Activity recommendation method and device, equipment, medium and program product

Abstract

The invention provides an activity recommendation method and device, equipment, a medium and a program product, and can be applied to the technical field of artificial intelligence or financial science and technology. The activity recommendation method comprises the steps that user portraits of multiple users are input into a pre-trained behavior prediction model, predicted behavior data representing the multiple users in a preset duration are obtained, and the user portraits comprise historical transaction data of the users and user tags; obtaining reference portrait data which is pre-constructed for the reference user and is related to the theme of the target activity; performing similarity calculation on the predicted behavior data and the reference portrait data to obtain activity consistency data used for representing the topic correlation degree of the predicted behavior of the user and the target activity; determining a target user based on the activity consistency data, the user portraits of the multiple users, the predicted behavior data and the feature vector of the target activity; and recommending the target activity to the target user.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence or the technical field of financial technology, and more particularly to an activity recommendation method, device, equipment, medium and program product. BACKGROUND

[0002] In current credit card activity recommendation, mainly rely on through the pre-set rules or based on coarse-grained label to carry out crowd targeting. But this kind of traditional recommendation method although can realize the preliminary screening of target customers to a certain extent, but there is obvious limitation, leading to the recommendation accuracy of high response user is not high. SUMMARY

[0003] In view of the above problems, the present application provides an activity recommendation method, device, equipment, medium and program product.

[0004] According to a first aspect of the present application, an activity recommendation method is provided, comprising: inputting user portraits of a plurality of users into a pre-trained behavior prediction model to obtain predicted behavior data representing the plurality of users within a preset time length, the user portrait including historical transaction data and user labels of the user; obtaining reference portrait data related to the theme of a target activity pre-constructed for a reference user; performing similarity calculation on the predicted behavior data and the reference portrait data to obtain activity consistency data representing the degree of relevance between the predicted behavior of the user and the theme of the target activity; determining a target user based on the activity consistency data, the user portraits of the plurality of users, the predicted behavior data and the feature vector of the target activity; recommending the target activity to the target user.

[0005] According to an embodiment of the present application, the determination of the target user based on the activity consistency data, the user portraits of the plurality of users, the predicted behavior data and the feature vector of the target activity comprises: inputting the user portraits of the plurality of users into a pre-trained classification model to obtain response prediction data representing the probability of each user in the plurality of users responding to the activity; inputting the user portraits of the plurality of users, the predicted behavior data and the feature vector of the target activity into a pre-trained response prediction model to obtain a user activity response data group; determining the target user based on the activity consistency data, the response prediction data and the user activity response data group.

[0006] According to an embodiment of the present application, the target user is determined based on the activity consistency data, the response prediction data and the user activity response data set, comprising: for each user in the plurality of users, determining a response score of the user to the target activity based on the activity consistency data, the response prediction data and the user activity response data corresponding to the user; and screening the plurality of users based on the response score to obtain the target user.

[0007] According to an embodiment of the present application, the reference portrait data is constructed by: obtaining target historical response user data representing user interaction behavior with an activity; and clustering the target historical response user data to obtain the reference portrait data.

[0008] According to an embodiment of the present application, the target historical response user data comprises a user response probability, and the clustering of the target historical response user data to obtain the reference portrait data comprises: clustering the target historical response user data to obtain a plurality of clustering clusters; determining a target clustering cluster based on the user response probability of each clustering cluster in the plurality of clustering clusters; and reducing the dimension of the target historical response user data corresponding to the target clustering cluster to obtain the reference portrait data of a preset dimension.

[0009] According to an embodiment of the present application, the target historical response user data is determined by: screening historical response user data based on the similarity of the target activity and a plurality of historical activities to obtain the target historical response user data.

[0010] According to an embodiment of the present application, the reference portrait data comprises M activity target user vectors, the prediction behavior data comprises M prediction probability behavior vectors, M is an integer greater than or equal to 1, and the similarity calculation of the prediction behavior data and the reference portrait data to obtain the activity consistency data representing the correlation degree of the predicted behavior of the user and the theme of the target activity comprises: calculating the dot product of the mth activity target user vector and the mth prediction probability behavior vector to obtain the mth user probability dot product, m is an integer greater than 1 and less than or equal to M; calculating the product of the modulus of the mth activity target user vector and the modulus of the mth prediction probability behavior vector to obtain the mth user probability product; obtaining the mth vector similarity based on the mth user probability dot product and the mth user probability product to obtain a vector similarity set, the vector similarity set comprising M vector similarities; and determining the activity consistency data based on the M vector similarities.

[0011] The second aspect of the present application provides an activity recommendation device, comprising: an input module configured to input user portraits of a plurality of users into a pre-trained behavior prediction model to obtain predicted behavior data representing predicted behaviors of the plurality of users within a preset time period, wherein the user portraits comprise historical transaction data and user labels of the users; an acquisition module configured to acquire reference portrait data related to a theme of a target activity for a reference user; a similarity calculation module configured to calculate a similarity between the predicted behavior data and the reference portrait data to obtain activity consistency data representing a degree of relevance between predicted behaviors of the users and the theme of the target activity; a target user determination module configured to determine a target user based on the activity consistency data, the user portraits of the plurality of users, the predicted behavior data, and a feature vector of the target activity; and a recommendation module configured to recommend the target activity to the target user.

