Product recommendation scheme determination method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a product recommendation scheme determination method and device, electronic equipment and a storage medium, and relates to the technical field of financial science and technology. The method comprises the steps of obtaining a type and an evaluation index of a target object recommended by a to-be-recommended product in response to first input; according to the types of the target objects, historical behavior data of the multiple target objects and historical identity data of the multiple target objects are obtained; based on the historical behavior data of the plurality of target objects and the historical identity data of the plurality of target objects, evaluating a plurality of to-be-selected recommendation schemes obtained in advance by adopting an evaluation index to obtain a target prediction value corresponding to each to-be-selected recommendation scheme; and screening the plurality of to-be-selected recommendation schemes according to the target predicted values corresponding to the plurality of to-be-selected recommendation schemes to obtain a target recommendation scheme. According to the embodiment of the invention, the evaluation efficiency of the product recommendation scheme can be improved.

Description

Technical Field

[0001] The present application relates to the field of fintech, and particularly to a method, device, electronic device and storage medium for determining a product recommendation scheme. Background Art

[0002] When enterprises and individuals quickly develop their businesses and promote and market their products, they often need to adopt various evaluation methods to analyze the advantages and disadvantages of recommendation schemes. Currently, the advantages and disadvantages of recommendation schemes are mainly analyzed by simulating the application scenarios of the recommendation schemes. However, in the process of product marketing, there are interferences from external factors (such as seasonal fluctuations, marketing activities or competitor behaviors) and internal factors (such as insufficient sample size or subjective judgment errors of marketers). When testing multiple variables simultaneously, it is difficult to determine which variable causes the change in the result, resulting in low evaluation efficiency of the product recommendation scheme. Summary of the Invention

[0003] The main purpose of the embodiments of the present application is to propose a method, device, electronic device and storage medium for determining a product recommendation scheme, aiming to improve the evaluation efficiency of the product recommendation scheme.

[0004] To achieve the above object, in the first aspect of the embodiments of the present application, a method for determining a product recommendation scheme is proposed. The method includes:

[0005] In response to a first input, obtaining the type of the target object to whom the product to be recommended is recommended and an evaluation index, where the evaluation index is used to evaluate the effect of recommending the product to be recommended to the target object;

[0006] According to the type of the target object, obtaining the historical behavior data of multiple target objects and the historical identity data of multiple target objects;

[0007] Based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects, using the evaluation index to evaluate multiple pre-obtained candidate recommendation schemes, and obtaining a target prediction value corresponding to each candidate recommendation scheme, where the target prediction value includes at least one of a predicted conversion rate, a predicted purchase rate, a predicted number of visits, and a predicted long-term user retention rate;

[0008] Screening the multiple candidate recommendation schemes according to the target prediction values corresponding to the multiple candidate recommendation schemes to obtain a target recommendation scheme.

[0009] In some embodiments, the screening the multiple candidate recommendation schemes according to the target prediction values corresponding to the multiple candidate recommendation schemes to obtain a target recommendation scheme includes:

[0010] If the target prediction value includes the predicted value of the conversion rate, the predicted value of the purchase rate, the predicted value of the number of visits, or the predicted value of the long-term user retention rate, then the candidate recommendation plan corresponding to the largest target prediction value among the target prediction values corresponding to multiple candidate recommendation plans is used as the target recommendation plan;

[0011] Or,

[0012] If the target prediction value includes the predicted value of the conversion rate, the predicted value of the purchase rate, the predicted value of the number of visits, or the predicted value of the long-term user retention rate, then the candidate recommendation plan corresponding to the target prediction value greater than the preset value among the target prediction values corresponding to multiple candidate recommendation plans is used as the target recommendation plan.

[0013] In some embodiments, the screening of multiple candidate recommendation plans according to the target prediction values corresponding to the multiple candidate recommendation plans to obtain the target recommendation plan includes:

[0014] If the target prediction value includes at least N dimensions among the predicted value of the conversion rate, the predicted value of the purchase rate, the predicted value of the number of visits, and the predicted value of the long-term user retention rate, then for each candidate recommendation plan, the predicted values included in the target prediction value corresponding to the candidate recommendation plan are weighted and summed to obtain the weighted score corresponding to the candidate recommendation plan, where N is an integer greater than or equal to 2 and less than or equal to 4;

[0015] The candidate recommendation plan corresponding to the largest weighted score among the weighted scores corresponding to the multiple candidate recommendation plans is used as the target recommendation plan;

[0016] Or,

[0017] If the target prediction value includes at least N dimensions among the predicted value of the conversion rate, the predicted value of the purchase rate, the predicted value of the number of visits, and the predicted value of the long-term user retention rate, then for each dimension, the candidate recommendation plan corresponding to the largest predicted value in the dimension is used as the intermediate recommendation plan, and N intermediate recommendation plans are obtained;

[0018] In response to the second input, the target recommendation plan among the N intermediate recommendation plans is determined, and the second input is used to select the target recommendation plan.

[0019] In some embodiments, the evaluating of multiple candidate recommendation plans obtained in advance by using the evaluation index based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects to obtain the target prediction value corresponding to each candidate recommendation plan includes:

[0020] Evaluate the multiple candidate recommendation solutions using the evaluation metrics based on the historical behavior data of the multiple target objects and the historical identity data of the multiple target objects through a pre-trained solution evaluation model, and obtain the target prediction value corresponding to each candidate recommendation solution;

[0021] Among them, the solution evaluation model is trained through multiple training data, and the multiple training data include the historical behavior data of multiple reference objects, the historical identity data of the multiple reference objects, the multiple candidate recommendation solutions, and the actual values of the conversion rate, the actual value of the purchase rate, the actual value of the number of visits, and the actual value of the long-term user retention rate corresponding to each candidate recommendation solution.

[0022] In some embodiments, evaluating the multiple candidate recommendation solutions using the evaluation metrics based on the historical behavior data of the multiple target objects and the historical identity data of the multiple target objects through a pre-trained solution evaluation model, and obtaining the target prediction value corresponding to each candidate recommendation solution includes:

[0023] Evaluate the multiple candidate recommendation solutions using the evaluation metrics based on the historical behavior data of the multiple target objects and the historical identity data of the multiple target objects through the solution evaluation model, and obtain the initial prediction value corresponding to each candidate recommendation solution and the confidence level corresponding to the initial prediction value;

[0024] If the confidence level is less than the preset confidence level, adjust the type of the target object, and jump to the step of obtaining the historical behavior data of the multiple target objects and the historical identity data of the multiple target objects according to the type of the target object for execution until the confidence level is greater than or equal to the preset confidence level;

[0025] If the confidence level is greater than or equal to the preset confidence level, use the initial prediction value corresponding to each candidate recommendation solution as the target prediction value corresponding to each candidate recommendation solution.

