A push method and device of a target object

Abstract

The specification provides a push method and device of a target object. The method comprises: obtaining behavior data of a target user, behavior data of each reference user in a reference user set, supply and demand data of each object to be pushed in an object to be pushed set, and multi-dimensional environmental data; determining a target prediction score of the target user for each object to be pushed in the object to be pushed set according to the behavior data of the target user, the behavior data of each reference user in the reference user set, the supply and demand data of each object to be pushed in the object to be pushed set, and the multi-dimensional environmental data; determining at least one object to be pushed to the target user according to the target prediction score of the target user for each object to be pushed in the object to be pushed set; and pushing the at least one object to the target user. Through the above scheme, the score of each object can be predicted, and the target object to be pushed to the target user can be accurately determined, thereby improving user experience and operation efficiency.

Description

Technical Field

[0001] This manual belongs to the field of big data technology, and in particular relates to a method and device for pushing data to a target object. Background Technology

[0002] A product recommendation system is used to recommend products to users. By pushing products that are suitable for users, the system can reduce the time users spend selecting products and increase the turnover rate of products.

[0003] However, existing product recommendation systems often encounter unstructured data problems when processing data. The existence of unstructured data leads to various different data structures, making it impossible to process the data uniformly and efficiently, which in turn results in lower accuracy of product recommendations.

[0004] Currently, there is no effective solution for improving the accuracy of product recommendation systems. Summary of the Invention

[0005] This application provides a method and apparatus for pushing to a target object, which determines the score of the object to be pushed by multi-dimensional data, thereby effectively improving the accuracy of recommendations.

[0006] This application provides a method and apparatus for pushing a target object, which is implemented as follows:

[0007] A method for pushing data to a target object, comprising:

[0008] Acquire behavioral data of target users, behavioral data of each reference user in the reference user set, supply and demand data of each target object in the target object set, and multi-dimensional environmental data;

[0009] Based on the target user's behavioral data, the behavioral data of each reference user in the reference user set, the supply and demand data of each target object in the target object set, and multi-dimensional environmental data, the target user's target prediction score for each target object in the target object set is determined.

[0010] Based on the target user's target prediction score for each target object in the target object set, determine at least one object to be pushed to the target user;

[0011] Push the at least one object to the target user.

[0012] In one implementation, based on the target user's behavioral data, the behavioral data of each reference user in the reference user set, the supply and demand data of each target object in the target object set, and multi-dimensional environmental data, the target user's target prediction score for each target object in the target object set is determined, including:

[0013] Based on the target user's behavioral data and the behavioral data of each reference user in the reference user set, determine the first predicted score of each target user in the target user set to be pushed to.

[0014] Based on the target user's behavioral data, the supply and demand data of each target object in the target object set, and multi-dimensional environmental data, the second predicted score of the target user for each target object in the target object set is determined.

[0015] Based on the first predicted score of each target object in the target object set and the second predicted score of each target object in the target object set, the target predicted score of each target object in the target object set is determined.

[0016] In one implementation, based on the target user's behavioral data and the behavioral data of each reference user in the reference user set, a first predicted score for each target user in the target user set is determined, including:

[0017] Based on the behavioral data of the target user and the behavioral data of each reference user in the reference user set, the similarity between the target user and each reference user in the reference user set is determined.

[0018] Based on the similarity between the target user and each reference user in the reference user set, one or more users in the reference user set whose similarity to the target user is higher than a preset threshold are selected as the neighborhood user set.

[0019] Based on the target user's behavioral data, the behavioral data of each neighboring user in the neighboring user set, and multi-dimensional environmental data, determine the similarity between the evaluation environment of the target user when evaluating each target object and the evaluation environment of each neighboring user in the neighboring user set when evaluating each target object.

[0020] Based on the target user's behavioral data, the behavioral data of each reference user in the reference user set, the similarity between the target user and each neighboring user in the neighboring user set, and the similarity between the evaluation environment of the target user when evaluating each target object and the evaluation environment of each neighboring user in the neighboring user set when evaluating each target object, the first predicted score of the target user for each target object in the target object set is determined.

[0021] In one implementation, determining the similarity between the target user and each reference user in the reference user set based on the target user's behavioral data and the behavioral data of each reference user in the reference user set includes:

[0022] Based on the target user's behavioral data, determine the target user's evaluation value for each target object to be pushed to, and determine the average value of each target object to be pushed to in the set of target objects to be pushed to;

[0023] Based on the behavioral data of each reference user in the reference user set, determine the evaluation value of each reference user in the reference user set for each object to be pushed, and determine the average value of the evaluation values ​​of each object to be pushed in the set formed by each reference user in the reference user set for the objects to be pushed.

[0024] The similarity between the target user and each reference user in the reference user set is determined based on the target user's rating of each target object, the average rating of each target object in the set formed by the target user, the rating of each target object by each reference user in the reference user set, and the average rating of each target object in the set formed by each reference user in the reference user set.

[0025] In one implementation, the similarity between the evaluation environment of the target user and that of each neighboring user in the neighboring user set is determined based on the target user's behavioral data, the behavioral data of each neighboring user in the neighboring user set, and multi-dimensional environmental data, including:

[0026] The behavioral data of the target user is correlated with the multi-dimensional environmental data to obtain the contextual environment dataset when the target user evaluates each object to be pushed, and to determine the weight value of each contextual environment in the contextual environment dataset when the target user evaluates each object to be pushed.

[0027] The behavioral data of each neighboring user in the neighboring user set is correlated with the multi-dimensional environmental data to obtain the contextual environment dataset when each neighboring user in the neighboring user set evaluates each object to be pushed, and to determine the weight value of each contextual environment data in the contextual environment dataset when each neighboring user in the neighboring user set evaluates each object to be pushed.

[0028] Based on the contextual environment dataset when the target user evaluates each object to be pushed, the weight value of each contextual environment in the contextual environment dataset when the target user evaluates each object to be pushed, the contextual environment dataset when each neighboring user in the neighboring user set evaluates each object to be pushed, and the weight value of each contextual environment data in the contextual environment dataset when each neighboring user in the neighboring user set evaluates each object to be pushed, the similarity between the evaluation environment when the target user evaluates the current object to be pushed and the evaluation environment when the current neighboring user in the neighboring user set evaluates the current object to be pushed is determined.

[0029] In one implementation, based on the target user's behavioral data, the supply and demand data of each target object in the target object set, and multi-dimensional environmental data, a second predicted score for each target object in the target object set is determined, including:

[0030] Based on the target user's behavior data, one or more objects that have been consumed in the target user's behavior data are taken as the historical consumption object set;

[0031] Based on the supply and demand data of each push object in the historical consumer object set, the target user's preference for each historical consumer object in the historical consumer object set, and multi-dimensional environmental data, the user's preference for each historical consumer object in the historical consumer object set is determined.

[0032] Based on the supply and demand data of each target object in the target object set and multi-dimensional environmental data, the target user's preference for each target object in the target object set is determined.

[0033] Based on the target user's preference for each historical consumer in the historical consumer set and each target consumer in the target set, the similarity between each target consumer in the target set and each historical consumer in the historical consumer set is determined.

[0034] Based on the target user's behavioral data and the similarity between each push object in the target object set and each historical consumption object in the historical consumption object set, a second predicted score for each push object in the target user's target object set is determined.

[0035] In one implementation, determining the target user's preference for each object in the push-to-object set based on supply and demand data and multi-dimensional environmental data includes:

[0036] The supply and demand data of each object to be pushed in the set of objects to be pushed are associated with the multi-dimensional environmental data to obtain the context environment data associated with each object to be pushed, which serves as the context environment dataset corresponding to each object to be pushed.

[0037] Determine the amount of context environment data in the context environment dataset corresponding to each object to be pushed;

[0038] Determine the association frequency between each context environment data and each push object in the context environment dataset corresponding to each push object;

[0039] Based on the association frequency between each context environment data in the context environment dataset corresponding to each push object and each push object, determine the first weight value of each context environment data in the context environment dataset corresponding to each push object.

[0040] Based on the number of context environment data in the context environment dataset corresponding to each target object and the first weight value of each context environment data, the target user's preference for each target object in the target object set is determined under each context environment data in the context environment dataset corresponding to each target object.

[0041] In one implementation, determining the target user's preference for each historical consumer in the historical consumer set is based on the historical consumer set, the supply and demand data of each target in the target set, and multi-dimensional environmental data, including:

[0042] The supply and demand data of each object to be pushed in the set of objects to be pushed are associated with the multi-dimensional environmental data to obtain the context environment data associated with each object to be pushed, which serves as the context environment dataset corresponding to each object to be pushed.