[0012] According to an embodiment of the present application, the target user determination module comprises: a classification input submodule configured to input the user portraits of the plurality of users into a pre-trained classification model to obtain response prediction data representing response probabilities of each of the plurality of users to an activity; a prediction input submodule configured to input the user portraits of the plurality of users, the predicted behavior data, and the feature vector of the target activity into a pre-trained response prediction model to obtain a user activity response data set; and a determination submodule configured to determine the target user based on the activity consistency data, the response prediction data, and the user activity response data set.

[0013] According to an embodiment of the present application, the determination submodule comprises: a response score determination unit configured to determine, for each of the plurality of users, a response score of the user to the target activity based on the activity consistency data, the response prediction data, and the user activity response data corresponding to the user; and a screening unit configured to screen the plurality of users based on the response scores to obtain the target user.

[0014] According to an embodiment of the present application, the device further comprises: a historical user data acquisition module configured to acquire target historical response user data representing interaction behaviors of users and activities; and a clustering module configured to cluster the target historical response user data to obtain the reference portrait data.

[0015] According to an embodiment of the present application, the target historical response user data comprises a user response probability, the clustering module comprises: a clustering submodule for clustering the target historical response user data to obtain a plurality of clustering clusters; a target clustering cluster determination submodule for determining a target clustering cluster based on the user response probability of each clustering cluster in the plurality of clustering clusters; and a dimension reduction submodule for dimension reducing the target historical response user data corresponding to the target clustering cluster to obtain the reference portrait data of a preset dimension.

[0016] According to an embodiment of the present application, the device further comprises: a historical activity data acquisition module, a historical user data filtering module for filtering historical response user data based on the similarity of the target activity and a plurality of historical activities to obtain the target historical response user data.

[0017] According to an embodiment of the present application, the reference portrait data comprises M activity target user vectors, the predicted behavior data comprises M predicted probability behavior vectors, M is an integer greater than or equal to 1, the similarity calculation module comprises: a dot product calculation submodule for calculating the dot product of the mth activity target user vector and the mth predicted probability behavior vector to obtain the mth user probability dot product, m is an integer greater than 1 and less than or equal to M; a product calculation submodule for calculating the product of the modulus of the mth activity target user vector and the modulus of the mth predicted probability behavior vector to obtain the mth user probability product; a similarity calculation submodule for obtaining the mth vector similarity based on the mth user probability dot product and the mth user probability product to obtain a vector similarity group, the vector similarity group comprises M vector similarities; and a determination submodule for determining the activity consistency data based on the M vector similarities.

[0018] The third aspect of the present application provides an electronic device comprising: one or more processors; a memory for storing one or more computer programs, wherein the one or more processors execute the one or more computer programs to implement the steps of the method.

[0019] The fourth aspect of the present application further provides a computer-readable storage medium having a computer program or instructions stored thereon, wherein the computer program or instructions are executed by a processor to implement the steps of the method.

[0020] The fifth aspect of the present application further provides a computer program product comprising a computer program or instructions, wherein the computer program or instructions are executed by a processor to implement the steps of the method.

[0021] According to the embodiment of the present application, by inputting the user portrait of the user into the pre-trained behavior prediction model, the behavior of the user within a preset time length can be accurately predicted, and meanwhile, in combination with the reference portrait data related to the target activity theme, the correlation degree between the predicted behavior of the user and the target activity theme is calculated, so as to filter out the user most interested in the activity. Compared with the traditional rule filtering or coarse-grained label orientation, the present application not only considers the static features of the user, but also fuses the future behavior trend prediction of the user, so that the high-response user group can be more accurately identified, the invalid recommendation is reduced, and the participation rate and conversion rate of the activity are improved. BRIEF DESCRIPTION OF DRAWINGS

[0022] The above and other objects, features and advantages of the present application will become more apparent from the following description when taken in conjunction with the accompanying drawings, in which:

[0023] Figure 1 An application scenario diagram of an activity recommendation method and device according to an embodiment of the present application is schematically shown;

[0024] Figure 2 A flowchart of an activity recommendation method according to an embodiment of the present application is schematically shown;

[0025] Figure 3 A flowchart of determining a target user according to an embodiment of the present application is schematically shown;

[0026] Figure 4 A structural block diagram of an activity recommendation device according to an embodiment of the present application is schematically shown; and

[0027] Figure 5 A block diagram of an electronic device suitable for implementing an activity recommendation method according to an embodiment of the present application is schematically shown. DETAILED DESCRIPTION

[0028] Hereinafter, embodiments of the present application will be described with reference to the accompanying drawings. It is to be understood, however, that the description is merely exemplary and is intended to provide a thorough understanding of the present application. The following description, given together with the accompanying drawings, is intended to provide a thorough understanding of the present application. However, it is apparent that one or more embodiments can be implemented without the specific details, as is obvious to those skilled in the art.

