obtained from the above solution is distributed to the user. For example, a marketing user list can be generated for all users to whom marketing strategies are to be issued. The list includes users who need marketing and the corresponding marketing strategies that need to be issued to users. The marketing strategies are then distributed according to the list, thereby achieving precise marketing while also optimizing the allocation and efficient use of marketing resources, thereby reducing marketing costs.

[0064] The marketing strategy distribution method of the embodiment of the present application adopts a preset marketing strategy evaluation model to recommend marketing strategies, which significantly improves the accuracy and personalization level of marketing strategy recommendations, enhances user experience and satisfaction, and improves customer loyalty and repurchase rate. In addition, combined with the company's budget constraints and marketing goals, the recommendation problem is transformed into a multi-option knapsack problem solution, ensuring the maximization of customer value and marketing effect under limited resources.

[0065] In some embodiments of the present application, the preset marketing strategy evaluation model is trained in the following manner: constructing a user's marketing strategy data set; and training using a causal forest algorithm based on the user's marketing strategy data set to obtain the preset marketing strategy evaluation model.

[0066] When training a preset marketing strategy evaluation model, the embodiment of the present application can first construct a user's marketing strategy data set as a training sample data set, and then based on the ensemble learning idea of ​​the causal forest, take the causal forest model as the core model for training to obtain a trained preset marketing strategy evaluation model.

[0067] The causal forest model is based on random forests and achieves data homogeneity / no confusion in the local feature space by repeatedly partitioning the feature space (Recursive Partitioning). This partitioning method enables the model to estimate treatment effects in different feature subspaces, thereby achieving accurate estimation of heterogeneous treatment effects.

[0068] In the intelligent marketing scenario of the embodiment of the present application, the causal forest model can be used to estimate the difference in conversion effects of different marketing strategies for users, thereby achieving more accurate marketing decisions. In specific applications, the feature data to be predicted by the user can be input into multiple causal trees respectively, and the predicted value of the conversion rate gain output by each causal tree can be obtained. The predicted values ​​of the conversion rate gain output by all causal trees are averaged to obtain the final conversion rate gain of the user.

[0069] In the training process of the marketing strategy evaluation model, the embodiment of the present application adopts the causal forest integrated learning idea, constructs multiple causal trees and makes voting decisions, thereby improving the robustness and prediction accuracy of the model. In addition, by dynamically adjusting the model parameters and structure, it is possible to achieve rapid response to market changes and user behavior changes, and maintain the cutting-edge and timeliness of marketing strategy recommendations.

[0070] In some embodiments of the present application, constructing a user's marketing strategy data set includes: obtaining a user list, the user list including a historical marketing user list and a historical non-marketing user list for each marketing strategy in a marketing strategy list within a preset time period; collecting feature data of each user in the user list, and constructing a user feature vector based on the feature data of each user; calculating a marketing evaluation score corresponding to each user in the user list; and constructing the user's marketing strategy data set based on the user feature vector, the marketing strategy list, and the marketing evaluation score.

[0071] When constructing a user's marketing strategy dataset D, the embodiment of the present application can first obtain a marketing strategy list T = {T1, T2, ..., T K}, and sample the users who have never been intervened by any marketing strategy in the past period of time, such as three months, to obtain a list of historical non-marketed users. The sampling volume should preferably be at the same order of magnitude as the average number of users of all marketing strategies.

[0072] It should be noted that the above-collected historical marketing user list and historical non-marketing user list are extracted based on A / B test. The historical non-marketing user list is a specially reserved control group, which is statistically homogeneous with the historical marketing user list. In actual application, the marketing group and the control group will be randomly selected from the total user list to ensure their homogeneity. Then the marketing group will be randomly divided to carry out various strategic marketing. This does not conflict with the existing marketing process. It only extracts a small part of the traffic for A / B test to obtain a data set that can be used to build a marketing strategy evaluation model.

[0073] Then collect the user's feature data in multiple dimensions, such as basic attribute data such as age, gender, region, etc., data related to product transactions such as contract signing, asset level, medium- and long-term transactions, product holdings, etc., and online behavior data of users, etc. Preprocess and feature engineer the user's feature data in multiple dimensions, such as handling null values, outliers, etc. in the data, and vectorize the categorical variables to construct a user feature vector X that meets the model input requirements.