[0043] Determine the association frequency between each context environment data in the context environment dataset corresponding to each object to be pushed and each historical consumer object in the historical consumer object set;

[0044] Based on the association frequency between each context environment data in the context environment dataset corresponding to each object to be pushed and each historical consumer object, determine the second weight value of each context environment data in the context environment dataset corresponding to each object to be pushed.

[0045] Based on the number of context environment data in the context environment dataset corresponding to each target object and the second weight value of each context environment data, the target user's preference for each historical consumer object in the historical consumer object set is determined under each context environment data in the context environment dataset corresponding to each target object.

[0046] In one implementation, based on the target user's behavioral data and the similarity between each push object in the target object set and each historical consumer object in the historical consumer object set, a second predicted score for each push object in the target user's target object set is determined, including:

[0047] Determine the membership degree of the target user's evaluation value to each target object in the target user's behavioral data;

[0048] Based on the similarity between each push object in the target object set and each historical consumption object in the historical consumption object set, and the membership degree of the target user's evaluation value of each push object in the target user's behavior data, the second predicted rating of the target user for each push object in the target object set is determined.

[0049] In one implementation, determining the predicted score of each target user in the set of target users to be pushed to is based on the first predicted score of each target user in the set of target users to be pushed to and the second predicted score of each target user in the set of target users to be pushed to, including:

[0050] Determine the rating adjustment factors for the first and second predicted scores;

[0051] Based on the rating adjustment factor, the first predicted rating of each target object in the target user's target object set, and the second predicted rating of each target object in the target user's target object set, the target predicted rating of each target object in the target user's target object set is determined.

[0052] A device for pushing a target object, comprising:

[0053] The acquisition module is used to acquire behavioral data of target users, behavioral data of each reference user in the reference user set, supply and demand data of each object to be pushed in the object to be pushed set, and multi-dimensional environmental data.

[0054] The calculation module is used to determine the target prediction score of the target user for each object in the target object set based on the target user's behavior data, the behavior data of each reference user in the reference user set, the supply and demand data of each object to be pushed in the target object set, and multi-dimensional environmental data.

[0055] The determination module is used to determine at least one object to be pushed to the target user based on the target prediction score of each object in the target user's set of objects to be pushed;

[0056] A push module is used to push the at least one object to the target user.

[0057] An electronic device includes a processor and a memory for storing processor-executable instructions, wherein the processor, when executing the instructions, implements the steps of any of the methods described above.

[0058] A computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implement the steps of any of the methods described above.

[0059] This application provides a method and apparatus for pushing target objects. It acquires behavioral data of a target user, behavioral data of reference users in a reference user set, supply and demand data of objects to be pushed in a target object set, and multi-dimensional environmental data. Based on the behavioral data of the target user, the behavioral data of reference users in the reference user set, the supply and demand data of objects to be pushed in the target object set, and the multi-dimensional environmental data, it determines the target user's predicted target score for each object to be pushed in the target object set. Based on the predicted target score of each object to be pushed in the target object set, it determines at least one object to push to the target user. The at least one object is then pushed to the target user. By converting unstructured data from various dimensions of the user into scores, unstructured data from various dimensions can be processed uniformly, achieving a transformation from unstructured to structured data before pushing target objects. This solves the technical problem of low data utilization caused by the ineffective use of existing unstructured data, and achieves the technical effect of improving the matching degree between push objects and users based on unstructured data. Attached Figure Description

[0060] To more clearly illustrate the embodiments of this specification, the accompanying drawings used in the embodiments will be briefly introduced below. The drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0061] Figure 1 This is a flowchart of one embodiment of the push method for the target object provided in this application;

[0062] Figure 2 This is a flowchart illustrating a method for determining a first predicted score in one embodiment of the push method for the target object provided in this application;

[0063] Figure 3 This is a flowchart illustrating a method for determining a second predicted score in one embodiment of the target object push method provided in this application;

[0064] Figure 4 This is a flowchart illustrating one embodiment of the product recommendation system provided in this application;

[0065] Figure 5 This is a hardware structure block diagram of an electronic device for a target object push method provided in this application;

[0066] Figure 6This is a schematic diagram of the module structure of one embodiment of the target object push device provided in this application. Detailed Implementation

[0067] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0068] To address the technical problem of low recommendation accuracy in recommendation systems due to the inability to uniformly process data from various dimensions under unstructured data conditions, this application provides a method and apparatus for pushing target objects. Although this application provides method operation steps or apparatus structures as shown in the following embodiments or accompanying drawings, more or fewer operation steps or module units may be included in the method or apparatus based on conventional or non-inventive methods. In steps or structures where there is no logically necessary causal relationship, the execution order of these steps or the module structure of the apparatus is not limited to the execution order or module structure described in the embodiments and accompanying drawings of this application. When the method or module structure is applied in actual devices or terminal products, it can be executed sequentially or in parallel according to the method or module structure shown in the embodiments or accompanying drawings (e.g., in a parallel processor or multi-threaded processing environment, or even a distributed processing environment).

[0069] Specifically, such as Figure 1 As shown, the above-mentioned method for pushing to the target object may include the following steps:

[0070] S101: Obtain behavioral data of target users, behavioral data of each reference user in the reference user set, supply and demand data of each target object in the target object set, and multi-dimensional environmental data.

[0071] The behavioral data can include user evaluation behavior, purchase behavior, etc.; the supply and demand data can include transaction records and transaction locations of the target object; and the multi-dimensional environmental data can include weather, temperature, time, etc.

[0072] Furthermore, the acquired behavioral data of target users, behavioral data of reference users in the reference user set, supply and demand data of each target object in the target object set, and multi-dimensional environmental data can be preprocessed. Specifically, the unstructured data in the behavioral data of target users, behavioral data of reference users in the reference user set, supply and demand data of each target object in the target object set, and multi-dimensional environmental data can be scored and classified using the analytic hierarchy process (AHP), so that unstructured data can be processed uniformly. Dimensionality reduction can also be performed on high-dimensional data to simplify subsequent data processing.

[0073] Specifically, missing data in the behavioral data of target users and reference users in the reference user set can be supplemented based on the demographic data of each reference user in the target user set, in order to solve the data sparsity problem and improve the accuracy of target audience push.

[0074] Specifically, the demographic similarity between two users can be calculated using the following formula:

[0075]

[0076] Where cSim is the demographic similarity, and D a D is the demographic feature vector of the first user. b Let m be the demographic feature vector of the second user, where m is the dimension of the demographic feature and M is the size of the dimension of the demographic feature.

[0077] Specifically, after calculating the demographic similarity between any two users in the target user and reference user sets, for users with missing data, the missing items can be supplemented by combining the behavioral data of one or more users whose demographic similarity with the missing data user is greater than a second preset threshold. For example, when user A lacks evaluation data for product B, 10 users with a demographic similarity greater than 0.8 with user A are selected, and the average of the evaluation values ​​of these 10 users for product B is taken as user A's evaluation value for product B. Alternatively, the evaluation value with the highest frequency among the evaluation values ​​of these 10 users for product B can be taken as user A's evaluation value for product B. In actual implementation, an appropriate second preset threshold and data determination method can be selected according to the actual situation and needs, which will not be elaborated here.

[0078] S102: Based on the target user's behavior data, the behavior data of each reference user in the reference user set, the supply and demand data of each target object in the target object set, and multi-dimensional environmental data, determine the target user's target prediction score for each target object in the target object set.

[0079] Specifically, determining the target user's predicted score for each target in the push notification set can be divided into three parts. First, based on the target user's behavioral data, the behavioral data of each reference user in the reference user set, and multi-dimensional environmental data, the target user's first predicted score for each target in the push notification set can be determined. Then, based on the target user's behavioral data, the supply and demand data of each target in the push notification set, and multi-dimensional environmental data, the target user's second predicted score for each target in the push notification set can be determined. Finally, based on the target user's first and second predicted scores for each target in the push notification set, the target user's final predicted score for each target in the push notification set can be determined.

[0080] Furthermore, the first predictive score is primarily determined based on the similarity between the target user and a user group with shared interests; specifically, for example... Figure 2 As shown, determining the first predicted score for each target user in the target user set based on the target user's behavioral data, the behavioral data of each reference user in the reference user set, and multi-dimensional environmental data may include the following steps:

[0081] S201: Based on the behavioral data of the target user and the behavioral data of each reference user in the reference user set, determine the similarity between the target user and each reference user in the reference user set.