[0029] The terms used herein are merely used to describe specific embodiments, and are not intended to limit the present application. The terms "include", "comprise" and the like used herein indicate the presence of the described features, steps, operations and / or components, but do not exclude the presence or addition of one or more other features, steps, operations or components.

[0030] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0031] When using expressions such as "at least one of A, B and C", they should generally be interpreted in accordance with the meaning that is commonly understood by those skilled in the art (e.g., "a system having at least one of A, B and C" should include, but is not limited to, a system having A alone, a system having B alone, a system having C alone, a system having A and B, a system having A and C, a system having B and C, and / or a system having A, B and C, etc.).

[0032] It should be noted that the activity recommendation methods, apparatus, equipment, media, and program products of this application can be used in the fields of fintech or artificial intelligence, or in any field other than fintech. This application does not limit the application fields of the activity recommendation methods, apparatus, equipment, media, and program products.

[0033] In the technical solution of this application, the user information (including but not limited to user personal information, user image information, user device information, such as location information) and data (including but not limited to data used for analysis, stored data, and displayed data) involved are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of related data all comply with relevant laws, regulations, and standards, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entry points for users to choose to authorize or refuse.

[0034] In scenarios involving automated decision-making using personal information, the methods, devices, and systems provided in this application all offer users corresponding entry points for choosing to agree to or reject the automated decision-making results. If the user chooses to reject, the process proceeds to the expert decision-making stage. Here, "automated decision-making" refers to the activity of automatically analyzing and evaluating an individual's behavioral habits, interests, or economic, health, and credit status through computer programs, and then making a decision. Here, "expert decision-making" refers to the activity of making decisions by personnel who specialize in a particular field, possess specialized experience, knowledge, and skills, and have reached a certain level of professional expertise.

[0035] In today's credit card activity recommendation, the mainstream approach is to rely on pre-set rules or to conduct crowd targeting based on coarse-grained labels. The pre-set rule screening is to screen the target customers by setting some fixed conditions. The coarse-grained label targeting is to divide the crowd based on some basic attributes or approximate behavior characteristics of the user. Although this traditional recommendation method can achieve preliminary screening of target customers to some extent, it has obvious limitations, resulting in low recommendation accuracy for high-response users.

[0036] Therefore, the embodiments of the present application provide an activity recommendation method, comprising: inputting user portraits of a plurality of users into a pre-trained behavior prediction model to obtain predicted behavior data representing the behaviors of the plurality of users within a preset time period, wherein the user portrait comprises historical transaction data and user labels of the user; obtaining reference portrait data related to the theme of a target activity for a reference user; performing similarity calculation on the predicted behavior data and the reference portrait data to obtain activity consistency data representing the degree of relevance between the predicted behavior of the user and the theme of the target activity; determining a target user based on the activity consistency data, the user portraits of the plurality of users, the predicted behavior data, and a feature vector of the target activity; and recommending the target activity to the target user.

[0037] Figure 1 The application scenario of the activity recommendation method and device according to the embodiments of the present application is schematically shown.

[0038] As shown in Figure 1 The application scenario 100 according to the embodiments can include a terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105. The network 104 is used as a medium to provide a communication link between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. The network 104 can include various connection types, such as wired, wireless communication links, or optical fiber cables, etc.

[0039] The user can use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 through the network 104 to receive or send messages, etc. Various communication client applications can be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only as examples).

[0040] The first terminal device 101, the second terminal device 102, and the third terminal device 103 can be various electronic devices with display screens and supporting web browsing, including but not limited to smartphones, tablet computers, laptop computers, and desktop computers, etc.

[0041] Server 105 can be a server that provides various services, such as a backend management server that supports websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103 (this is just an example). The backend management server can analyze and process data such as received user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal devices.

[0042] It should be noted that the activity recommendation method provided in this application embodiment can generally be executed by server 105. Correspondingly, the activity recommendation device provided in this application embodiment can generally be located in server 105. The activity recommendation method provided in this application embodiment can also be executed by a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or server 105. Correspondingly, the activity recommendation device provided in this application embodiment can also be located in a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or server 105.

[0043] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0044] The following will be based on Figure 1 The described scene, through Figures 2-3 The activity recommendation method according to the embodiments of this application will be described in detail.

[0045] Figure 2 A flowchart illustrating an activity recommendation method according to an embodiment of this application is shown schematically.

[0046] like Figure 2 As shown, the activity recommendation method in this embodiment includes operations S210 to S250.

[0047] In operation S210, user profiles of multiple users are input into a pre-trained behavior prediction model to obtain predicted behavior data representing multiple users within a preset time period.

[0048] The user profile includes the user's historical transaction data and user tags.

[0049] According to an embodiment of the present application, the user portrait can be a high-dimensional vector, and the user portrait can be constructed based on user behavior characteristics, user time series behavior embedding, user life cycle, etc. For example, the user portrait can be constructed based on the user's recent average consumption amount, consumption category distribution, login frequency, etc. After construction, the feature vector can be further selected and dimensionally reduced to generate a user portrait with a preset dimension. The preset dimension can be 100, for example.