[0074] The embodiment of the present application innovatively designs a multi-dimensional feature fusion mechanism, which not only considers the basic attribute information of users, but also deeply mines multi-dimensional information such as the user's asset level, product holdings, historical purchase behavior, and online behavior data. In addition, through algorithms such as feature selection and weight allocation, the feature combination with the greatest impact on marketing strategy recommendations is automatically screened out and effectively integrated to form comprehensive and accurate user portrait data.

[0075] In addition, the embodiment of the present application also designs an indicator for measuring and evaluating the marketing effect of the marketing strategy on the user and calculating the marketing evaluation score Y, and reflects the intervention effect of the marketing strategy on the user by calculating the marketing evaluation score corresponding to each user in the user list. The specific indicator used to calculate the marketing evaluation score can be flexibly set by those skilled in the art in combination with the actual application scenario, and is not specifically limited here.

[0076] Finally, based on the above user feature vector X, marketing strategy list T and marketing evaluation score Y, the final user marketing strategy data set D = (X, T, Y) is constructed.

[0077] In some embodiments of the present application, calculating the marketing evaluation score corresponding to each user in the user list includes: calculating the marketing evaluation score corresponding to the users in the historical marketing user list based on the product purchase rate and asset management scale improvement rate of the users in the historical marketing user list within a preset time after marketing; calculating the marketing evaluation score corresponding to the users in the historical non-marketing user list based on the product purchase rate and asset management scale improvement rate of the users in the historical non-marketing user list within a preset time without marketing.

[0078] When calculating the marketing evaluation score, the embodiment of the present application selected two indicators for comprehensive calculation, one is the product purchase rate, and the other is the AUM (Assets Under Management) improvement rate.

[0079] The product purchase rate can be understood as the ratio of the user's corresponding product purchase amount to the opening asset balance. For users in the historical marketing user list, the product purchase rate within the preset time after the marketing can be understood as the ratio of the actual purchase or investment amount for a specific financial product or service to the opening asset balance within a certain period of time after the marketing activity, such as within 7 days.

[0080] This metric reflects the impact of marketing activities on the direct purchasing behavior of the target customer group. A high purchase rate usually means that the marketing activities have successfully stimulated customers' purchasing interest and prompted them to take action.

[0081] For users in the historical marketing user list, the AUM increase rate within the preset time after the marketing refers to the growth rate of the asset management scale within a certain period of time after the marketing activity, such as within 7 days. In the financial field, AUM is an important indicator to measure the total scale of assets managed by asset management companies or financial institutions. The AUM increase rate reflects the contribution of marketing activities to the growth of asset scale.

[0082] Taking into account the actual business scenarios, some users who are not affected by marketing strategies may also purchase related financial products within a certain period of time. Therefore, the calculation of product purchase rate and AUM increase rate will also be involved. However, the product purchase rate and AUM increase rate here have nothing to do with marketing strategies and are only the results calculated based on the natural growth rate.

[0083] For the product purchase rate and AUM increase rate of users in the above-mentioned historical marketing user list within a period of time after marketing, as well as the product purchase rate and AUM increase rate of users in the historical non-marketing user list within a period of time without marketing, the final marketing evaluation score Y can be calculated by weighted summation of the product purchase rate and AUM increase rate respectively. The specific weight distribution can be flexibly set according to actual needs and is not specifically limited here.

[0084] In some embodiments of the present application, the method of training the marketing strategy dataset of the user using a causal forest algorithm to obtain the preset marketing strategy evaluation model includes: constructing multiple data subsets based on the marketing strategy dataset of the user; splitting each subset according to a preset splitting criterion, and calculating the difference between the maximum marketing evaluation score corresponding to the two nodes after the split and the maximum marketing evaluation score corresponding to the node before the split; and when the splitting stop condition is met, stopping the splitting and obtaining multiple causal trees as the preset marketing strategy evaluation model.

[0085] When training the model, N samples are randomly selected with replacement from the user's marketing strategy dataset D to form a subset Si. This process is repeated Nt times to construct N_t data subsets. The sampling process ensures that the samples of each marketing strategy are selected in proportion.