[0082] The similarity between the target user and each reference user in the reference user set is determined based on the target user's evaluation values ​​of each target object in the target user's behavioral data and the evaluation values ​​of each reference user in the reference user set's behavioral data of each target object. Specifically, determining the similarity between the target user and each reference user in the reference user set may include the following steps:

[0083] Based on the target user's behavioral data, determine the target user's evaluation value for each target object to be pushed to, and determine the average value of each target object to be pushed to in the set of target objects to be pushed to;

[0084] Based on the behavioral data of each reference user in the reference user set, determine the evaluation value of each reference user in the reference user set for each object to be pushed, and determine the average value of the evaluation values ​​of each object to be pushed in the set formed by each reference user in the reference user set for the objects to be pushed.

[0085] The similarity between the target user and each reference user in the reference user set is determined based on the target user's rating of each target object, the average rating of each target object in the set formed by the target user, the rating of each target object by each reference user in the reference user set, and the average rating of each target object in the set formed by each reference user in the reference user set.

[0086] In implementation, the similarity between the target user and each reference user in the reference user set can be calculated using the following formula:

[0087]

[0088] Among them, u a Indicates the target user, u b Sim1(u) represents the reference user in the reference user set. a ,u b ) represents the similarity between the target user and the reference user, i t I represents the current object to be pushed, and I represents a collection of multiple objects to be pushed. This represents the target user's rating of the currently targeted object from the target user's behavioral data. This represents the average rating of each target object in the set formed by the target user's behavioral data regarding the target object. This represents the rating of the reference object to the current target object based on the reference user's behavior data. This represents the average rating of each target object in the set formed by the reference user's behavior data.

[0089] S202: Based on the similarity between the target user and each reference user in the reference user set, one or more users in the reference user set whose similarity to the target user is higher than a preset threshold are selected as a neighborhood user set.

[0090] S203: Based on the target user's behavior data, the behavior data of each neighboring user in the neighboring user set, and multi-dimensional environmental data, determine the similarity of the evaluation environment between the target user's evaluation of each object to be pushed and the evaluation environment of each neighboring user in the neighboring user set when evaluating each object to be pushed.

[0091] Specifically, the similarity of the evaluation environment between the target user's evaluation of each target object and the evaluation environment of each neighboring user in the neighboring user set when evaluating each target object can be determined according to the following steps, including:

[0092] The behavioral data of the target user is correlated with the multi-dimensional environmental data to obtain the contextual environment dataset when the target user evaluates each object to be pushed, and to determine the weight value of each contextual environment in the contextual environment dataset when the target user evaluates each object to be pushed.

[0093] The behavioral data of each neighboring user in the neighboring user set is correlated with the multi-dimensional environmental data to obtain the contextual environment dataset when each neighboring user in the neighboring user set evaluates each object to be pushed, and to determine the weight value of each contextual environment data in the contextual environment dataset when each neighboring user in the neighboring user set evaluates each object to be pushed.

[0094] Based on the contextual environment dataset when the target user evaluates each object to be pushed, the weight value of each contextual environment in the contextual environment dataset when the target user evaluates each object to be pushed, the contextual environment dataset when each neighboring user in the neighboring user set evaluates each object to be pushed, and the weight value of each contextual environment data in the contextual environment dataset when each neighboring user in the neighboring user set evaluates each object to be pushed, the similarity between the evaluation environment when the target user evaluates the current object to be pushed and the evaluation environment when the current neighboring user in the neighboring user set evaluates the current object to be pushed is determined.

[0095] Furthermore, in implementation, the similarity between the evaluation environment of the target user when evaluating the current object to be pushed and the evaluation environment of the current neighboring users in the neighboring user set when evaluating the current object to be pushed can be calculated according to the following formula:

[0096]

[0097] Where WJsim represents the similarity of the evaluation environment between the target user and each neighboring user in the neighboring user set, and C a C represents the contextual information used by the target user to rate the currently pushed object. d This represents the contextual environment data when neighboring users evaluate the current push notification object, where w represents the weight value of the contextual environment data when evaluating each push notification object, and W = {w1, w2, ..., w...} l ,...,w L},and L represents the total amount of contextual data when the target user and neighboring users evaluate the current object to be pushed.

[0098] S204: Based on the similarity between the target user and each reference user in the reference user set, and the similarity between the target user and each neighboring user in the neighboring user set in the evaluation environment, determine the first predicted score of the target user for each object to be pushed in the target object set.

[0099] Specifically, the first predicted score for each target user in the target user set can be calculated using the following formula:

[0100]

[0101] Among them, P CCFG (u a i t ) represents the target user's first predicted rating of the currently targeted object, u d N represents the neighboring users of the neighboring user set. a This represents the set of neighboring users.

[0102] Specifically, based on the first predicted score of each target object in the target object set, the scope of the target objects can be initially determined. For example, a preset number of target objects with higher first predicted scores can be selected from the target object set, or target objects with first predicted scores greater than a third preset threshold can be selected. In the subsequent process of determining the second predicted score, the second predicted score can be determined only for the target objects filtered by the first predicted score. However, it is worth noting that in actual implementation, the decision on whether to perform preliminary filtering of target objects based on the first predicted score, the number of target objects obtained from the filtering, and the filtering method can be determined according to actual needs. This application does not impose any limitations on these aspects.

[0103] Furthermore, the contextual data weight values ​​for evaluating each target object in step S203 above can be determined according to the following steps:

[0104] S1: Associate the target user's behavioral data with multi-dimensional environmental data to obtain the target user's historical consumption data and satisfaction under various environmental data conditions;

[0105] S2: Based on the target user's historical consumption data and satisfaction under various environmental data, construct a training set with the actual ratings of each target user to be pushed to;

[0106] S3: Determine the initial weight values ​​of the contextual environment data when evaluating each target object;

[0107] S4: Based on the initial weight values ​​of the contextual environment data, determine the similarity between the contextual environment data of the target user and each user in the user neighborhood set when they evaluate the target product;

[0108] S5: Calculate the first predicted rating of the target user for each target product based on the similarity between the target user and the contextual environment data of each user in the user neighborhood set when they evaluate the target product.

[0109] S6: Based on the target user's first predicted rating for each target product and the actual ratings for each product in the training set, calculate the appropriate score according to the following formula:

[0110]

[0111] Where fs is the appropriate score, T R For the training set, Act(i t Pred(i) represents the target user's actual rating of the pushed content. t The first predicted rating for the target user regarding the pushed content is 0.

[0112] S7: If the appropriate score does not meet the first preset condition, adjust the weight values ​​of each context environment data, recalculate the first predicted score, and then recalculate the appropriate score based on the first predicted score and the actual score in the training set until the appropriate score meets the first preset condition or the number of iterations reaches the first preset value.

[0113] S8: If the appropriate score meets the first preset condition or the number of iterations reaches the first preset value, terminate the iteration and use the weight value of the last calculated appropriate score as the final weight value of each context environment data.

[0114] Specifically, the first preset condition can be that the calculated appropriate score needs to be between 0.8 and 0.95, and the first preset value of the number of iterations can be 2, 3, etc. However, it is worth noting that in actual implementation, the above-mentioned first preset condition and number of iterations can be determined according to the actual situation and on-site requirements to determine a suitable appropriate score range and the first preset value, which is not limited in this application.

[0115] By determining the weight values ​​of each contextual data using the above method, the importance and impact of each contextual data on the target user can be fully considered, making the final weight values ​​of each contextual data more fair and reasonable.

[0116] Furthermore, a second predicted score for the target user can be determined based on the similarity between the target user's historical consumption and the target to be pushed to, thereby further improving the accuracy of the target predicted score. Specifically, for example... Figure 3As shown, determining the second predicted score for each target user in the target user set based on the target user's behavioral data, the supply and demand data of each target user in the target user set, and multi-dimensional environmental data may include the following steps:

[0117] S301: Based on the target user's behavior data, one or more objects that have been consumed in the target user's behavior data are taken as a set of historical consumption objects.

[0118] S302: Based on the supply and demand data of each push object in the historical consumer object set, the push object set, and multi-dimensional environmental data, determine the target user's preference for each historical consumer object in the historical consumer object set.

[0119] In a specific multi-dimensional environment with rich data, the importance of each historical consumption item to the target user can be determined by assessing the target user's preference for each historical consumption item. Specifically, the target user's preference for each historical consumption item in the set can be determined by the following steps:

[0120] The supply and demand data of each object to be pushed in the set of objects to be pushed are associated with the multi-dimensional environmental data to obtain the context environment data associated with each object to be pushed, which serves as the context environment dataset corresponding to each object to be pushed.