[0050] For ease of understanding, the time series behavior embedding is explained as follows:

[0051] Time series behavior embedding is a technique that encodes behavior patterns and time-dependent relationships in time series data into a low-dimensional vector space. It maps each data point in the time series and its historical information to an embedding vector, thereby capturing dynamic behavior and patterns in the time series.

[0052] In operation S220, reference portrait data related to the theme of the target activity is obtained for the reference user.

[0053] According to an embodiment of the present application, the reference portrait data can be obtained by clustering high-response groups in historical response user data, or can be generated or filtered by key features specified by relevant staff. The present application does not limit this.

[0054] In operation S230, similarity calculation is performed on the predicted behavior data and the reference portrait data to obtain activity consistency data representing the degree of relevance between the predicted behavior of the user and the theme of the target activity.

[0055] According to an embodiment of the present application, the consistency between the predicted behavior data and the reference portrait data can be determined by calculating the cosine similarity, Euclidean distance, Manhattan distance, etc. between the two, and the present application does not limit this.

[0056] In operation S240, the target user is determined based on the activity consistency data, the user portraits of the plurality of users, the predicted behavior data, and the feature vector of the target activity.

[0057] In operation S250, the target user is determined based on the activity consistency data, the user portraits of the plurality of users, the predicted behavior data, and the feature vector of the target activity.

[0058] For example, for user A, the determined predicted behavior data includes "going overseas, going to a restaurant, applying for installment", and for user B, the determined predicted behavior data includes "going to a restaurant, credit card increase", etc. It can be determined that user A is the target user of the target activity "overseas credit card preferential treatment".

[0059] According to the embodiment of the present application, by inputting the user portrait of the user into the pre-trained behavior prediction model, the behavior of the user within a preset time length can be accurately predicted, and meanwhile, in combination with the reference portrait data related to the target activity theme, the correlation degree between the predicted behavior of the user and the target activity theme is calculated, so as to filter out the user who is most interested in the activity. Compared with the traditional rule filtering or coarse-grained label orientation, the present application not only considers the static features of the user, but also combines the future behavior trend prediction of the user, so that the high-response user group can be more accurately identified, the invalid recommendation is reduced, and the participation rate and conversion rate of the activity are improved.

[0060] According to the embodiment of the present application, the above determining the target user based on the activity consistency data, the user portraits of the plurality of users, the predicted behavior data and the feature vector of the target activity comprises: inputting the user portraits of the plurality of users into a pre-trained classification model to obtain response prediction data representing the probability of each user in the plurality of users responding to the activity; inputting the user portraits of the plurality of users, the predicted behavior data and the feature vector of the target activity into a pre-trained response prediction model to obtain a user activity response data group; and determining the target user based on the activity consistency data, the response prediction data and the user activity response data group.

[0061] According to the embodiment of the present application, the above classification model and response prediction model can be a logistic regression model, an extreme gradient boosting model or a bidirectional encoder representation model, etc. The probability that the user is a high-response user can be determined by inputting the user portrait into the pre-trained classification model, and then the response probability of the plurality of users to the target activity can be determined by the user portrait, the predicted behavior data and the feature vector of the target activity.

[0062] According to the embodiment of the present application, the response probability of the user is determined by the classification model, and further, the user portraits of the plurality of users, the predicted behavior data and the feature vector of the target activity are comprehensively analyzed, and then the target user who is highly responsive to the target activity is filtered out, and the target activity is recommended to the target user, so as to improve the activity recommendation effect, and at the same time, since the target activity meeting the demand of the target user is recommended to the target user, the user experience and satisfaction are improved.

[0063] According to the embodiment of the present application, the above determining the target user based on the activity consistency data, the response prediction data and the user activity response data group comprises: for each user in the plurality of users, determining a response score of the user to the target activity based on the activity consistency data, the response prediction data and the user activity response data corresponding to the user; and filtering the plurality of users based on the response score to obtain the target user.

[0064] According to the embodiment of the present application, the response score of the user to the target activity can be determined according to the activity consistency data, the response prediction data and the user activity response data corresponding to the user through the pre-trained scoring model, or can be determined according to the preset rule, and the present application does not make any limitation in this regard.

[0065] The response score of the user to the target activity can be calculated by the following formula (1).

[0066] Score(u, a) = a · Sim(P'_u, T_a) + b · Prob(u e A_target) + g · R_u(a) (1)

[0067] Wherein, Score(u, a) represents the response score of the user u to the target activity a, a, b, g are weight parameters, P'_u represents the predicted behavior data of the user u, T_a represents the reference portrait data of the target activity a, Sim(P'_u, T_a) represents the activity consistency data of the predicted behavior of the user u and the target activity a, Prob(u e A_target) represents the response prediction data of the user u to the activity, and R_u(a) represents the user activity response data of the user u to the activity a.