[0026] In some embodiments, obtaining the historical behavior data of the multiple target objects and the historical identity data of the multiple target objects according to the type of the target object includes:

[0027] Determine multiple initial objects according to the type of the target object;

[0028] Obtain the recommended product white list of the multiple initial objects;

[0029] For each initial object, when the type of the product in the recommended product white list of the initial object includes the type of the product to be recommended, determine the initial object as the target object;

[0030] Obtain the historical behavior data of multiple said target objects and the historical identity data of multiple said target objects.

[0031] In some embodiments, the historical behavior data includes at least one of the historical purchase information of the target object, the historical access information of the target object, and the historical search information of the target object, and the historical identity data includes at least one of the age of the target object, the gender of the target object, and the region where the target object is located.

[0032] To achieve the above object, a second aspect of the embodiments of the present application proposes a device for determining a product recommendation scheme, the device includes:

[0033] A response module, configured to, in response to a first input, obtain the type of the target object to which the product to be recommended is recommended and an evaluation index, where the evaluation index is used to evaluate the effect of recommending the product to be recommended to the target object;

[0034] An acquisition module, configured to obtain the historical behavior data of multiple target objects and the historical identity data of multiple said target objects according to the type of the target object;

[0035] An evaluation module, configured to, based on the historical behavior data of multiple said target objects and the historical identity data of multiple said target objects, use the evaluation index to evaluate multiple pre-obtained candidate recommendation schemes, and obtain a target prediction value corresponding to each said candidate recommendation scheme, where the target prediction value includes at least one of a predicted value of conversion rate, a predicted value of purchase rate, a predicted value of access times, and a predicted value of user long-term retention rate;

[0036] A screening module, configured to screen multiple said candidate recommendation schemes according to the target prediction values corresponding to multiple said candidate recommendation schemes, and obtain a target recommendation scheme.

[0037] To achieve the above object, a third aspect of the embodiments of the present application proposes an electronic device, the electronic device includes a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the product recommendation scheme determination method described in the first aspect above is implemented.

[0038] To achieve the above object, a fourth aspect of the embodiments of the present application proposes a computer-readable storage medium, the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the product recommendation scheme determination method described in the first aspect above is implemented.

[0039] The method, device, electronic device and storage medium for determining a product recommendation solution proposed in this application obtain the type and evaluation indicators of the target object for which the product to be recommended is recommended in response to a first input, and then obtain the historical behavior data of multiple target objects and the historical identity data of multiple target objects according to the type of the target object. Subsequently, based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects, the evaluation indicators are used to evaluate multiple pre-obtained candidate recommendation solutions to obtain the target prediction value corresponding to each candidate recommendation solution. Finally, multiple candidate recommendation solutions are screened according to the target prediction values corresponding to multiple candidate recommendation solutions to obtain the target recommendation solution. Through the above steps, by evaluating multiple candidate recommendation solutions and then determining the target recommendation solution according to the target prediction value corresponding to each candidate recommendation solution obtained by the evaluation, the recommendation solution with better recommendation effect among multiple candidate recommendation solutions can be determined quickly, and the evaluation efficiency of the product recommendation solution is improved. Description of the Drawings

[0040] Figure 1 is a flowchart of the method for determining a product recommendation solution provided by an embodiment of this application;

[0041] Figure 2 is Figure 1 the flowchart of step S102 in

[0042] Figure 3 is another flowchart of the method for determining a product recommendation solution provided by an embodiment of this application;

[0043] Figure 4 is yet another flowchart of the method for determining a product recommendation solution provided by an embodiment of this application;

[0044] Figure 5 is a schematic structural diagram of the device for determining a product recommendation solution provided by an embodiment of this application;

[0045] Figure 6 is a schematic hardware structure diagram of the electronic device provided by an embodiment of this application. Detailed Embodiments

[0046] In order to make the purpose, technical solutions and advantages of this application clearer, the following further describes this application in detail with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not used to limit this application.

[0047] It should be noted that although the functional modules are divided in the schematic diagram of the device and the logical sequence is shown in the flowchart, in some cases, the steps shown or described can be executed in a different module division from that in the device or a different sequence from that in the flowchart. Terms such as "first" and "second" in the specification, claims and the above-mentioned drawings are used to distinguish similar objects and do not necessarily describe a specific order or sequence.

[0048] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0049] First, several terms involved in this application are analyzed as follows:

[0050] Product to be recommended: refers to a product that has not been recommended to the user but has the possibility of being recommended to the user. The product to be recommended can be a financial product, such as deposit financial products, loan financial products, insurance products, and wealth management products, etc.

[0051] White list: A white list refers to a list of entities approved or trusted by the user. Entities in the white list are considered safe and can be allowed to access specific resources, services, or perform certain operations.

[0052] To solve the problems of the prior art, the embodiments of this application provide a method, device, electronic device, and storage medium for determining a product recommendation scheme, aiming to improve the evaluation efficiency of the product recommendation scheme.

[0053] The method, device, electronic device, and storage medium for determining a product recommendation scheme provided by the embodiments of this application are specifically described through the following embodiments. First, the method for determining a product recommendation scheme in the embodiments of this application is described.

[0054] The embodiments of this application can obtain and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is to use a digital computer or a machine controlled by a digital computer to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to obtain the best results of theory, method, technology, and application system.

[0055] Artificial intelligence basic technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, and mechatronics. Artificial intelligence software technologies mainly include several major directions such as computer vision technology, robotics, biometric technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0056] The product recommendation scheme determination method provided by the embodiments of this application relates to the field of fintech. The product recommendation scheme determination method provided by the embodiments of this application can be applied to a terminal, or to a server side, or can be software running on a terminal or a server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, or can be configured as a server cluster or a distributed system composed of multiple physical servers, or can be configured as a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks, and big data and artificial intelligence platforms; the software can be an application that implements the product recommendation scheme determination method, etc., but is not limited to the above forms.