[0121] Determine the association frequency between each context environment data in the context environment dataset corresponding to each object to be pushed and each historical consumer object in the historical consumer object set;

[0122] Based on the association frequency between each context environment data in the context environment dataset corresponding to each object to be pushed and each historical consumer object, determine the second weight value of each context environment data in the context environment dataset corresponding to each object to be pushed.

[0123] Based on the number of context environment data in the context environment dataset corresponding to each target object and the second weight value of each context environment data, the target user's preference for each historical consumer object in the historical consumer object set is determined under each context environment data in the context environment dataset corresponding to each target object.

[0124] Specifically, the preference of the aforementioned target users for each historical consumption item in the consumption object set can be calculated using the following formula:

[0125]

[0126] Among them, c th This represents the context environment data in the context environment dataset corresponding to the object to be pushed. This represents the target user's preference for current historical consumption objects within the current context of the data. j Represents the current historical consumption object, rank2(c th ) represents the second weight value of the context environment data in the context environment dataset corresponding to the current object to be pushed, |H| represents the number of context environment data in the context environment dataset corresponding to the current object to be pushed, and ρ>1 is the Gaussian parameter.

[0127] S303: Based on the supply and demand data of each object to be pushed in the object set and multi-dimensional environmental data, determine the target user's preference for each object to be pushed in the object set.

[0128] In a specific multi-dimensional environment with rich data, the importance of each historical consumption item to the target user can be determined by identifying the target user's preference for each item in the target user set. Specifically, the target user's preference for each item in the target user set can be determined in the following ways:

[0129] The supply and demand data of each object to be pushed in the set of objects to be pushed are associated with the multi-dimensional environmental data to obtain the context environment data associated with each object to be pushed, which serves as the context environment dataset corresponding to each object to be pushed.

[0130] Determine the amount of context environment data in the context environment dataset corresponding to each object to be pushed;

[0131] Determine the association frequency between each context environment data and each push object in the context environment dataset corresponding to each push object;

[0132] Based on the association frequency between each context environment data in the context environment dataset corresponding to each push object and each push object, determine the first weight value of each context environment data in the context environment dataset corresponding to each push object.

[0133] Based on the number of context environment data in the context environment dataset corresponding to each target object and the first weight value of each context environment data, the target user's preference for each target object in the target object set is determined under each context environment data in the context environment dataset corresponding to each target object.

[0134] Specifically, the preference of the target user for each object in the target object set can be calculated using the following formula:

[0135]

[0136] Among them, c thThis represents the context environment data in the context environment dataset corresponding to the object to be pushed. This represents the target user's preference for the currently pushed object within the current context of the data. t This represents the current object to be pushed, rank1(c h ) represents the first weight value of the context environment data in the context environment dataset corresponding to the current object to be pushed, |H| represents the number of context environment data in the context environment dataset corresponding to the current object to be pushed, and ρ>1 is the Gaussian parameter.

[0137] S304: Based on the target user's preference for each historical consumer in the historical consumer set and each target object in the target object set, determine the similarity between each target object in the target object set and each historical consumer in the historical consumer set.

[0138] When implementing this, the similarity between each push object in the set of objects to be pushed and each historical consumer object in the set of historical consumers can be calculated using the following formula:

[0139]

[0140] Among them, Sim2(i t i j The similarity between the current target object and each historical target object in the historical target object set is denoted as .

[0141] S305: Based on the target user's behavioral data and the similarity between each push object in the target object set and each historical consumption object in the historical consumption object set, determine the target user's second predicted score for each push object in the target object set.

[0142] Because the upper and lower bounds of the rating levels for each target user's evaluation of the objects to be pushed may be inconsistent in the target user's behavioral data, this inconsistency leads to uncertainty in the target user's evaluation values ​​for each target user's evaluation values. Therefore, to avoid the impact of this uncertainty, it is necessary to determine the membership degree of the target user's evaluation values ​​for each target user's evaluation values ​​in the target user's behavioral data. Specifically, the membership degree of the target user's evaluation values ​​for each target user's evaluation values ​​in the target user's behavioral data can be calculated using the following formula:

[0143]

[0144] Wherein, min represents the lower limit of the rating level when the target user evaluates the current push object, and max represents the upper limit of the rating level when the target user evaluates the current push object.

[0145] After determining the membership degree of the target user's rating values ​​for each target object in the target user's behavioral data, based on the similarity between each target object in the target object set and each historical consumer object in the historical consumer object set, and the membership degree of the target user's rating values ​​for each target object in the target user's behavioral data, a second predicted rating for each target object in the target object set is determined, including:

[0146] The second predicted score for each target user in the target user set is calculated using the following formula:

[0147]

[0148] Among them, P CRM (u a i t Sim2(i) represents the target user's second predicted rating of the currently targeted object. t i j ) represents the similarity between each push object in the target object set and each historical consumer object in the historical consumer object set, I J It represents a set of multiple historical consumption objects. This represents the membership degree of the target user's rating value to each target object in the target user's behavioral data.

[0149] The aforementioned scheme effectively combines target user behavior data, supply and demand data of the objects to be pushed, and multi-dimensional environmental data to determine the target user's preference for each object to be pushed and each historical consumption object in the target user's behavior data. This accurately determines the similarity between each object in the push object set and each historical consumption object in the historical consumption object set. Then, it fully considers the uncertainty of the target user's evaluation value of each object in the push object set to improve the accuracy of the second predictive score. The second predictive score can further optimize the object push results. For example, based on the objects to be pushed selected by the first predictive score, the second predictive score can be used to further filter the objects to be pushed, thereby improving the accuracy of object push. However, it is worth noting that in actual implementation, whether to further filter the objects to be pushed using the second predictive score, the number of objects to be pushed obtained through filtering, and the filtering method can be determined according to actual needs; this application does not impose any limitations on these aspects. Furthermore, before determining the target predicted score based on the aforementioned first and second predicted scores, in order to adjust the contribution of the first and second predicted scores, a score adjustment factor for the first and second predicted scores can be determined first. Then, based on the score adjustment factor, the first predicted score of each target object in the target user's target object set, and the second predicted score of each target object in the target user's target object set, the target predicted score for each target object in the target user's target object set is determined, including:

[0150] The target predicted score for each target user in the target user set can be calculated using the following formula:

[0151] P CRCF (u a i t )=αP CCFG (u a i t )+(1-α)P CRM (u a i t )

[0152] Among them, P CRCF (u a i t ) represents the target user's predicted rating of the currently targeted object, α represents the rating adjustment factor, and P CCFG (u a i t P represents the target user's first predicted rating of the currently targeted object. CRM (u a i t ) represents the target user's second predicted rating of the object to be pushed to.

[0153] Specifically, under different multi-dimensional environmental data, the first prediction score is calculated based on the similarity between different users, while the second prediction score is calculated based on the similarity between the target object and the target user's historical consumption objects. Therefore, the first and second prediction scores have different focuses. The target prediction score is determined by combining the similarity between users and the similarity between the target object and the target user's historical consumption objects. It can simultaneously leverage the advantages of the first and second prediction scores, improve the accuracy of the target object push, and provide customized target object push services for the target user.

[0154] S103: Based on the target user's target prediction score for each object in the set of objects to be pushed to, determine at least one object to be pushed to the target user.

[0155] Specifically, based on the target user's target prediction score for each target object in the target object set, one or more objects can be selected and pushed to the target user according to the second preset condition. For example, three objects with target prediction scores higher than the fifth preset threshold can be pushed to the target user for selection, or the object with the highest target prediction score can be pushed directly to the target user. In actual implementation, the appropriate number of objects to be pushed and the second preset condition can be set according to the actual situation and needs, which is not limited in this application.

[0156] S104: Push the at least one object to the target user.

[0157] In the example above, the system acquires behavioral data of the target user, behavioral data of reference users in the reference user set, supply and demand data of objects to be pushed to in the object-to-push set, and multi-dimensional environmental data. Based on these data, the system determines the target user's predicted score for each object in the object-to-push set. Then, based on these predicted scores, it determines at least one object to push to the target user. Finally, it pushes this at least one object to the target user. This method allows for the prediction of scores for each object, accurately determining the target object to push to the target user, thereby improving user experience and operational efficiency.

[0158] The above method will be described below with reference to a specific embodiment. However, it is worth noting that this specific embodiment is only for better illustration of this application and does not constitute an improper limitation of this application.