[0068] According to the embodiment of the present application, the weight parameters a, b, g can be optimized by Bayesian optimization, grid search, A / B test result feedback, etc. In the case of large parameter space and limited computing resources, Bayesian optimization can be selected to optimize the weight parameters. In the case of low parameter dimension and sufficient computing resources, grid search can be selected to optimize the weight parameters. In the case of real-time feedback, A / B test result feedback can be used to optimize the weight parameters.

[0069] According to the embodiment of the present application, when the plurality of users are screened, the plurality of users can be sorted from high to low according to the response score, and the first preset number of users can be selected as the target user.

[0070] According to the embodiment of the present application, by determining the response score data of the plurality of users to the target activity, and then determining the target user, the plurality of data dimensions and the optimized weight parameters are comprehensively considered, which can more accurately identify the target user with high response probability to the target activity. At the same time, according to different application scenarios and resource conditions, the appropriate weight parameter optimization method can be selected, and then the accuracy and timeliness of the response score data can be ensured, and the activity recommendation effect can be improved.

[0071] According to the embodiment of the present application, the above-mentioned reference portrait data is constructed by the following method: obtaining target historical response user data representing the interaction behavior of the user and the activity; clustering the target historical response user data to obtain the reference portrait data.

[0072] According to the embodiments of the present application, the target historical response user data can include recent transaction records of a plurality of users, bill information, user basic attributes, channel preferences, user response data, and the like. Further, the target historical response user data can be subjected to preprocessing operations such as standardization, missing data completion, normalization, and the like to improve the target historical response user data.

[0073] According to the embodiments of the present application, the target historical response user data includes user response probabilities, the target historical response user data is clustered to obtain reference portrait data, including: clustering the target historical response user data to obtain a plurality of clustering clusters; determining a target clustering cluster based on the user response probabilities of each clustering cluster in the plurality of clustering clusters; and reducing the dimension of the target historical response user data corresponding to the target clustering cluster to obtain reference portrait data of a preset dimension.

[0074] According to the embodiments of the present application, the target historical response user data can be clustered by using a K-means algorithm, a hierarchical clustering algorithm, or a spectral clustering algorithm, and the present application does not limit the clustering algorithm. Further, the most relevant clustering cluster to the target behavior (for example, the response probability) can be selected from the plurality of clustering clusters for further analysis, so as to determine the target clustering cluster.

[0075] According to the embodiments of the present application, the target historical response user data reflecting the behavior characteristics and preferences of the user is clustered, the user is divided into a plurality of clustering clusters with similar behavior characteristics, and then the most relevant clustering cluster to the target behavior is selected from the plurality of clustering clusters, so as to improve the pertinence and effectiveness of the analysis. Further, the high-dimensional target historical response user data is reduced in dimension, which facilitates subsequent analysis and provides support, and reduces the computational cost.

[0076] According to the embodiments of the present application, the target historical response user data is determined by the following method: the historical response user data is filtered based on the similarity groups of the target activity and a plurality of historical activities to obtain the target historical response user data.

[0077] According to the embodiments of the present application, the historical activity data can be filtered according to the similarity groups of the historical activities and the target activity before the target historical response user data is determined, so as to select the target historical response user data more similar to the target activity, and then the reference portrait data determined is more suitable for the target activity.

[0078] According to the embodiment of the present application, the reference image data includes M active target user vectors, the predicted behavior data includes M predicted probability behavior vectors, M is an integer greater than or equal to 1, the similarity calculation is performed on the predicted behavior data and the reference image data, and the activity consistency data for representing the correlation degree of the predicted behavior of the user and the theme of the target activity is obtained, including: calculating the dot product of the mth active target user vector and the mth predicted probability behavior vector to obtain the mth user probability dot product, m is an integer greater than 1 and less than or equal to M; calculating the product of the modulus of the mth active target user vector and the modulus of the mth predicted probability behavior vector to obtain the mth user probability product; obtaining the mth vector similarity based on the mth user probability dot product and the mth user probability product to obtain a vector similarity group, the vector similarity group includes M vector similarities; and determining the activity consistency data based on the M vector similarities.

[0079] According to the embodiment of the present application, the similarity calculation can be performed by the following formula (2).

[0080] (2)

[0081] Wherein, S m represents the mth vector similarity, A m represents the mth active target user vector, B m represents the mth predicted probability behavior vector.

[0082] Figure 3 The flowchart for determining the target user according to the embodiment of the present application is schematically shown.

[0083] As Figure 3As shown, the process begins with building a user portrait using user data, which involves integrating multi-dimensional information such as basic information, behavioral habits, and preferences of the user to form a comprehensive description of the user's characteristics. Next, reference portrait data is built based on historical response data of the target, extracting user characteristics that have a significant response to the target activity through analysis of historical data. Then, the built user portrait is input into the behavior prediction model to predict the user's possible behavior patterns, generating prediction behavior data; at the same time, the user portrait is input into the classification model to predict the user's possible response probability to a specific activity, generating response prediction data. Next, similarity calculation is performed on the reference portrait data and the prediction behavior data to evaluate the relevance of the user's predicted behavior to the theme of the target activity, obtaining activity consistency data. In addition, combining the prediction behavior data, the user portrait, and the feature vector of the target activity, the specific response of the user to the target activity is analyzed, obtaining user activity response data. Finally, based on the activity consistency data, the response prediction data, and the user activity response data, a response score is calculated for each user to reflect the potential response likelihood of the user to the target activity, and the target user group is determined according to these scores. This systematic process helps to improve the accuracy and efficiency of marketing activities, optimize resource allocation, and improve user experience.