[0057] This application can be used in many general or specific computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics devices, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and so on. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. This application can also be practiced in a distributed computing environment, where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0058] It should be noted that in each specific embodiment of this application, when it comes to relevant processing that needs to be carried out based on data related to the user's identity or characteristics such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first. Moreover, the collection, use, and processing of these data will comply with relevant laws, regulations, and standards. In addition, when the embodiments of this application need to obtain the user's sensitive personal information, the user's separate permission or separate consent will be obtained through methods such as pop-up windows or jumping to a confirmation page. After clearly obtaining the user's separate permission or separate consent, the necessary user-related data for the normal operation of the embodiments of this application will be obtained.

[0059] Figure 1 is the flowchart of the product recommendation scheme determination method provided by the embodiments of this application. Please refer to Figure 1。The product recommendation scheme determination method provided by the embodiments of the present application can be applied to an electronic device. Figure 1 The method in [[ ]] may include but is not limited to steps S101 to S104.

[0060] Step S101: In response to a first input, obtain the type of the target object for which the product to be recommended is recommended and an evaluation index, where the evaluation index is used to evaluate the effect of recommending the product to be recommended to the target object;

[0061] Step S102: According to the type of the target object, obtain the historical behavior data of multiple target objects and the historical identity data of multiple target objects;

[0062] Step S103: Based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects, use the evaluation index to evaluate multiple pre-obtained candidate recommendation schemes, and obtain a target prediction value corresponding to each candidate recommendation scheme, where the target prediction value includes at least one of a predicted value of conversion rate, a predicted value of purchase rate, a predicted value of the number of visits, and a predicted value of the long-term user retention rate;

[0063] Step S104: Screen the multiple candidate recommendation schemes according to the target prediction values corresponding to the multiple candidate recommendation schemes to obtain a target recommendation scheme.

[0064] Specifically, the first input includes the type of the target object and the evaluation index. In response to the first input, the type of the target object for which the product to be recommended is recommended and the evaluation index can be obtained. The product to be recommended is a financial product, such as deposit financial products, loan financial products, insurance products, and wealth management products, etc. The type of the target object refers to the type of users to whom the product to be recommended needs to be recommended. How to specifically divide the type of the target object can be determined according to the actual situation. For example, it can be divided according to gender, divided according to income level, or classified according to user behavior, or divided in other ways, which is not limited here. The evaluation index can be at least one of conversion rate, purchase rate, the number of visits, and the long-term user retention rate. Among them, the conversion rate is the ratio of the number of target objects who complete registration to the total number of target objects recommended. The purchase rate is the ratio of the number of target objects who complete purchase to the total number of target objects recommended. The number of visits refers to the number of times the target object views the information related to the product to be recommended. The long-term user retention rate is the ratio of the target objects who maintain activity and usage frequency within a preset time period to the total number of target objects recommended.

[0065] According to the types of target objects, historical behavior data of multiple target objects and historical identity data of multiple target objects can be obtained. Specifically, after obtaining relevant data of a user (including historical behavior data and historical identity data), the relevant data of the user can be labeled using classification tags to determine the type to which the user belongs, and the relevant data of the user can be classified and stored in a database according to the type to which the user belongs. In this way, after obtaining the type of the target object, the type of the target object can be matched with the types stored in the database to determine multiple users corresponding to the successfully matched type, that is, multiple target objects, and then the historical behavior data of multiple target objects and the historical identity data of multiple target objects stored in the database can be obtained. In one example, the historical behavior data may include at least one of the historical purchase information of the target object, the historical access information of the target object, and the historical search information of the target object, and the historical identity data may include at least one of the age of the target object, the gender of the target object, and the region where the target object is located.

[0066] Moreover, multiple candidate recommendation schemes can be obtained in advance. For example, the multiple candidate recommendation schemes may include a scheme for recommending a product to be recommended by phone, a scheme for recommending a product to be recommended by advertising, and a scheme for recommending a product to be recommended by online promotion. Based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects, multiple candidate recommendation schemes can be evaluated using evaluation metrics to obtain the target prediction values corresponding to each candidate recommendation scheme. The target prediction values include at least one of the predicted value of the conversion rate, the predicted value of the purchase rate, the predicted value of the number of visits, and the predicted value of the long-term user retention rate. In one example, if the evaluation metric is the conversion rate, multiple target objects can be divided into three groups, and based on the historical behavior data and historical identity data of one group of target objects, the conversion rate of a candidate recommendation scheme can be predicted by looking up a table or through a pre-constructed scheme evaluation model, so as to obtain the predicted value of the conversion rate corresponding to the candidate recommendation scheme. Or, if the evaluation metrics are the conversion rate and the number of visits, multiple target objects can be divided into three groups, and based on the historical behavior data and historical identity data of one group of target objects, the conversion rate and the number of visits of a candidate recommendation scheme can be predicted by looking up a table or through a scheme evaluation model to obtain the predicted value of the conversion rate and the predicted value of the number of visits corresponding to the candidate recommendation scheme. In this way, the evaluation of each candidate recommendation scheme can be completed to obtain the target prediction values corresponding to each candidate recommendation scheme.

[0067] Subsequently, based on the target prediction values corresponding to multiple candidate recommendation solutions, the multiple candidate recommendation solutions can be screened to obtain the target recommendation solution. In one example, the multiple candidate recommendation solutions include candidate recommendation solution 1, candidate recommendation solution 2, and candidate recommendation solution 3, and the evaluation metric is the conversion rate. At this time, the target prediction values include the predicted values of the conversion rate. After comparing the predicted values of the conversion rate corresponding to the multiple candidate recommendation solutions, the largest predicted value of the conversion rate (e.g., 20%) among the predicted values of the conversion rate corresponding to the multiple candidate recommendation solutions can be determined, and the candidate recommendation solution corresponding to the largest predicted value of the conversion rate (e.g., candidate recommendation solution 2) is determined as the target recommendation solution.

[0068] Steps S101 to S104 illustrated in the embodiments of the present application, by responding to the first input, obtain the type of the target object for which the product to be recommended is recommended and the evaluation metric, and then according to the type of the target object, obtain the historical behavior data of multiple target objects and the historical identity data of multiple target objects. Subsequently, based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects, use the evaluation metric to evaluate multiple candidate recommendation solutions obtained in advance to obtain the target prediction value corresponding to each candidate recommendation solution. Finally, screen the multiple candidate recommendation solutions according to the target prediction values corresponding to the multiple candidate recommendation solutions to obtain the target recommendation solution. The above steps can quickly determine the recommendation solution with better recommendation effect among the multiple candidate recommendation solutions by evaluating the multiple candidate recommendation solutions and then determining the target recommendation solution according to the target prediction value corresponding to each candidate recommendation solution obtained by the evaluation, improving the evaluation efficiency of the product recommendation solution.