[0159] Specifically, such as Figure 4As shown, a context-aware recommendation method optimized by a fuzzy genetic algorithm in this example may include the following steps:

[0160] S401: Collect behavioral data of target users, behavioral data of various reference users, supply and demand data of goods, and multi-dimensional environmental data.

[0161] S402: Preprocess the user's behavior data, the behavior data of each reference user, the supply and demand data of goods, and the multi-dimensional environmental data.

[0162] Specifically, in order to comprehensively evaluate unstructured data such as products and convenience store environment, a comprehensive scoring mechanism based on the analytic hierarchy process was introduced to score and classify unstructured data in user behavior data, behavior data of various reference users, supply and demand data of products, and multi-dimensional environmental data.

[0163] Furthermore, since a large number of information dimensions increases the load on the model and affects the performance of the algorithm, in order to simplify the calculation, the high-dimensional information in user behavior data, behavior data of various reference users, supply and demand data of goods, and multi-dimensional environmental data can be reduced in dimensionality through reduction methods.

[0164] Furthermore, data sparsity can negatively impact the accuracy of product recommendation systems. When available data only supports comprehensive evaluation of a subset of products, and these evaluations are not comprehensive enough in terms of dimensions, data sparsity occurs. This sparsity affects the quality of the interaction matrix between users, products, and the contextual environment, leading to lower accuracy and reliability of the recommendation results. To address data sparsity, demographic data based on member attribute data can be used to predict and fill in missing data items among users, products, and the contextual environment.

[0165] Specifically, demographic similarity between users can be calculated using the following formula:

[0166]

[0167] Among them, cosineSim(D a D b ) represents the demographic similarity, D a For user u a The demographic feature vector, D b For user u b The demographic feature vector, where j is the dimension of the demographic feature and n is the size of the dimension of the demographic feature.

[0168] Furthermore, since users with similar demographic characteristics often have similar preferences, after calculating the demographic similarity between users, based on a user with missing data, the target user's rating score for the target product can be predicted based on the average rating score of users with a demographic similarity greater than a second preset threshold for the target product. Similarly, the rating scores of the target user and each reference user for each product can be supplemented using the above method.

[0169] S403: Based on the preprocessed data, predict the target user's rating for each product.

[0170] This example employs a context-aware product recommendation system architecture based on fuzzy genetic algorithms. This architecture combines the advantages of collaborative filtering and recursive algorithms, aiming to improve prediction accuracy and address the problems encountered by traditional two-dimensional recommendation techniques when integrating contextual data. By applying fuzzy genetic algorithms to a context-aware recursive collaborative filtering model, the specific impact of different contextual data on user ratings can be distinguished. Specifically, the context-aware product recommendation system architecture based on fuzzy genetic algorithms may include:

[0171] S4031: Calculate the target user's first predicted rating for each target product using a collaborative filtering model.

[0172] Specifically, the first predicted rating of each target product by the target user is calculated using a collaborative filtering model, including:

[0173] S1: Based on the rating scores and average rating scores of each product given by the target user and each reference user, the similarity between the target user and each reference user is calculated according to the following formula, including:

[0174]

[0175] Among them, u a For target users, u b For reference users, Sim1(u a ,u b Let ) represent the similarity between the target user and the reference user, and I represent the product set, i t For the current product, The rating score of the target user for the current product is derived from the target user's behavioral data. This represents the average rating score given by the target users to each product, based on their behavioral data. To reference user behavior data, we need to consider the user's rating score for the current product. The average rating score given by users to the current product is used as a reference in user behavior data.

[0176] S2: Based on the similarity between the target user and each reference user, at least one reference user whose similarity to the target user is greater than a preset threshold is taken as the target user's neighborhood user set.

[0177] S3: Associate the target user's behavioral data with multi-dimensional environmental data to determine the contextual environmental data when the target user evaluates each product;

[0178] S4: Associate the behavioral data of reference users with multi-dimensional environmental data to determine the contextual environmental data when each neighboring user in the neighboring user set evaluates each product;

[0179] S5: Based on the contextual environment data of the target user and each neighboring user when evaluating each product, calculate the similarity between the contextual environment data of the target user and each neighboring user when evaluating each product according to the following formula:

[0180]

[0181] Among them, u d For neighboring users in the neighboring user set, WJsim represents the similarity between the target user and neighboring users in the contextual environment data when they evaluate the current product, and C represents the similarity between the target user and neighboring users in the contextual environment data. a C provides contextual information about the target user's evaluation of the current product. d This refers to the contextual environment data when neighboring users rate the current product, where W = {w1, w2, ..., w...} l ,...,w L} represents the weight value of the contextual data when evaluating the current product, and L represents the total amount of contextual data when the target user and neighboring users evaluate the current product.

[0182] S6: Based on the similarity between the target user and each neighboring user in the neighboring user set, and the similarity between the target user and the contextual data of each neighboring user in the neighboring user set when evaluating the current product, calculate the target user's first predicted rating for each product according to the following formula:

[0183]

[0184] Among them, P CCFG (u a i t ) represents the target user's first predicted rating for the current product, N. a This refers to the set of neighboring users of the target user.

[0185] Furthermore, the weight values ​​of each context data in step S5 above can be determined according to the following steps:

[0186] Based on the target users' historical consumption data and satisfaction levels across various sites and environments, as well as product data, a training set with actual ratings for each product is constructed.

[0187] Determine the initial weight values ​​for each context data in the context environment dataset;

[0188] Based on the initial weight values, the similarity between the target user and the contextual environment data of each neighboring user when evaluating each product is calculated.

[0189] Based on the similarity between the target user and the contextual environment data of each neighboring user when evaluating each product, the first predicted rating of the target user for each product is calculated.

[0190] Based on the target user's first predicted rating for each product and the actual ratings for each product in the training set, the appropriate score is calculated according to the following formula:

[0191]

[0192] Where fs is the appropriate score, T R For the training set, Act(i t For target users, product i t The actual rating, Pred(i t For target users, product i t The first predicted score.

[0193] If the appropriate score does not meet the first preset condition, adjust the weight values ​​of each context environment data, recalculate the first predicted score, and then recalculate the appropriate score based on the first predicted score and the actual score obtained in the training set, until the appropriate score meets the first preset condition or the number of iterations reaches the first preset value.

[0194] If the appropriate score meets the first preset condition or the number of iterations reaches the first preset value, the iteration is terminated, and the weight value of the last calculated appropriate score is used as the final weight value of each context environment data.

[0195] S4032: Calculate the target user's second predicted rating for each target product using a recursive algorithm model.

[0196] To effectively integrate the contextual data frequently accessed by target users with their preferences for specific products, thereby improving the accuracy of predictive ratings, this example analyzes different contextual data related to the target user's past purchasing behavior. For instance, the product recommendation system can recommend suitable products based on combinations of environmental data reflecting the target user's past preferences, such as time (morning, afternoon, evening, night), travel purpose (daily commuting, weekend outing, long-distance freight, holiday long-distance travel), and weather (sunny, cloudy, rainy / snowy)}.

[0197] Specifically, we can first determine the importance of each product to the target user under different contextual data environments, including:

[0198] S1: Associate the supply and demand data of each commodity with multi-dimensional environmental data to obtain the contextual environment data associated with each commodity, which serves as the contextual environment dataset for each commodity.

[0199] S2: Based on the contextual dataset corresponding to each product, calculate the importance of each product to the target user under different contextual data according to the following formula:

[0200]

[0201] Among them, c pk For the current context environment data, To determine the importance of the current product to the target user within the current context, rank1(c pk ) represents the first-level position of the current context data, and 1 ≤ rank1(c pk )≤|L|, where |L is the size of the context dataset corresponding to the current product, and ρ>1 is the Gaussian parameter.

[0202] Specifically, the first-level position of the aforementioned current contextual data can be determined based on the frequency of association between the current contextual data and the current product.

[0203] S3: Obtain the target user's past purchasing behavior based on the target user's behavioral data;

[0204] S4: Based on the target user's past purchasing behavior and the contextual dataset for each product, calculate the importance of each product previously purchased by the target user to the target user under different contextual data using the following formula:

[0205]

[0206] in, To determine the importance of previously purchased items to the target user within the current context, rank2(c) is used.pk ) represents the second-level position of the context condition, and 1 ≤ rank2(c pk )≤|L|.

[0207] Specifically, the second-level position of the aforementioned current contextual data can be determined based on the frequency of association between the current contextual data and previously purchased products.