[0084] Based on the above activity recommendation method, the application also provides an activity recommendation device. The following will be combined with Figure 4 to describe the device in detail.

[0085] Figure 4 The structure block diagram of the activity recommendation device according to the embodiment of the application is schematically shown.

[0086] As Figure 4 shown, the activity recommendation device 400 of this embodiment includes an input module 410, an acquisition module 420, a similarity calculation module 430, a target user determination module 440, and a recommendation module 450.

[0087] The input module 410 is configured to input the user portraits of a plurality of users into a pre-trained behavior prediction model to obtain prediction behavior data representing the predicted behaviors of the plurality of users within a preset time length, the user portraits including historical transaction data and user labels of the users. In an embodiment, the input module 410 can be configured to perform the operation S210 described above, and details are not repeated here.

[0088] The acquisition module 420 is configured to acquire reference portrait data related to the theme of the target activity that is pre-built for a reference user. In an embodiment, the acquisition module 420 can be configured to perform the operation S220 described above, and details are not repeated here.

[0089] The similarity calculation module 430 is configured to perform similarity calculation on the predicted behavior data and the reference portrait data to obtain activity consistency data representing the relevance of the predicted behavior of the user to the theme of the target activity. In an embodiment, the similarity calculation module 430 can be configured to perform operation S230 described above, and details are not repeated here.

[0090] The target user determination module 440 is configured to determine the target user based on the activity consistency data, the user portraits of the plurality of users, the predicted behavior data, and the feature vector of the target activity. In an embodiment, the target user determination module 440 can be configured to perform operation S240 described above, and details are not repeated here.

[0091] The recommendation module 450 is configured to recommend the target activity to the target user. In an embodiment, the recommendation module 450 can be configured to perform operation S250 described above, and details are not repeated here.

[0092] According to an embodiment of the present application, the target user determination module comprises: a classification input sub-module configured to input the user portraits of the plurality of users into a pre-trained classification model to obtain response prediction data representing the probability of each user in the plurality of users responding to the activity; a prediction input sub-module configured to input the user portraits of the plurality of users, the predicted behavior data, and the feature vector of the target activity into a pre-trained response prediction model to obtain a user activity response data set; and a determination sub-module configured to determine the target user based on the activity consistency data, the response prediction data, and the user activity response data set.

[0093] According to an embodiment of the present application, the determination sub-module comprises: a response score determination unit configured to, for each user in the plurality of users, determine a response score of the user to the target activity based on the activity consistency data, the response prediction data, and the user activity response data corresponding to the user; and a screening unit configured to screen the plurality of users based on the response scores to obtain the target user.

[0094] According to an embodiment of the present application, the apparatus further comprises: a historical user data acquisition module configured to acquire target historical response user data representing the interaction behavior of a user with an activity; and a clustering module configured to cluster the target historical response user data to obtain the reference portrait data.

[0095] According to an embodiment of the present application, the target historical response user data comprises a user response probability, the clustering module comprises: a clustering submodule for clustering the target historical response user data to obtain a plurality of clustering clusters; a target clustering cluster determination submodule for determining a target clustering cluster based on the user response probability of each clustering cluster in the plurality of clustering clusters; and a dimension reduction submodule for dimension reduction on the target historical response user data corresponding to the target clustering cluster to obtain the reference portrait data of a preset dimension.

[0096] According to an embodiment of the present application, the device further comprises: a historical activity data acquisition module, a historical user data filtering module for filtering historical response user data based on the similarity of the target activity and a plurality of historical activities to obtain the target historical response user data.

[0097] According to an embodiment of the present application, the reference portrait data comprises M activity target user vectors, the predicted behavior data comprises M predicted probability behavior vectors, M is an integer greater than or equal to 1, the similarity calculation module comprises: a dot product calculation submodule for calculating the dot product of the mth activity target user vector and the mth predicted probability behavior vector to obtain the mth user probability dot product, m is an integer greater than 1 and less than or equal to M; a product calculation submodule for calculating the product of the modulus of the mth activity target user vector and the modulus of the mth predicted probability behavior vector to obtain the mth user probability product; a similarity calculation submodule for obtaining the mth vector similarity based on the mth user probability dot product and the mth user probability product to obtain a vector similarity group, the vector similarity group comprises M vector similarities; and a determination submodule for determining the activity consistency data based on the M vector similarities.