[0069] In addition, an impact analysis report can also be output. The impact analysis report includes the type of the target object, the evaluation metric, and the target prediction value corresponding to each candidate recommendation solution for subsequent data comparison and analysis.

[0070] Figure 2 Yes Figure 1 is the flowchart of step S102 in Figure 2 . In some embodiments, step S102, according to the type of the target object, obtain the historical behavior data of multiple target objects and the historical identity data of multiple target objects, includes:

[0071] Step S201, determine multiple initial objects according to the type of the target object;

[0072] Step S202, obtain the recommended product white list of multiple initial objects;

[0073] Step S203: For each of the initial objects, when the types of the products in the recommended product white list of the initial object include the type of the product to be recommended, determine the initial object as the target object.

[0074] Step S204: Obtain the historical behavior data of multiple target objects and the historical identity data of multiple target objects.

[0075] According to the type of the target object, the historical behavior data of multiple target objects and the historical identity data of multiple target objects can be obtained. Specifically, according to the type of the target object, the type of the target object can be matched with the types stored in the database to determine multiple users corresponding to the successfully matched types, that is, multiple initial objects. Subsequently, the recommended product white lists of multiple initial objects can be obtained, that is, the lists of the types of products that multiple initial objects allow to be recommended to them. And the type of the product to be recommended can be determined, such as deposit type, loan type, insurance type or other types. For each initial object, when the types of the products in the recommended product white list of the initial object include the type of the product to be recommended, it is determined at this time that the product to be recommended can be recommended to the initial object, so the initial object is determined as the target object. In this way, multiple target objects can be determined from multiple initial objects, and then the historical behavior data of multiple target objects and the historical identity data of multiple target objects can be obtained. Through the above steps, according to the type of the target object, the historical behavior data of multiple target objects and the historical identity data of multiple target objects can be obtained to evaluate multiple candidate recommendation schemes based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects subsequently.

[0076] In some embodiments, the historical behavior data includes at least one of the historical purchase information of the target object, the historical access information of the target object, and the historical search information of the target object, and the historical identity data includes at least one of the age of the target object, the gender of the target object, and the region where the target object is located.

[0077] Specifically, the historical behavior data may include at least one of the historical purchase information of the target object, the historical access information of the target object, and the historical search information of the target object. The historical identity data includes at least one of the age of the target object, the gender of the target object, and the region where the target object is located. Among them, the historical purchase information of the target object refers to the information related to the purchase behavior of the target object within the historical time period, such as product information, transaction information, and promotion information, etc. The historical access information of the target object refers to the information related to the access behavior of the target object within the historical time period, which is usually used to describe the interaction between the target object and resources such as websites, applications, databases, servers, etc. The historical access information of the target object may include the access behavior data of the target object within the historical time period (such as access time, access page, stay duration, and click stream data, etc.), access source data, and interaction data. The historical search information of the target object refers to the information related to the search behavior of the target object within the historical time period (such as search terms, search time, and search result click behavior, etc.), which can be used to assist in understanding the needs and behaviors of the target object. Based on the foregoing historical behavior data and historical identity data, the effect of recommending the product to be recommended using the candidate recommendation scheme can be evaluated.

[0078] In some embodiments, step S103, based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects, using the evaluation index to evaluate multiple candidate recommendation schemes obtained in advance, and obtaining the target prediction value corresponding to each candidate recommendation scheme, includes:

[0079] Through a pre-trained scheme evaluation model, based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects, using the evaluation index to evaluate multiple candidate recommendation schemes, and obtaining the target prediction value corresponding to each candidate recommendation scheme;

[0080] Among them, the scheme evaluation model is trained through multiple training data. The multiple training data include the historical behavior data of multiple reference objects, the historical identity data of multiple reference objects, multiple candidate recommendation schemes, and the actual values of the conversion rate, purchase rate, number of accesses, and long-term user retention rate corresponding to each candidate recommendation scheme.

[0081] Based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects, evaluation metrics can be used to evaluate multiple candidate recommendation schemes, so as to obtain the target prediction value corresponding to each candidate recommendation scheme. Specifically, an initial evaluation model can be trained in advance through multiple training data, that is, the initial evaluation model is used to classify multiple training data, and the initial evaluation model is fine-tuned according to the output data of the initial evaluation model to obtain a scheme evaluation model. The multiple training data are data labeled with classification labels, and the multiple training data include the historical behavior data of multiple reference objects, the historical identity data of multiple reference objects, multiple candidate recommendation schemes, and the actual values of the conversion rate, purchase rate, number of visits, and long-term user retention rate corresponding to each candidate recommendation scheme. In addition, the multiple training data may also include other recommendation schemes other than the multiple candidate recommendation schemes, and the actual values of the conversion rate, purchase rate, number of visits, and long-term user retention rate corresponding to the other recommendation schemes. In this way, the scheme evaluation model can be applied to scenarios where other recommendation schemes are evaluated.

[0082] After the scheme evaluation model is constructed, the historical behavior data of multiple target objects, the historical identity data of multiple target objects, the evaluation metrics, and multiple candidate recommendation schemes can be input into the scheme evaluation model, so as to obtain the target prediction value corresponding to each candidate recommendation scheme. In one example, if the evaluation metric is the conversion rate, through the scheme evaluation model, based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects, the predicted value of the conversion rate corresponding to each candidate recommendation scheme can be determined. In another example, if the evaluation metric is the long-term user retention rate, through the scheme evaluation model, based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects, the predicted value of the long-term user retention rate corresponding to each candidate recommendation scheme can be determined. Through the above steps, the target prediction value corresponding to each candidate recommendation scheme can be quickly determined by using the scheme evaluation model, thereby further improving the evaluation efficiency of the product recommendation scheme.