[0208] S5: To effectively integrate contextual conditions and target users' preferences for specific products, the similarity between each product in different contextual environments and each product previously purchased by the target user can be calculated using the following formula, based on the importance of each product to the target user in different contextual environments and the importance of each product previously purchased by the target user:

[0209]

[0210] S6: Based on the similarity between each product and each product previously purchased by the target user under different contextual conditions, calculate the target user's second predicted rating for each product according to the following formula:

[0211]

[0212] Among them, P CRM (u a i t I represents the target user's second predicted rating for the current product. a This is a collection of all the items that the target user has previously purchased. This refers to the membership degree of the target user's rating score for the current product in the target user's behavioral data.

[0213] Specifically, considering that target users may rate different products differently under different contextual data, for example, in some product recommendation systems, the upper limit of the rating score may be 5 points and the lower limit may be 1 point, while in other product recommendation systems, the upper limit may be 3 points and the lower limit may be 0 points. To avoid the impact of the uncertainty of the upper and lower limits of the rating score on the final predicted score, the membership degree of the target user's rating of each product in the target user's behavioral data can be introduced into the second predicted rating model mentioned above. Therefore, the membership degree of the target user's rating of each product can be determined according to the following formula:

[0214]

[0215] in, The membership degree of the target user's rating of the current product, min is the lower limit of the rating level when the target user evaluates the current product, and max is the upper limit of the rating level when the target user evaluates the current product.

[0216] S4033: Calculate the target user's target prediction score for each target product using a hybrid collaborative filtering and recursive algorithm.

[0217] To combine the collaborative filtering and recursive models mentioned above, compensating for their shortcomings while leveraging their strengths to further improve the accuracy of product recommendation results, the contribution of the two models can be adjusted using a factor α. Specifically, the value of α can be obtained through validation using test set data. Furthermore, the optimized hybrid collaborative filtering and recursive algorithm can be expressed as:

[0218] P CRCF (u a i t )=αP CCFG (u a i t )+(1-α)P CRM (u a i t )

[0219] Among them, P CRCF (u a i t ) represents the target user's predicted rating for the current product, α represents the rating adjustment factor, and P CCFG (u a i t P represents the target user's first predicted rating for the current product. CRM (u a i t This represents the target user's second predicted rating for the current product.

[0220] S4034: Generate a product recommendation list based on the target prediction score of each product.

[0221] In the example above, the analytic hierarchy process (AHP) is used to classify unstructured data by rating, and the fuzzy genetic algorithm is used to optimize the context-aware hybrid recommendation technology based on collaborative filtering and recursive algorithms. Based on the target user's purchase history, behavioral characteristics, and real-time contextual data (e.g., weather, holidays, time), the recommendation results of the product recommendation system are accurately determined, thereby improving the customer's shopping experience and purchase conversion rate.

[0222] The methods and embodiments provided in the above-described embodiments of this application can be executed in a mobile terminal, computer terminal, or similar computing device. Taking operation on an electronic device as an example... Figure 5 This is a hardware structure block diagram of an electronic device for a target object push method provided in this application. For example... Figure 5 As shown, the electronic device 10 may include one or more (only one is shown in the figure) processors 02 (processors 02 may include, but are not limited to, microprocessors MCUs or programmable logic devices FPGAs, etc.), a memory 04 for storing data, and a transmission module 06 for communication functions. Those skilled in the art will understand that... Figure 5 The structure shown is for illustrative purposes only and does not limit the structure of the electronic device described above. For example, electronic device 10 may also include... Figure 5 The more or fewer components shown, or having the same Figure 5 The different configurations shown.

[0223] The memory 04 can be used to store software programs and modules of application software, such as the program instructions / modules corresponding to the target object push method in this embodiment. The processor 02 executes various functional applications and data processing by running the software programs and modules stored in the memory 04, thereby realizing the target object push method of the application described above. The memory 04 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 04 may further include memory remotely located relative to the processor 02, and these remote memories can be connected to the electronic device 10 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0224] The transmission module 06 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of the electronic device 10. In one example, the transmission module 06 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission module 06 may be a Radio Frequency (RF) module, used for wireless communication with the Internet.

[0225] At the software level, the aforementioned target object push device can be as follows: Figure 6 As shown, it includes:

[0226] The acquisition module 601 is used to acquire the behavioral data of the target user, the behavioral data of each reference user in the reference user set, the supply and demand data of each push object in the push object set, and multi-dimensional environmental data.

[0227] The calculation module 602 is used to determine the target prediction score of the target user for each object in the target object set based on the target user's behavior data, the behavior data of each reference user in the reference user set, the supply and demand data of each object to be pushed in the object to be pushed set, and multi-dimensional environmental data.

[0228] The determining module 603 is used to determine at least one object to be pushed to the target user based on the target prediction score of each object in the target user's set of objects to be pushed.

[0229] The push module 604 is used to push the at least one object to the target user.

[0230] In one implementation, the calculation module 602 determines the target prediction score for each target object in the target object set based on the target user's behavior data, the behavior data of each reference user in the reference user set, the supply and demand data of each target object in the target object set, and multi-dimensional environmental data, including:

[0231] Based on the target user's behavioral data and the behavioral data of each reference user in the reference user set, determine the first predicted score of each target user in the target user set to be pushed to.

[0232] Based on the target user's behavioral data, the supply and demand data of each target object in the target object set, and multi-dimensional environmental data, the second predicted score of the target user for each target object in the target object set is determined.

[0233] Based on the first predicted score of each target object in the target object set and the second predicted score of each target object in the target object set, the target predicted score of each target object in the target object set is determined.

[0234] In one implementation, based on the target user's behavioral data and the behavioral data of each reference user in the reference user set, a first predicted score for each target user in the target user set is determined, including:

[0235] S1: Based on the behavioral data of the target user and the behavioral data of each reference user in the reference user set, determine the similarity between the target user and each reference user in the reference user set.

[0236] In one implementation, determining the similarity between the target user and each reference user in the reference user set based on the target user's behavioral data and the behavioral data of each reference user in the reference user set includes:

[0237] Based on the target user's behavioral data, determine the target user's evaluation value for each target object to be pushed to, and determine the average value of each target object to be pushed to in the set of target objects to be pushed to;

[0238] Based on the behavioral data of each reference user in the reference user set, determine the evaluation value of each reference user in the reference user set for each object to be pushed, and determine the average value of the evaluation values ​​of each object to be pushed in the set formed by each reference user in the reference user set for the objects to be pushed.

[0239] The similarity between the target user and each reference user in the reference user set is determined based on the target user's rating of each target object, the average rating of each target object in the set formed by the target user, the rating of each target object by each reference user in the reference user set, and the average rating of each target object in the set formed by each reference user in the reference user set.

[0240] S2: Based on the similarity between the target user and each reference user in the reference user set, one or more users in the reference user set whose similarity to the target user is higher than a preset threshold are selected as the neighborhood user set.

[0241] S3: Based on the target user's behavior data, the behavior data of each neighboring user in the neighboring user set, and multi-dimensional environmental data, determine the similarity between the evaluation environment of the target user when evaluating each object to be pushed and the evaluation environment of each neighboring user in the neighboring user set when evaluating each object to be pushed.

[0242] In one implementation, the similarity of the evaluation environment between the target user and each neighboring user in the neighboring user set can be determined by the following steps:

[0243] The behavioral data of the target user is correlated with the multi-dimensional environmental data to obtain the contextual environment dataset when the target user evaluates each object to be pushed, and to determine the weight value of each contextual environment in the contextual environment dataset when the target user evaluates each object to be pushed.

[0244] The behavioral data of each neighboring user in the neighboring user set is correlated with the multi-dimensional environmental data to obtain the contextual environment dataset when each neighboring user in the neighboring user set evaluates each object to be pushed, and to determine the weight value of each contextual environment data in the contextual environment dataset when each neighboring user in the neighboring user set evaluates each object to be pushed.

[0245] Based on the contextual environment dataset of the target user evaluating each target object, the weight values ​​of each contextual environment in the contextual environment dataset of the target user evaluating each target object, the contextual environment dataset of each neighboring user in the neighboring user set evaluating each target object, and the weight values ​​of each contextual environment data in the contextual environment dataset of each neighboring user in the neighboring user set evaluating each target object, the similarity between the evaluation environment of the target user evaluating the current target object and the evaluation environment of the current neighboring user in the neighboring user set evaluating the current target object is determined.