[0086] Then for each subset Si of the N_t data subsets, at each node, randomly select m feature coordinates for splitting. Assume that the current node is node n , split into left and right nodes node l and node r , and their sample numbers are N n , N l , N r , the splitting criterion can be expressed as follows:

[0087]

[0088] Among them, K is the number of marketing strategies, Y is the marketing evaluation score, and Δ is the conversion rate gain before and after the split, that is, the gain of the marketing evaluation score. The splitting method that maximizes this gain is selected for splitting.

[0089] Find the splitting method that can maximize the expected response increase among the above m features, and stop splitting when any of the following conditions is met:

[0090] 1) The number of samples processed is less than the minimum number of splits (min split) defined by the user;

[0091] 2) There is no partition that can lead to a non-negative gain in the estimated Δ;

[0092] 3) All samples in a node have the same response value.

[0093] For any given customer X i , Marketing strategy k , the prediction result is the average of all causal tree predictions.

[0094] The embodiment of the present application introduces a causal forest algorithm as the core model. Compared with the traditional decision tree, the causal forest model of the embodiment of the present application adopts a more complex splitting criterion when splitting nodes, and comprehensively considers multiple dimensions such as information gain, similarity between marketing strategies, and user response prediction.

[0095] In some embodiments of the present application, constructing and solving a multi-option knapsack problem based on the conversion rate gain of the user under different marketing strategies and a preset marketing budget to obtain a target marketing strategy corresponding to the user includes: constructing multiple constraints of the multi-option knapsack problem based on the conversion rate gain of the user under different marketing strategies and a preset marketing budget; solving the multi-option knapsack problem using a solution algorithm based on the multiple constraints of the multi-option knapsack problem to obtain the target marketing strategy corresponding to the user.

[0096] Based on the prediction results of users under different marketing strategies output by the marketing strategy evaluation model of the aforementioned embodiment, maximizing the product conversion rate gain under the budget constraint is modeled as a multiple choice knapsack problem (MCKP for short), and a personalized distribution plan for user marketing strategies is obtained. The multiple choice knapsack problem is a variant of the knapsack problem, which introduces additional constraints, making the problem more complex and flexible. The multiple choice knapsack problem is solved by adopting a certain solution algorithm, such as solving its dual problem, and finally the target marketing strategy corresponding to the user is obtained. The specific solution algorithm to be adopted can be flexibly selected by technicians in this field according to actual needs, and is not specifically limited here.

[0097] In some embodiments of the present application, the multiple constraints for constructing a multi-option knapsack problem based on the user's conversion rate gain under different marketing strategies and a preset marketing budget include: constructing a constraint for maximizing the conversion rate gain based on the user's conversion rate gain under different marketing strategies and a marketing strategy distribution matrix; constructing a constraint for a preset marketing budget based on the single-user marketing cost of each marketing strategy and the marketing strategy distribution matrix and a preset marketing budget; and constructing a constraint on the number of single-user marketing strategy distributions based on the marketing strategy distribution matrix.

[0098] The multiple constraints of the multi-option knapsack problem constructed in the embodiment of the present application may include the following aspects:

[0099]

[0100] Where T = (t jk ) is the marketing strategy distribution matrix, t jk =1 means that k marketing strategy is issued to user j, t jk=0 means that k marketing strategy is not issued to user j, and a single user can only issue one marketing strategy at most; P = (p jk ) is the prediction result of the marketing strategy evaluation model, i k is the single customer marketing cost of marketing strategy k.

[0101] Therefore, formula (2) indicates that the conversion rate gain of users under various marketing strategies is maximized, formula (3) indicates that the total cost of issuing marketing strategies to all users must not exceed the maximum budget limit COST_LIMIT, and formula (4) indicates that a single customer can only be issued one marketing strategy at most to avoid wasting marketing resources and reducing customer satisfaction. Of course, this value can also be flexibly adjusted according to demand.