[0083] Figure 3 is another flowchart of the product recommendation scheme determination method provided by the embodiments of the present application. Please refer to Figure 3 . In some embodiments, the step of evaluating multiple candidate recommendation schemes by using the evaluation metrics based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects through a pre-trained scheme evaluation model to obtain the target prediction value corresponding to each candidate recommendation scheme includes:

[0084] Step S301, in response to a first input, obtain the type of the target object for which the product to be recommended is recommended and the evaluation metrics, where the evaluation metrics are used to evaluate the effect of recommending the product to be recommended to the target object;

[0085] Step S302, according to the type of the target object, obtain the historical behavior data of multiple target objects and the historical identity data of multiple target objects;

[0086] Step S303, through the solution evaluation model, based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects, use the evaluation metrics to evaluate multiple candidate recommendation solutions, and obtain the initial prediction value corresponding to each candidate recommendation solution and the confidence level corresponding to the initial prediction value;

[0087] Step S304, if the confidence level is less than the preset confidence level, adjust the type of the target object, and jump to the step of obtaining the historical behavior data of multiple target objects and the historical identity data of multiple target objects according to the type of the target object for execution, until the confidence level is greater than or equal to the preset confidence level;

[0088] Step S305, if the confidence level is greater than or equal to the preset confidence level, use the initial prediction value corresponding to each candidate recommendation solution as the target prediction value corresponding to each candidate recommendation solution.

[0089] Step S306, screen multiple candidate recommendation solutions according to the target prediction values corresponding to multiple candidate recommendation solutions to obtain a target recommendation solution.

[0090] For the specific details of the above Step S301, Step S302, and Step S306, please refer to Step S101, Step S102, and Step S104 respectively, which will not be elaborated here.

[0091] Through a solution evaluation model, based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects, evaluation metrics can be used to evaluate multiple candidate recommendation solutions, so as to obtain the target prediction value corresponding to each candidate recommendation solution. Specifically, the historical behavior data of multiple target objects, the historical identity data of multiple target objects, the evaluation metrics, and multiple candidate recommendation solutions can be input into the solution evaluation model, so as to obtain the initial prediction value corresponding to each candidate recommendation solution and the confidence level corresponding to the initial prediction value. After obtaining the confidence level corresponding to the initial prediction value, the confidence level corresponding to the initial prediction value can be compared with the preset confidence level. If the confidence level corresponding to any one of the initial prediction values corresponding to multiple candidate recommendation solutions is less than the preset confidence level, the type of the target object is adjusted, and the steps of obtaining the historical behavior data of multiple target objects and the historical identity data of multiple target objects according to the type of the target object are re-executed until the confidence levels corresponding to all the initial prediction values corresponding to multiple candidate recommendation solutions are greater than or equal to the preset confidence level. If the confidence levels corresponding to all the initial prediction values corresponding to multiple candidate recommendation solutions are greater than or equal to the preset confidence level, the initial prediction value corresponding to each candidate recommendation solution is used as the target prediction value corresponding to each candidate recommendation solution. In this way, the target prediction value corresponding to each candidate recommendation solution can be evaluated, which is convenient for subsequent screening of multiple candidate recommendation solutions based on the target prediction values corresponding to multiple candidate recommendation solutions.

[0092] In some embodiments, step S104 of screening multiple candidate recommendation solutions according to the target prediction values corresponding to the multiple candidate recommendation solutions to obtain a target recommendation solution includes:

[0093] If the target prediction value includes a predicted value of conversion rate, a predicted value of purchase rate, a predicted value of visit times, or a predicted value of user long-term retention rate, the candidate recommendation solution corresponding to the largest target prediction value among the target prediction values corresponding to the multiple candidate recommendation solutions is used as the target recommendation solution;

[0094] Or,

[0095] If the target prediction value includes a predicted value of conversion rate, a predicted value of purchase rate, a predicted value of visit times, or a predicted value of user long-term retention rate, the candidate recommendation solution corresponding to the target prediction value greater than the preset value among the target prediction values corresponding to the multiple candidate recommendation solutions is used as the target recommendation solution.

[0096] When the target predicted value includes any one of the predicted value of conversion rate, the predicted value of purchase rate, the predicted value of number of visits, and the predicted value of long-term user retention rate, according to the target predicted values corresponding to multiple candidate recommendation schemes, two methods can be adopted to screen the multiple candidate recommendation schemes to obtain the target recommendation scheme. Specifically, the target predicted values corresponding to the multiple candidate recommendation schemes can be compared, and the candidate recommendation scheme corresponding to the largest target predicted value among the target predicted values corresponding to the multiple candidate recommendation schemes can be used as the target recommendation scheme. Or, the target predicted values corresponding to the multiple candidate recommendation schemes can be compared with a preset value, and the candidate recommendation scheme corresponding to the target predicted value greater than the preset value can be used as the target recommendation scheme. In this way, when the target predicted value only includes any one of the predicted value of conversion rate, the predicted value of purchase rate, the predicted value of number of visits, and the predicted value of long-term user retention rate, the target recommendation scheme can be screened out from the multiple candidate recommendation schemes.

[0097] Figure 4 is another flowchart of the product recommendation scheme determination method provided by the embodiments of the present application. Please refer to Figure 4 . In some embodiments, step S104, screening the multiple candidate recommendation schemes according to the target predicted values corresponding to the multiple candidate recommendation schemes to obtain the target recommendation scheme, includes:

[0098] Step S401, in response to a first input, obtain the type of the target object to which the product to be recommended is recommended and the evaluation index, where the evaluation index is used to evaluate the effect of recommending the product to be recommended to the target object;

[0099] Step S402, according to the type of the target object, obtain the historical behavior data of multiple target objects and the historical identity data of the multiple target objects;

[0100] Step S403, based on the historical behavior data of the multiple target objects and the historical identity data of the multiple target objects, use the evaluation index to evaluate the multiple candidate recommendation schemes obtained in advance to obtain the target predicted value corresponding to each candidate recommendation scheme, where the target predicted value includes at least one of the predicted value of conversion rate, the predicted value of purchase rate, the predicted value of number of visits, and the predicted value of long-term user retention rate;

[0101] For the specific implementation of the above steps S401 - S403, please refer to steps S101 - S103, which will not be elaborated here.

[0102] Step S404, if the target prediction value includes at least N dimensions among the predicted value of conversion rate, the predicted value of purchase rate, the predicted value of number of visits, and the predicted value of long-term user retention rate, then for each of the candidate recommendation schemes, perform a weighted sum calculation on the predicted values included in the target prediction value corresponding to the candidate recommendation scheme to obtain the weighted score corresponding to the candidate recommendation scheme, where N is an integer greater than or equal to 2 and less than or equal to 4;

[0103] Step S405, take the candidate recommendation scheme corresponding to the maximum weighted score among the weighted scores corresponding to multiple candidate recommendation schemes as the target recommendation scheme;

[0104] Or,

[0105] Step S406, if the target prediction value includes at least N dimensions among the predicted value of conversion rate, the predicted value of purchase rate, the predicted value of number of visits, and the predicted value of long-term user retention rate, then for each of the dimensions, take the candidate recommendation scheme corresponding to the maximum predicted value in the dimension as the intermediate recommendation scheme to obtain N intermediate recommendation schemes;

[0106] Step S407, in response to the second input, determine the target recommendation scheme among the N intermediate recommendation schemes, where the second input is used to select the target recommendation scheme.