[0246] S4: Based on the target user's behavior data, the behavior data of each reference user in the reference user set, the similarity between the target user and each neighboring user in the neighboring user set, and the similarity between the evaluation environment of the target user when evaluating each target object and the evaluation environment of each neighboring user in the neighboring user set when evaluating each target object, determine the first predicted score of the target user for each target object in the target object set.

[0247] In one implementation, based on the target user's behavioral data, the supply and demand data of each target object in the target object set, and multi-dimensional environmental data, a second predicted score for each target object in the target object set is determined, including:

[0248] S1: Based on the target user's behavior data, select one or more objects that have been consumed in the target user's behavior data as a set of historical consumption objects.

[0249] S2: Based on the supply and demand data of each push object in the historical consumer object set, the push object set, and multi-dimensional environmental data, determine the target user's preference for each historical consumer object in the historical consumer object set.

[0250] In one implementation, determining the target user's preference for each historical consumer in the historical consumer set is based on the historical consumer set, the supply and demand data of each target in the target set, and multi-dimensional environmental data, including:

[0251] The supply and demand data of each object to be pushed in the set of objects to be pushed are associated with the multi-dimensional environmental data to obtain the context environment data associated with each object to be pushed, which serves as the context environment dataset corresponding to each object to be pushed.

[0252] Determine the association frequency between each context environment data in the context environment dataset corresponding to each object to be pushed and each historical consumer object in the historical consumer object set;

[0253] Based on the association frequency between each context environment data in the context environment dataset corresponding to each object to be pushed and each historical consumer object, determine the second weight value of each context environment data in the context environment dataset corresponding to each object to be pushed.

[0254] Based on the number of context environment data in the context environment dataset corresponding to each target object and the second weight value of each context environment data, the target user's preference for each historical consumer object in the historical consumer object set is determined under each context environment data in the context environment dataset corresponding to each target object.

[0255] S3: Based on the supply and demand data of each object in the object set to be pushed and multi-dimensional environmental data, determine the target user's preference for each object in the object set to be pushed.

[0256] In one implementation, determining the target user's preference for each object in the push-to-object set based on supply and demand data and multi-dimensional environmental data includes:

[0257] The supply and demand data of each object to be pushed in the set of objects to be pushed are associated with the multi-dimensional environmental data to obtain the context environment data associated with each object to be pushed, which serves as the context environment dataset corresponding to each object to be pushed.

[0258] Determine the amount of context environment data in the context environment dataset corresponding to each object to be pushed;

[0259] Determine the association frequency between each context environment data and each push object in the context environment dataset corresponding to each push object;

[0260] Based on the association frequency between each context environment data in the context environment dataset corresponding to each push object and each push object, determine the first weight value of each context environment data in the context environment dataset corresponding to each push object.

[0261] Based on the number of context environment data in the context environment dataset corresponding to each target object and the first weight value of each context environment data, the target user's preference for each target object in the target object set is determined under each context environment data in the context environment dataset corresponding to each target object.

[0262] S4: Based on the target user's preference for each historical consumer in the historical consumer set and each target object in the target object set, determine the similarity between each target object in the target object set and each historical consumer in the historical consumer set.

[0263] S5: Based on the target user's behavioral data and the similarity between each push object in the target object set and each historical consumption object in the historical consumption object set, determine the target user's second predicted score for each push object in the target object set.

[0264] In one implementation, the second predicted score of each target user in the target user set can be determined as follows:

[0265] Determine the membership degree of the target user's evaluation value to each target object in the target user's behavioral data;

[0266] Based on the similarity between each push object in the target object set and each historical consumption object in the historical consumption object set, and the membership degree of the target user's evaluation value of each push object in the target user's behavior data, the second predicted rating of the target user for each push object in the target object set is determined.

[0267] In one implementation, the predicted score of each target user in the target user set can be determined as follows:

[0268] Determine the rating adjustment factors for the first and second predicted scores;

[0269] Based on the rating adjustment factor, the first predicted rating of each target object in the target user's target object set, and the second predicted rating of each target object in the target user's target object set, the target predicted rating of each target object in the target user's target object set is determined.

[0270] This application also provides a specific implementation of an electronic device capable of implementing all steps of the target object push method in the above embodiments. The electronic device specifically includes: a processor, a memory, a communication interface, and a bus; wherein the processor, memory, and communication interface communicate with each other via the bus; the processor is used to call a computer program in the memory, and when the processor executes the computer program, it implements all steps of the target object push method in the above embodiments. For example, when the processor executes the computer program, it implements the following steps:

[0271] Step 1: Obtain behavioral data of the target user, behavioral data of each reference user in the reference user set, supply and demand data of each target object in the target object set, and multi-dimensional environmental data;

[0272] Step 2: Based on the target user's behavioral data, the behavioral data of each reference user in the reference user set, the supply and demand data of each target object in the target object set, and multi-dimensional environmental data, determine the target user's target prediction score for each target object in the target object set.

[0273] Step 3: Based on the target user's target prediction score for each target object in the target object set, determine at least one object to be pushed to the target user;

[0274] Step 4: Push the at least one object to the target user.

[0275] Embodiments of this application also provide a computer-readable storage medium capable of implementing all steps of the target object push method in the above embodiments. The computer-readable storage medium stores a computer program that, when executed by a processor, implements all steps of the target object push method in the above embodiments. For example, when the processor executes the computer program, it implements the following steps:

[0276] Step 1: Obtain behavioral data of the target user, behavioral data of each reference user in the reference user set, supply and demand data of each target object in the target object set, and multi-dimensional environmental data;

[0277] Step 2: Based on the target user's behavioral data, the behavioral data of each reference user in the reference user set, the supply and demand data of each target object in the target object set, and multi-dimensional environmental data, determine the target user's target prediction score for each target object in the target object set.

[0278] Step 3: Based on the target user's target prediction score for each target object in the target object set, determine at least one object to be pushed to the target user;

[0279] Step 4: Push the at least one object to the target user.

[0280] As described above, this application embodiment obtains the target user's behavioral data, the behavioral data of each reference user in the reference user set, the supply and demand data of each target object in the target object set, and multi-dimensional environmental data; based on the target user's behavioral data, the behavioral data of each reference user in the reference user set, the supply and demand data of each target object in the target object set, and the multi-dimensional environmental data, it determines the target user's predicted score for each target object in the target object set; based on the target user's predicted score for each target object in the target object set, it determines at least one object to be pushed to the target user; and it pushes the at least one object to the target user. This scheme can predict the scores of each object, accurately determine the target object to be pushed to the target user, thereby improving user experience and operational efficiency.

[0281] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on its differences from other embodiments. In particular, hardware + program embodiments are relatively simple in description because they are fundamentally similar to method embodiments; relevant parts can be referred to the descriptions in the method embodiments.

[0282] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0283] While this application provides the method operation steps as described in the embodiments or flowcharts, more or fewer operation steps may be included based on conventional or non-inventive labor. The order of steps listed in the embodiments is merely one possible execution order among many and does not represent the only execution order. In actual device or client product execution, the methods shown in the embodiments or drawings can be executed sequentially or in parallel (e.g., in a parallel processor or multi-threaded processing environment).

[0284] While this specification provides method operation steps as described in the embodiments or flowcharts, more or fewer operation steps may be included based on conventional or non-inventive means. The order of steps listed in the embodiments is merely one possible execution order among many and does not represent the only execution order. In actual device or end product execution, the methods shown in the embodiments or drawings may be executed sequentially or in parallel (e.g., in a parallel processor or multi-threaded processing environment, or even a distributed data processing environment). The terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, product, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, product, or apparatus. Without further limitations, the presence of other identical or equivalent elements in the process, method, product, or apparatus that includes said elements is not excluded.

[0285] For ease of description, the above devices are described in terms of function, divided into various modules. Of course, in implementing the embodiments of this specification, the functions of each module can be implemented in one or more software and / or hardware components, or a module that performs the same function can be implemented by a combination of multiple sub-modules or sub-units. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, or indirect coupling or communication connection between devices or units, and may be electrical, mechanical, or other forms.

[0286] Those skilled in the art will also know that, besides implementing the controller using purely computer-readable program code, the same functions can be achieved by logically programming the method steps, making the controller function as logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers (PLCs), and embedded microcontrollers. Therefore, such a controller can be considered a hardware component, and the devices within it used to implement various functions can also be considered structures within that hardware component. Alternatively, the devices used to implement various functions can be considered as both software modules implementing the method and structures within a hardware component.