[0102] The marketing strategy distribution matrix is ​​the matrix to be solved by the multi-option knapsack problem. The multi-option knapsack problem is an optimization problem. The goal is to find a strategy distribution matrix that can maximize the conversion rate gain while satisfying the constraints. Therefore, based on the above constraints, the above problem is solved by solving its dual problem to obtain the optimal T = (t jk ).

[0103] In summary, the key points and technical effects of the distribution method of the marketing strategy of this application mainly include:

[0104] 1) Multi-dimensional feature fusion mechanism: This application innovatively designs a multi-dimensional feature fusion mechanism, which not only considers the basic attribute information of users, but also deeply mines multi-dimensional information such as users' asset level, product holdings, historical purchase behavior, and online behavior data. Through feature selection and weight allocation algorithms, the feature combination with the greatest impact on marketing strategy recommendations is automatically screened out and effectively integrated to form a comprehensive and accurate user portrait.

[0105] 2) Causal forest model construction: The causal forest algorithm is introduced as the core model. Compared with the traditional decision tree, the intelligent causal tree of this application adopts a more complex splitting criterion when splitting nodes, and comprehensively considers multiple dimensions such as information gain, similarity between strategies and user response prediction. During the model training process, the integrated learning idea of ​​causal forest is adopted to construct multiple intelligent causal trees and make voting decisions to improve the robustness and prediction accuracy of the model. In addition, it is also possible to achieve rapid response to market changes and user behavior changes by dynamically adjusting model parameters and structures, and maintain the cutting-edge and timeliness of marketing strategy recommendations.

[0106] 3) Personalized marketing strategy recommendation strategy: Based on the constructed multi-dimensional user portrait and causal forest model, this application proposes a personalized marketing strategy recommendation strategy. This strategy can automatically match and recommend the most suitable marketing strategy for each user's unique characteristics and needs, enhance user experience and satisfaction, and improve customer loyalty and repurchase rate. At the same time, combined with the company's budget constraints and marketing goals, the recommendation problem is transformed into a multi-option knapsack problem solution, achieving the optimal allocation and efficient use of marketing resources, and ensuring the maximization of customer value and marketing effect under limited resources.

[0107] The present application embodiment provides a distribution device 200 for a marketing strategy, such as Figure 2 As shown, a schematic diagram of the structure of a distribution device for a marketing strategy in an embodiment of the present application is provided. The distribution device 200 for the marketing strategy includes: an acquisition unit 210, a prediction unit 220, a construction and solution unit 230, and a distribution unit 240, wherein:

[0108] An acquisition unit 210 is used to acquire characteristic data of a user;

[0109] The prediction unit 220 is used to predict the conversion rate gain of the user under different marketing strategies based on the characteristic data of the user and using a preset marketing strategy evaluation model;

[0110] A construction and solution unit 230 is used to construct and solve a multi-option knapsack problem according to the conversion rate gain of the user under different marketing strategies and a preset marketing budget, so as to obtain a target marketing strategy corresponding to the user;

[0111] The distribution unit 240 is used to distribute the target marketing strategy corresponding to the user to the user.

[0112] In some embodiments of the present application, the preset marketing strategy evaluation model is trained in the following manner: constructing a user's marketing strategy data set; and training using a causal forest algorithm based on the user's marketing strategy data set to obtain the preset marketing strategy evaluation model.

[0113] In some embodiments of the present application, the preset marketing strategy evaluation model is trained in the following manner: obtaining a user list, the user list including a historical marketing user list and a historical non-marketing user list for each marketing strategy in a marketing strategy list within a preset time period; collecting feature data of each user in the user list, and constructing a user feature vector based on the feature data of each user; calculating the marketing evaluation score corresponding to each user in the user list; and constructing a marketing strategy data set for the user based on the user feature vector, the marketing strategy list, and the marketing evaluation score.

[0114] In some embodiments of the present application, the preset marketing strategy evaluation model is trained in the following manner: based on the product purchase rate and asset management scale improvement rate of the users in the historical marketing user list within a preset time after marketing, the marketing evaluation score corresponding to the users in the historical marketing user list is calculated; based on the product purchase rate and asset management scale improvement rate of the users in the historical non-marketing user list within a preset time without marketing, the marketing evaluation score corresponding to the users in the historical non-marketing user list is calculated.