[0107] When the target prediction value includes at least N dimensions among the predicted value of conversion rate, the predicted value of purchase rate, the predicted value of number of visits, and the predicted value of long-term user retention rate, according to the target prediction values corresponding to multiple candidate recommendation schemes, two methods can be used to screen multiple candidate recommendation schemes to obtain the target recommendation scheme. Among them, N is an integer greater than or equal to 2 and less than or equal to 4. Specifically, for each candidate recommendation scheme, a weighted sum calculation can be performed on the predicted values included in the target prediction value corresponding to the candidate recommendation scheme to obtain the weighted score corresponding to the candidate recommendation scheme. In one example, if the target prediction value includes the predicted value of conversion rate, the predicted value of purchase rate, and the predicted value of number of visits, then for each candidate recommendation scheme, multiply the predicted value of conversion rate corresponding to the candidate recommendation scheme by the first weight corresponding to the conversion rate to obtain the first operation result, multiply the predicted value of purchase rate corresponding to the candidate recommendation scheme by the second weight corresponding to the purchase rate to obtain the second operation result, and multiply the predicted value of number of visits corresponding to the candidate recommendation scheme by the third weight corresponding to the number of visits to obtain the third operation result, and add the first operation result, the second operation result, and the third operation result to obtain the weighted score corresponding to the candidate recommendation scheme. After determining the weighted score corresponding to each candidate recommendation scheme, take the candidate recommendation scheme corresponding to the maximum weighted score among the weighted scores corresponding to multiple candidate recommendation schemes as the target recommendation scheme.

[0108] Alternatively, in each dimension, the candidate recommendation plan corresponding to the largest predicted value in each dimension is used as the intermediate recommendation plan, and N intermediate recommendation plans are obtained. In another example, if the target predicted value includes two dimensions, namely the predicted value of the conversion rate and the predicted value of the purchase rate, then, in the dimension of the predicted value of the conversion rate, the candidate recommendation plan corresponding to the largest predicted value of the conversion rate can be used as an intermediate recommendation plan. Similarly, in the dimension of the predicted value of the purchase rate, the candidate recommendation plan corresponding to the largest predicted value of the purchase rate can be used as an intermediate recommendation plan. It should be noted that in any dimension, if there are M candidate recommendation plans with equal predicted values, and the predicted values of the M candidate recommendation plans are the largest predicted values in the dimension, at this time, these M candidate recommendation plans can all be used as intermediate recommendation plans. The second input is used to select the target recommendation plan. After obtaining N intermediate recommendation plans, the second input can be received, and in response to the second input, the target recommendation plan among the N intermediate recommendation plans can be determined.

[0109] Through the above steps, when the target predicted value includes at least N dimensions among the predicted value of the conversion rate, the predicted value of the purchase rate, the predicted value of the number of visits, and the predicted value of the long-term user retention rate, the target recommendation plan can be screened out from multiple candidate recommendation plans.

[0110] Figure 5 is a schematic structural diagram of the product recommendation plan determination device provided by the embodiments of the present application. Please refer to Figure 5 Embodiments of the present application further provide a product recommendation plan determination device 500, which can implement the above product recommendation plan determination method. The device 500 includes:

[0111] A response module 501, configured to obtain the type of the target object to which the product to be recommended is recommended and the evaluation index in response to the first input, where the evaluation index is used to evaluate the effect of recommending the product to be recommended to the target object;

[0112] An acquisition module 502, configured to obtain the historical behavior data of multiple target objects and the historical identity data of multiple target objects according to the type of the target object;

[0113] An evaluation module 503, configured to evaluate multiple candidate recommendation plans obtained in advance by using the evaluation index based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects, and obtain the target predicted value corresponding to each candidate recommendation plan, where the target predicted value includes at least one of the predicted value of the conversion rate, the predicted value of the purchase rate, the predicted value of the number of visits, and the predicted value of the long-term user retention rate;

[0114] A screening module 504, configured to screen multiple candidate recommendation schemes according to the target prediction values corresponding to the multiple candidate recommendation schemes, so as to obtain a target recommendation scheme.

[0115] In some embodiments, the screening module 504 includes:

[0116] A first screening sub-module, configured to, if the target prediction value includes a predicted value of conversion rate, a predicted value of purchase rate, a predicted value of number of visits, or a predicted value of long-term user retention rate, use the candidate recommendation scheme corresponding to the largest target prediction value among the target prediction values corresponding to the multiple candidate recommendation schemes as the target recommendation scheme;

[0117] Or,

[0118] A second screening sub-module, configured to, if the target prediction value includes a predicted value of conversion rate, a predicted value of purchase rate, a predicted value of number of visits, or a predicted value of long-term user retention rate, use the candidate recommendation scheme corresponding to the target prediction value greater than a preset value among the target prediction values corresponding to the multiple candidate recommendation schemes as the target recommendation scheme.

[0119] In some embodiments, the screening module 504 includes:

[0120] An operator sub-module, configured to, if the target prediction value includes at least N dimensions among the predicted value of conversion rate, the predicted value of purchase rate, the predicted value of number of visits, and the predicted value of long-term user retention rate, for each candidate recommendation scheme, perform a weighted sum calculation on the predicted values included in the target prediction value corresponding to the candidate recommendation scheme to obtain a weighted score corresponding to the candidate recommendation scheme, where N is an integer greater than or equal to 2 and less than or equal to 4;

[0121] A third screening sub-module, configured to use the candidate recommendation scheme corresponding to the largest weighted score among the weighted scores corresponding to the multiple candidate recommendation schemes as the target recommendation scheme;

[0122] Or,

[0123] A fourth screening sub-module, configured to, if the target prediction value includes at least N dimensions among the predicted value of conversion rate, the predicted value of purchase rate, the predicted value of number of visits, and the predicted value of long-term user retention rate, for each dimension, use the candidate recommendation scheme corresponding to the largest predicted value in the dimension as an intermediate recommendation scheme, to obtain N intermediate recommendation schemes;

[0124] A first determination sub-module, configured to, in response to a second input, determine a target recommendation scheme among the N intermediate recommendation schemes, where the second input is used to select the target recommendation scheme.