[0098] According to an embodiment of the present application, any of the plurality of modules of the input module 410, the obtaining module 420, the similarity calculation module 430, the target user determination module 440 and the recommendation module 450 can be combined in one module, or any of the plurality of modules can be split into a plurality of modules. Alternatively, at least part of the function of one or more of the modules can be combined with at least part of the function of the other modules, and implemented in one module. According to an embodiment of the present application, at least one of the input module 410, the obtaining module 420, the similarity calculation module 430, the target user determination module 440 and the recommendation module 450 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on chip, a system on board, a system on package, an application specific integrated circuit (ASIC), or any other reasonable manner of integrating or packaging a circuit, etc. or implemented by hardware or firmware, or implemented in any one of software, hardware and firmware or in a proper combination of any of them. Alternatively, at least one of the input module 410, the obtaining module 420, the similarity calculation module 430, the target user determination module 440 and the recommendation module 450 can be at least partially implemented as a computer program module which, when executed, can perform the corresponding function.

[0099] Figure 5 A block diagram of an electronic device suitable for implementing the activity recommendation method according to an embodiment of the present application is schematically shown.

[0100] As shown in Figure 5 The electronic device 500 according to an embodiment of the present application includes a processor 501 which can perform various appropriate actions and processes according to a program stored in a read only memory (ROM) 502 or a program loaded from a storage portion 508 into a random access memory (RAM) 503. The processor 501 can include, for example, a general purpose microprocessor (e.g. a CPU), an instruction set processor and / or a related chipset and / or a special purpose microprocessor (e.g. an application specific integrated circuit (ASIC)), etc. The processor 501 can also include an on-board memory for cache use. The processor 501 can include a single processing unit or a plurality of processing units for performing different actions of the method processes according to an embodiment of the present application.

[0101] In the RAM 503, various programs and data required for the operation of the electronic device 500 are stored. The processor 501, the ROM 502, and the RAM 503 are connected to each other via the bus 504. The processor 501 performs various operations of the method flow according to the embodiments of the present application by executing the programs in the ROM 502 and / or the RAM 503. It should be noted that the programs can also be stored in one or more memories other than the ROM 502 and the RAM 503. The processor 501 can also perform various operations of the method flow according to the embodiments of the present application by executing the programs stored in the one or more memories.

[0102] According to the embodiments of the present application, the electronic device 500 can further include an input / output (I / O) interface 505, which is also connected to the bus 504. The electronic device 500 can further include one or more of the following components connected to the input / output (I / O) interface 505: an input part 506 including a keyboard, a mouse, and the like; an output part 507 including a cathode ray tube (CRT), a liquid crystal display (LCD), and the like, and a speaker, and the like; a storage part 508 including a hard disk, and the like; and a communication part 509 including a network interface card such as a LAN card, a modem, and the like. The communication part 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the input / output (I / O) interface 505 as necessary. A removable medium 511 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, and the like is mounted on the drive 510 as necessary, so that a computer program read therefrom is installed in the storage part 508 as necessary.

[0103] The present application also provides a computer readable storage medium, which can be included in the device / apparatus / system described in the above embodiments; or can exist separately without being assembled into the device / apparatus / system. The above computer readable storage medium carries one or more programs, when the one or more programs are executed, the method according to the embodiments of the present application is implemented.

[0104] According to an embodiment of the present application, the computer readable storage medium can be a non-transitory computer readable storage medium, for example, can include but not limited to: a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In this application, a computer readable storage medium can be any tangible medium that contains or stores a program for use by or in connection with an instruction execution system, apparatus, or device. For example, according to an embodiment of the present application, the computer readable storage medium can include one or more of the above-described ROM 502 and / or RAM 503 and / or a memory other than the ROM 502 and the RAM 503.

[0105] Embodiments of the present application also include a computer program product, which includes a computer program containing program codes for executing the methods shown in the flowcharts. When the computer program product is run in a computer system, the program codes are used to make the computer system implement the activity recommendation method provided by the embodiments of the present application.

[0106] The above-described functions defined in the system / device / apparatus of the embodiments of the present application are performed when the computer program is executed by the processor 501. According to an embodiment of the present application, the above-described system, device, module, unit, etc. can be implemented by computer program modules.

[0107] In one embodiment, the computer program can rely on a tangible storage medium such as an optical storage device, a magnetic storage device, etc. In another embodiment, the computer program can also be transmitted, distributed, and downloaded in the form of a signal on a network medium, and be downloaded and installed through the communication part 509, and / or be installed from the detachable medium 511. The program codes contained in the computer program can be transmitted by any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the foregoing.

[0108] In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 509, and / or be installed from the detachable medium 511. When the computer program is executed by the processor 501, the above-described functions defined in the system of the embodiments of the present application are performed. According to an embodiment of the present application, the above-described system, device, apparatus, module, unit, etc. can be implemented by computer program modules.