[0115] In some embodiments of the present application, the preset marketing strategy evaluation model is trained in the following manner: constructing multiple data subsets based on the user's marketing strategy data set; splitting each subset according to a preset splitting criterion, and calculating the difference between the maximum marketing evaluation scores corresponding to the two nodes after the split and the maximum marketing evaluation scores corresponding to the nodes before the split; when the conditions for stopping the splitting are met, stopping the splitting and obtaining multiple causal trees as the preset marketing strategy evaluation model.

[0116] In some embodiments of the present application, the construction and solution unit 230 is specifically used to: construct multiple constraints of a multi-option knapsack problem according to the conversion rate gain of the user under different marketing strategies and a preset marketing budget; based on the multiple constraints of the multi-option knapsack problem, use a solution algorithm to solve the multi-option knapsack problem to obtain the target marketing strategy corresponding to the user.

[0117] In some embodiments of the present application, the construction and solution unit 230 is specifically used to: construct constraints for maximizing conversion rate gains based on the conversion rate gains of the user under different marketing strategies and the marketing strategy distribution matrix; construct constraints for a preset marketing budget based on the single-user marketing cost of each marketing strategy and the marketing strategy distribution matrix and the preset marketing budget; and construct constraints for the number of single-user marketing strategy distributions based on the marketing strategy distribution matrix.

[0118] It can be understood that the above-mentioned marketing strategy distribution device can implement each step of the marketing strategy distribution method provided in the aforementioned embodiment, and the relevant explanations on the marketing strategy distribution method are applicable to the marketing strategy distribution device and will not be repeated here.

[0119] Figure 3 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present application. Figure 3At the hardware level, the electronic device includes a processor, and optionally an internal bus, a network interface, and a memory. The memory may include a memory, such as a high-speed random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage. Of course, the electronic device may also include hardware required for other services.

[0120] The processor, network interface and memory can be interconnected through an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0121] The memory is used to store the program. Specifically, the program may include a program code, and the program code includes a computer operation instruction. The memory may include a memory and a non-volatile memory, and provides instructions and data to the processor.

[0122] The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it, forming a distribution device for the marketing strategy at the logical level. The processor executes the program stored in the memory and is specifically used to perform the following operations:

[0123] Obtain user feature data;

[0124] According to the characteristic data of the user, using a preset marketing strategy evaluation model to predict the conversion rate gain of the user under different marketing strategies;

[0125] According to the conversion rate gain of the user under different marketing strategies and the preset marketing budget, a multi-option knapsack problem is constructed and solved to obtain the target marketing strategy corresponding to the user;

[0126] The target marketing strategy corresponding to the user is distributed to the user.

[0127] The above application Figure 1The method executed by the distribution device of the marketing strategy disclosed in the illustrated embodiment can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in the processor or an instruction in the form of software. The above processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as a hardware decoding processor for execution, or a combination of hardware and software modules in the decoding processor for execution. The software module can be located in a storage medium mature in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0128] The present application also provides a computer-readable storage medium, which stores one or more programs, wherein the one or more programs include instructions, which, when executed by an electronic device including multiple application programs, enable the electronic device to execute Figure 1 The method executed by the distribution device of the marketing strategy in the illustrated embodiment is specifically used to execute:

[0129] Obtain user feature data;

[0130] According to the characteristic data of the user, using a preset marketing strategy evaluation model to predict the conversion rate gain of the user under different marketing strategies;

[0131] According to the conversion rate gain of the user under different marketing strategies and the preset marketing budget, a multi-option knapsack problem is constructed and solved to obtain the target marketing strategy corresponding to the user;

[0132] The target marketing strategy corresponding to the user is distributed to the user.

[0133] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0134] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0135] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0136] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0137] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0138] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0139] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0140] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0141] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0142] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A method for distributing a marketing strategy, characterized in that: The distribution methods of the marketing strategy include: Obtain user feature data; According to the characteristic data of the user, using a preset marketing strategy evaluation model to predict the conversion rate gain of the user under different marketing strategies; According to the conversion rate gain of the user under different marketing strategies and the preset marketing budget, a multi-option knapsack problem is constructed and solved to obtain the target marketing strategy corresponding to the user; The target marketing strategy corresponding to the user is distributed to the user.