[0125] In some embodiments, the evaluation module 503 includes:

[0126] An evaluation sub-module, configured to evaluate multiple candidate recommendation schemes by using a pre-trained scheme evaluation model, based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects, and adopt the evaluation index to evaluate multiple candidate recommendation schemes, so as to obtain a target prediction value corresponding to each candidate recommendation scheme;

[0127] Wherein, the scheme evaluation model is trained by using multiple training data, and the multiple training data include the historical behavior data of multiple reference objects, the historical identity data of multiple reference objects, multiple candidate recommendation schemes, and the actual values of the conversion rate, the purchase rate, the number of visits, and the long-term user retention rate corresponding to each candidate recommendation scheme.

[0128] In some embodiments, the evaluation sub-module includes:

[0129] An evaluation unit, configured to evaluate multiple candidate recommendation schemes by using the scheme evaluation model, based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects, and adopt the evaluation index to evaluate multiple candidate recommendation schemes, so as to obtain an initial prediction value corresponding to each candidate recommendation scheme and the confidence level corresponding to the initial prediction value;

[0130] An adjustment unit, configured to, if the confidence level is less than a preset confidence level, adjust the type of the target object, and jump to the step of obtaining the historical behavior data of multiple target objects and the historical identity data of multiple target objects according to the type of the target object for execution, until the confidence level is greater than or equal to the preset confidence level;

[0131] A determination unit, configured to, if the confidence level is greater than or equal to the preset confidence level, use the initial prediction value corresponding to each candidate recommendation scheme as the target prediction value corresponding to each candidate recommendation scheme.

[0132] In some embodiments, the acquisition module 502 includes:

[0133] A second determination sub-module, configured to determine multiple initial objects according to the type of the target object;

[0134] A first acquisition sub-module, configured to acquire a recommended product white list of multiple initial objects;

[0135] A third determination sub-module, configured to, for each initial object, if the type of the product in the recommended product white list of the initial object includes the type of the product to be recommended, determine the initial object as the target object;

[0136] A second acquisition sub-module, configured to acquire historical behavior data of multiple said target objects and historical identity data of multiple said target objects.

[0137] For the specific implementation of the product recommendation solution determination device 500, reference may be made to the embodiments of the above-mentioned product recommendation solution determination method, which will not be elaborated here.

[0138] An embodiment of the present application further provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the above-mentioned product recommendation solution determination method. The electronic device can be any intelligent terminal including a tablet computer, an in-vehicle computer, etc.

[0139] Figure 6 is a schematic diagram of the hardware structure of the electronic device provided by the embodiment of the present application. Please refer to Figure 6 . The electronic device includes:

[0140] A processor 601, which can be implemented in ways such as a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided by the embodiments of the present application;

[0141] A memory 602, which can be implemented in forms such as a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 602 can store an operating system and other application programs. When implementing the technical solutions provided by the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 602, and the processor 601 is used to call and execute the product recommendation solution determination method of the embodiments of the present application;

[0142] An input / output interface 603, used to implement information input and output;

[0143] A communication interface 604, used to implement communication interaction between this device and other devices, and can implement communication through a wired method (such as USB, network cable, etc.) or through a wireless method (such as mobile network, WIFI, Bluetooth, etc.);

[0144] A bus 605, which transmits information between various components of the device (such as the processor 601, the memory 602, the input / output interface 603, and the communication interface 604);

[0145] Among them, the processor 601, the memory 602, the input / output interface 603, and the communication interface 604 are communicatively connected to each other inside the device through the bus 605.

[0146] An embodiment of the present application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-mentioned method for determining a product recommendation solution.

[0147] As a non-transitory computer-readable storage medium, the memory can be used to store non-transitory software programs and non-transitory computer-executable programs. In addition, the memory may include high-speed random access memory and may also include non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some embodiments, the memory may optionally include memories remotely provided with respect to the processor, and these remote memories can be connected to the processor through a network. Examples of the above-mentioned network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0148] The method, apparatus, electronic device, and storage medium for determining a product recommendation solution proposed in the present application obtain the type and evaluation metrics of the target object for which a product is to be recommended in response to a first input, then obtain the historical behavior data and historical identity data of multiple target objects according to the type of the target object, and then evaluate multiple pre-obtained candidate recommendation solutions based on the historical behavior data and historical identity data of the multiple target objects using the evaluation metrics to obtain a target prediction value corresponding to each candidate recommendation solution, and finally screen the multiple candidate recommendation solutions according to the target prediction values corresponding to the multiple candidate recommendation solutions to obtain a target recommendation solution. The above steps can quickly determine a recommendation solution with better recommendation effect among multiple candidate recommendation solutions by evaluating the multiple candidate recommendation solutions and then determining the target recommendation solution according to the target prediction value corresponding to each candidate recommendation solution obtained by the evaluation, thereby improving the evaluation efficiency of the product recommendation solution.

[0149] The embodiments described in the embodiments of the present application are for more clearly illustrating the technical solutions of the embodiments of the present application and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those skilled in the art will know that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of the present application are equally applicable to similar technical problems.

[0150] Those skilled in the art can understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application and may include more or fewer steps than those shown in the figures, or combine certain steps, or different steps.

[0151] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0152] Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, and their appropriate combinations.

[0153] As used in the specification of this application and the above-mentioned drawings, the terms "first", "second", "third", "fourth", etc. (if any) are used to distinguish similar objects and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances so that the embodiments of this application described here can be implemented in an order other than those illustrated or described here. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0154] It should be understood that in this application, "at least one (item)" means one or more, and "multiple" means two or more. "And / or" is used to describe the association relationship of associated objects and indicates that there can be three relationships. For example, "A and / or B" can mean: only A exists, only B exists, and both A and B exist at the same time. Among them, A and B can be singular or plural. The character " / " generally means that the associated objects before and after are in an "or" relationship. "At least one (one) of the following" or a similar expression means any combination of these items, including any combination of single items (ones) or plural items (ones). For example, at least one (one) of a, b, or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0155] In several embodiments provided by the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the above-mentioned units is only a logical function division. In actual implementation, there may be other division methods. 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 displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of devices or units can be in electrical, mechanical or other forms.

[0156] The units described above as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0157] In addition, in each embodiment of the present application, each functional unit can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0158] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in each embodiment of the present application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks, or optical discs that can store programs.