[0109] According to embodiments of the present application, program code for implementing the computer programs provided by embodiments of the present application can be written in any combination of one or more programming languages, and can be implemented in a high-level procedural and / or object-oriented programming language, and / or in assembly / machine language. Programming languages include, but are not limited to, Java, C++, python, "C", or the like. Program code can execute entirely on a user's computing device, partly on the user's device, as a stand-alone software package, partly on a remote computing device, or entirely on the remote computing device or server. In the latter scenario, the remote computing device can be connected to the user's computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computing device, such as through the Internet using an Internet Service Provider.

[0110] The computer program instructions can also be loaded onto a computer or other programmable information processing apparatus to cause a series of operations to be performed on the computer or other programmable information processing apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable information processing apparatus implement the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0111] Those skilled in the art will appreciate that features recited in the various embodiments of the present application can be combined and / or integrated in various combinations, even if such combinations have not been explicitly recited in the present application. In particular, the features recited in the various embodiments of the present application can be combined and / or integrated in various combinations, without departing from the spirit and teachings of the present application. All such combinations are within the scope of the present application.

Claims

1. An activity recommendation method, characterized in that, The method includes: The user profiles of multiple users are input into a pre-trained behavior prediction model to obtain predicted behavior data representing the multiple users within a preset time period. The user profiles include the user's historical transaction data and user tags. Obtain reference profile data pre-built for reference users that is related to the theme of the target activity; Similarity calculations are performed on the predicted behavior data and the reference profile data to obtain activity consistency data that characterizes the degree of relevance between the user's predicted behavior and the theme of the target activity; The target user is determined based on the activity consistency data, the user profiles of the multiple users, the predicted behavior data, and the feature vector of the target activity; Recommend the target activity to the target user.

2. The method according to claim 1, characterized in that, The step of determining the target user based on the activity consistency data, the user profiles of the multiple users, the predicted behavior data, and the feature vector of the target activity includes: The user profiles of the multiple users are input into a pre-trained classification model to obtain response prediction data that characterizes the probability of each user's response activity among the multiple users; The user profiles of the multiple users, the predicted behavior data, and the feature vector of the target activity are input into a pre-trained response prediction model to obtain a user activity response data set. The target user is determined based on the activity consistency data, the response prediction data, and the user activity response data set.

3. The method according to claim 2, characterized in that, The process of determining the target user based on the activity consistency data, the response prediction data, and the user activity response data group includes: For each of the plurality of users, a response score for the target activity is determined based on the activity consistency data, the response prediction data, and the user activity response data corresponding to the user. The target user is obtained by filtering the multiple users based on the response scores.

4. The method according to claim 1, characterized in that, The reference profile data is constructed using the following method: Acquire target historical response user data that characterizes user interaction behavior with activities; Clustering is performed on the target historical response user data to obtain the reference profile data.

5. The method according to claim 4, characterized in that, The target historical response user data includes user response probabilities, and the clustering of the target historical response user data to obtain the reference profile data includes: Clustering the target historical response user data yields multiple clusters; The target cluster is determined based on the user response probability of each cluster in the plurality of clusters; The target historical response user data corresponding to the target cluster is dimensionality reduced to obtain the reference profile data of the preset dimension.

6. The method according to claim 4, characterized in that, The target historical response user data is determined through the following method: Based on the similarity groups between the target activity and multiple historical activities, the historical response user data is filtered to obtain the target historical response user data.

7. The method according to claim 1, characterized in that, The reference profile data includes M target user vectors, and the predicted behavior data includes M predicted probability behavior vectors, where M is an integer greater than or equal to 1. The similarity calculation of the predicted behavior data and the reference profile data to obtain activity consistency data characterizing the degree of the relevance between the user's predicted behavior and the target activity's theme includes: Calculate the dot product of the m-th active target user vector and the m-th predicted probability behavior vector to obtain the m-th user probability dot product, where m is an integer greater than 1 and less than or equal to M; Calculate the product of the magnitude of the m-th active target user vector and the magnitude of the m-th predicted probability behavior vector to obtain the m-th user probability product; The m-th vector similarity is obtained based on the dot product of the m-th user probabilities and the product of the m-th user probabilities, to obtain a vector similarity group, which includes M vector similarities; The activity consistency data is determined based on the similarity of the M vectors.

8. An activity recommendation device, characterized in that, The device includes: The input module is used to input user profiles of multiple users into a pre-trained behavior prediction model to obtain predicted behavior data representing the multiple users within a preset time period. The user profiles include the user's historical transaction data and user tags. The acquisition module is used to acquire reference profile data pre-built for reference users that is related to the theme of the target activity; The similarity calculation module is used to perform similarity calculation on the predicted behavior data and the reference profile data to obtain activity consistency data that characterizes the degree of relevance between the user's predicted behavior and the topic of the target activity; The target user determination module is used to determine the target user based on the activity consistency data, the user profiles of the multiple users, the predicted behavior data, and the feature vector of the target activity. The recommendation module is used to recommend the target activity to the target user.

9. An electronic device, comprising: One or more processors; Memory, used to store one or more computer programs. The characteristic feature is that the one or more processors execute the one or more computer programs to implement the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program or instructions stored thereon, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the method according to any one of claims 1 to 7.

11. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the method according to any one of claims 1 to 7.

Images