2. The method for distributing marketing strategies according to claim 1, characterized in that: The preset marketing strategy evaluation model is trained in the following way: Build a user's marketing strategy dataset; The preset marketing strategy evaluation model is obtained by training the causal forest algorithm according to the marketing strategy data set of the user.

3. The method for distributing marketing strategies according to claim 2, characterized in that: The method of constructing a user's marketing strategy dataset includes: Obtain a user list, the user list including a historical marketing user list and a historical non-marketing user list for each marketing strategy in a marketing strategy list within a preset time period; Collecting feature data of each user in the user list, and constructing a user feature vector according to the feature data of each user; Calculate the marketing evaluation score corresponding to each user in the user list; A marketing strategy data set for the user is constructed according to the user feature vector, the marketing strategy list, and the marketing evaluation score.

4. The method for distributing marketing strategies according to claim 3, characterized in that: Calculating the marketing evaluation score corresponding to each user in the user list includes: Calculate the marketing evaluation scores corresponding to the users in the historical marketing user list according to the product purchase rate and asset management scale increase rate of the users in the historical marketing user list within a preset time after the marketing; The marketing evaluation scores corresponding to the users in the historical non-marketing user list are calculated based on the product purchase rate and asset management scale improvement rate of the users in the historical non-marketing user list within a preset time without marketing.

5. The method for distributing marketing strategies according to claim 2, characterized in that: The method of training the marketing strategy data set of the user using the causal forest algorithm to obtain the preset marketing strategy evaluation model includes: constructing a plurality of data subsets according to the marketing strategy dataset of the user; Each subset is split according to a preset splitting criterion, wherein the preset splitting criterion includes calculating the difference between the maximum value of the marketing evaluation scores corresponding to the two nodes after the splitting and the maximum value of the marketing evaluation scores corresponding to the nodes before the splitting; When the splitting stop condition is met, the splitting is stopped and multiple causal trees are obtained as the preset marketing strategy evaluation model.

6. The method for distributing marketing strategies according to claim 1, characterized in that: The multi-option knapsack problem is constructed and solved according to the conversion rate gain of the user under different marketing strategies and the preset marketing budget to obtain the target marketing strategy corresponding to the user, including: Construct multiple constraints for the multi-option knapsack problem based on the user's conversion rate gain under different marketing strategies and the preset marketing budget; Based on multiple constraints of the multi-option knapsack problem, the multi-option knapsack problem is solved using a solving algorithm to obtain a target marketing strategy corresponding to the user.

7. The method for distributing marketing strategies according to claim 6, characterized in that: The multiple constraints for constructing a multi-option knapsack problem based on the conversion rate gain of users under different marketing strategies and the preset marketing budget include: Constructing a constraint condition for maximizing the conversion rate gain according to the conversion rate gain of the user under different marketing strategies and a marketing strategy distribution matrix; According to the single-user marketing cost of each marketing strategy, the marketing strategy distribution matrix and the preset marketing budget, construct the constraint conditions of the preset marketing budget; A constraint condition on the number of marketing strategies distributed to a single user is constructed according to the marketing strategy distribution matrix.

8. A marketing strategy distribution device, characterized in that: The distribution device of the marketing strategy includes: An acquisition unit, used for acquiring characteristic data of a user; A prediction unit, configured to predict the conversion rate gain of the user under different marketing strategies using a preset marketing strategy evaluation model according to the characteristic data of the user; A construction and solution unit, used to construct and solve a multi-option knapsack problem according to the conversion rate gain of the user under different marketing strategies and a preset marketing budget, so as to obtain a target marketing strategy corresponding to the user; The distribution unit is used to distribute the target marketing strategy corresponding to the user to the user.

9. An electronic device, comprising: processor; as well as A memory arranged to store computer executable instructions, which when executed cause the processor to perform the method for distributing the marketing strategy of any one of claims 1 to 7.

10. A computer program product, comprising a computer program or instructions, wherein when the computer program or instructions are executed by a processor, the method for distributing the marketing strategy of any one of claims 1 to 7 is implemented.

Images