[0159] The preferred embodiments of the embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of rights of the embodiments of the present application. Any modifications, equivalent replacements, and improvements made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall be within the scope of rights of the embodiments of the present application.

Claims

1. A method for determining a product recommendation scheme, characterized in that The method includes: In response to a first input, obtaining the type of the target object for which the product to be recommended is recommended and the evaluation metrics, where the evaluation metrics are used to evaluate the effect of recommending the product to be recommended to the target object; According to the type of the target object, obtaining the historical behavior data of multiple target objects and the historical identity data of multiple target objects; Based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects, using the evaluation metrics to evaluate multiple pre-obtained candidate recommendation schemes, and obtaining the target prediction value corresponding to each candidate recommendation scheme, where the target prediction value includes at least one of the predicted value of the conversion rate, the predicted value of the purchase rate, the predicted value of the number of visits, and the predicted value of the long-term user retention rate; Screening multiple candidate recommendation schemes according to the target prediction values corresponding to multiple candidate recommendation schemes to obtain a target recommendation scheme.

2. The method according to claim 1, wherein The screening multiple candidate recommendation schemes according to the target prediction values corresponding to multiple candidate recommendation schemes to obtain a target recommendation scheme includes: If the target prediction value includes the predicted value of the conversion rate, the predicted value of the purchase rate, the predicted value of the number of visits, or the predicted value of the long-term user retention rate, then using the candidate recommendation scheme corresponding to the largest target prediction value among the target prediction values corresponding to multiple candidate recommendation schemes as the target recommendation scheme; Or, If the target prediction value includes the predicted value of the conversion rate, the predicted value of the purchase rate, the predicted value of the number of visits, or the predicted value of the long-term user retention rate, then using the candidate recommendation scheme corresponding to the target prediction value greater than the preset value among the target prediction values corresponding to multiple candidate recommendation schemes as the target recommendation scheme.

3. The method according to claim 1, wherein The screening multiple candidate recommendation schemes according to the target prediction values corresponding to multiple candidate recommendation schemes to obtain a target recommendation scheme includes: If the target prediction value includes at least N dimensions of the predicted value of the conversion rate, the predicted value of the purchase rate, the predicted value of the number of visits, and the predicted value of the long-term user retention rate, then for each candidate recommendation scheme, performing a weighted sum calculation on the predicted values included in the target prediction value corresponding to the candidate recommendation scheme to obtain the weighted score corresponding to the candidate recommendation scheme, where N is an integer greater than or equal to 2 and less than or equal to z; Using the candidate recommendation scheme corresponding to the largest weighted score among the weighted scores corresponding to multiple candidate recommendation schemes as the target recommendation scheme; Or, If the target prediction value includes at least N dimensions of the predicted value of the conversion rate, the predicted value of the purchase rate, the predicted value of the number of visits, and the predicted value of the long-term user retention rate, then for each dimension, using the candidate recommendation scheme corresponding to the largest predicted value in the dimension as an intermediate recommendation scheme, and obtaining N intermediate recommendation schemes; In response to a second input, determining the target recommendation scheme among the N intermediate recommendation schemes, where the second input is used to select the target recommendation scheme.

4. The method according to claim 1, wherein Based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects, using the evaluation index to evaluate multiple pre-acquired candidate recommendation schemes, and obtaining a target prediction value corresponding to each candidate recommendation scheme, including: Through a pre-trained scheme evaluation model, based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects, using the evaluation index to evaluate multiple candidate recommendation schemes, and obtaining a target prediction value corresponding to each candidate recommendation scheme; Among them, the scheme evaluation model is trained by multiple training data, and the multiple training data include the historical behavior data of multiple reference objects, the historical identity data of multiple reference objects, multiple candidate recommendation schemes, and the actual values of the conversion rate, the actual value of the purchase rate, the actual value of the number of visits, and the actual value of the long-term user retention rate corresponding to each candidate recommendation scheme.

5. The method according to claim 4, characterized in that The step of using the pre-trained scheme evaluation model to evaluate multiple candidate recommendation schemes based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects, and obtaining a target prediction value corresponding to each candidate recommendation scheme, includes: Through the scheme evaluation model, based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects, using the evaluation index to evaluate multiple candidate recommendation schemes, and obtaining an initial prediction value corresponding to each candidate recommendation scheme and the confidence level corresponding to the initial prediction value; If the confidence level is less than the preset confidence level, adjust the type of the target object, and jump to the step of obtaining the historical behavior data of multiple target objects and the historical identity data of multiple target objects according to the type of the target object for execution until the confidence level is greater than or equal to the preset confidence level; If the confidence level is greater than or equal to the preset confidence level, use the initial prediction value corresponding to each candidate recommendation scheme as the target prediction value corresponding to each candidate recommendation scheme.

6. The method according to claim 1, characterized in that, The step of obtaining the historical behavior data of multiple target objects and the historical identity data of multiple target objects according to the type of the target object includes: Determine multiple initial objects according to the type of the target object; Obtain the recommended product white list of multiple initial objects; For each initial object, when the type of the product in the recommended product white list of the initial object includes the type of the product to be recommended, determine the initial object as the target object; Obtain the historical behavior data of multiple target objects and the historical identity data of multiple target objects.

7. The method according to claim 1, characterized in that The historical behavior data includes at least one of the historical purchase information of the target object, the historical visit information of the target object, and the historical search information of the target object, and the historical identity data includes at least one of the age of the target object, the gender of the target object, and the region where the target object is located.

8. A device for determining a product recommendation scheme, characterized in that, The device includes: A response module, configured to obtain the type of the target object recommended by the product to be recommended and the evaluation metrics in response to the first input, where the evaluation metrics are used to evaluate the effect of recommending the product to be recommended to the target object; An acquisition module, configured to obtain the historical behavior data of multiple target objects and the historical identity data of multiple target objects according to the type of the target object; An evaluation module, configured to evaluate multiple pre-obtained candidate recommendation schemes based on the historical behavior data of multiple target objects and the historical identity data of multiple target objects, using the evaluation metrics, to obtain a target prediction value corresponding to each candidate recommendation scheme, where the target prediction value includes at least one of a predicted conversion rate, a predicted purchase rate, a predicted number of visits, and a predicted long-term user retention rate; A screening module, configured to screen multiple candidate recommendation schemes according to the target prediction values corresponding to the multiple candidate recommendation schemes to obtain a target recommendation scheme.

9. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the product recommendation scheme determination method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, the product recommendation scheme determination method according to any one of claims 1 to 7 is implemented.

Images