[0287] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0288] Those skilled in the art will understand that the embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, the embodiments of this specification can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, the embodiments of this specification can take the form of computer program products implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0289] The embodiments described in this specification can be described in the general context of computer-executable instructions, such as program modules, that are executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. The embodiments of this specification can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0290] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, system embodiments are basically similar to method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. In the description of this specification, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the embodiments in this specification. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described can be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0291] The above description is merely an embodiment of the present specification and is not intended to limit the embodiments of the present specification. For those skilled in the art, various modifications and variations can be made to the embodiments of the present specification. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the embodiments of the present specification should be included within the scope of the claims of the embodiments of the present specification.

Claims

1. A method for pushing content to a target object, characterized in that, include: Acquire behavioral data of target users, behavioral data of each reference user in the reference user set, supply and demand data of each target object in the target object set, and multi-dimensional environmental data; Based on the target user's behavioral data, the behavioral data of each reference user in the reference user set, the supply and demand data of each target object in the target object set, and multi-dimensional environmental data, the target user's target prediction score for each target object in the target object set is determined. Based on the target user's target prediction score for each target object in the target object set, determine at least one object to be pushed to the target user; Push the at least one object to the target user.

2. The method according to claim 1, characterized in that, Based on the target user's behavioral data, the behavioral data of each reference user in the reference user set, the supply and demand data of each target object in the target object set, and multi-dimensional environmental data, the target user's target prediction score for each target object in the target object set is determined, including: Based on the target user's behavioral data, the behavioral data of each reference user in the reference user set, and multi-dimensional environmental data, determine the first predicted score of each target user in the target user set to be pushed to; Based on the target user's behavioral data, the supply and demand data of each target object in the target object set, and multi-dimensional environmental data, the second predicted score of the target user for each target object in the target object set is determined. Based on the first predicted score of each target object in the target object set and the second predicted score of each target object in the target object set, the target predicted score of each target object in the target object set is determined.

3. The method according to claim 2, characterized in that, Based on the target user's behavioral data, the behavioral data of each reference user in the reference user set, and multi-dimensional environmental data, a first predicted score for each target user in the target user set is determined, including: Based on the behavioral data of the target user and the behavioral data of each reference user in the reference user set, the similarity between the target user and each reference user in the reference user set is determined. Based on the similarity between the target user and each reference user in the reference user set, one or more users in the reference user set whose similarity to the target user is higher than a preset threshold are selected as the neighborhood user set. Based on the target user's behavioral data, the behavioral data of each neighboring user in the neighboring user set, and multi-dimensional environmental data, determine the similarity between the evaluation environment of the target user when evaluating each target object and the evaluation environment of each neighboring user in the neighboring user set when evaluating each target object. Based on the target user's behavioral data, the behavioral data of each reference user in the reference user set, the similarity between the target user and each neighboring user in the neighboring user set, and the similarity between the evaluation environment of the target user when evaluating each target object and the evaluation environment of each neighboring user in the neighboring user set when evaluating each target object, the first predicted score of the target user for each target object in the target object set is determined.

4. The method according to claim 3, characterized in that, Based on the behavioral data of the target user and the behavioral data of each reference user in the reference user set, the similarity between the target user and each reference user in the reference user set is determined, including: Based on the target user's behavioral data, determine the target user's evaluation value for each target object to be pushed to, and determine the average value of each target object to be pushed to in the set of target objects to be pushed to; Based on the behavioral data of each reference user in the reference user set, determine the evaluation value of each reference user in the reference user set for each object to be pushed, and determine the average value of the evaluation values ​​of each object to be pushed in the set formed by each reference user in the reference user set for the objects to be pushed. The similarity between the target user and each reference user in the reference user set is determined based on the target user's rating of each target object, the average rating of each target object in the set formed by the target user, the rating of each target object by each reference user in the reference user set, and the average rating of each target object in the set formed by each reference user in the reference user set.

5. The method according to claim 3, characterized in that, Based on the target user's behavioral data, the behavioral data of each neighboring user in the neighboring user set, and multi-dimensional environmental data, determine the similarity between the evaluation environment of the target user when evaluating each target object and the evaluation environment of each neighboring user in the neighboring user set when evaluating each target object, including: The behavioral data of the target user is correlated with the multi-dimensional environmental data to obtain the contextual environment dataset when the target user evaluates each object to be pushed, and to determine the weight value of each contextual environment in the contextual environment dataset when the target user evaluates each object to be pushed. The behavioral data of each neighboring user in the neighboring user set is correlated with the multi-dimensional environmental data to obtain the contextual environment dataset when each neighboring user in the neighboring user set evaluates each object to be pushed, and to determine the weight value of each contextual environment data in the contextual environment dataset when each neighboring user in the neighboring user set evaluates each object to be pushed. Based on the contextual environment dataset when the target user evaluates each object to be pushed, the weight value of each contextual environment in the contextual environment dataset when the target user evaluates each object to be pushed, the contextual environment dataset when each neighboring user in the neighboring user set evaluates each object to be pushed, and the weight value of each contextual environment data in the contextual environment dataset when each neighboring user in the neighboring user set evaluates each object to be pushed, the similarity between the evaluation environment when the target user evaluates the current object to be pushed and the evaluation environment when the current neighboring user in the neighboring user set evaluates the current object to be pushed is determined.

6. The method according to claim 2, characterized in that, Based on the target user's behavioral data, the supply and demand data of each target object in the target object set, and multi-dimensional environmental data, a second predicted score for the target user is determined for each target object in the target object set, including: Based on the target user's behavior data, one or more objects that have been consumed in the target user's behavior data are taken as the historical consumption object set; Based on the supply and demand data of each push object in the historical consumer object set, the target user's preference for each historical consumer object in the historical consumer object set, and multi-dimensional environmental data, the user's preference for each historical consumer object in the historical consumer object set is determined. Based on the supply and demand data of each target object in the target object set and multi-dimensional environmental data, the target user's preference for each target object in the target object set is determined. Based on the target user's preference for each historical consumer in the historical consumer set and each target consumer in the target set, the similarity between each target consumer in the target set and each historical consumer in the historical consumer set is determined. Based on the target user's behavioral data and the similarity between each push object in the target object set and each historical consumption object in the historical consumption object set, a second predicted score for each push object in the target user's target object set is determined.

7. The method according to claim 6, characterized in that, Based on the supply and demand data and multi-dimensional environmental data of each target object in the target object set, the preference of the target user for each target object in the target object set is determined, including: The supply and demand data of each object to be pushed in the set of objects to be pushed are associated with the multi-dimensional environmental data to obtain the context environment data associated with each object to be pushed, which serves as the context environment dataset corresponding to each object to be pushed. Determine the amount of context environment data in the context environment dataset corresponding to each object to be pushed; Determine the association frequency between each context environment data and each push object in the context environment dataset corresponding to each push object; Based on the association frequency between each context environment data in the context environment dataset corresponding to each push object and each push object, determine the first weight value of each context environment data in the context environment dataset corresponding to each push object. Based on the number of context environment data in the context environment dataset corresponding to each target object and the first weight value of each context environment data, the target user's preference for each target object in the target object set is determined under each context environment data in the context environment dataset corresponding to each target object.

8. The method according to claim 6, characterized in that, Based on the historical consumer object set, the supply and demand data of each target object in the target object set, and multi-dimensional environmental data, determine the target user's preference for each historical consumer object in the historical consumer object set, including: The supply and demand data of each object to be pushed in the set of objects to be pushed are associated with the multi-dimensional environmental data to obtain the context environment data associated with each object to be pushed, which serves as the context environment dataset corresponding to each object to be pushed. Determine the association frequency between each context environment data in the context environment dataset corresponding to each object to be pushed and each historical consumer object in the historical consumer object set; Based on the association frequency between each context environment data in the context environment dataset corresponding to each object to be pushed and each historical consumer object, determine the second weight value of each context environment data in the context environment dataset corresponding to each object to be pushed. Based on the number of context environment data in the context environment dataset corresponding to each target object and the second weight value of each context environment data, the target user's preference for each historical consumer object in the historical consumer object set is determined under each context environment data in the context environment dataset corresponding to each target object.

9. The method according to claim 6, characterized in that, Based on the target user's behavioral data and the similarity between each push object in the target object set and each historical consumer object in the historical consumer object set, a second predicted score for each push object in the target user's target object set is determined, including: Determine the membership degree of the target user's evaluation value to each target object in the target user's behavioral data; Based on the similarity between each push object in the target object set and each historical consumption object in the historical consumption object set, and the membership degree of the target user's evaluation value of each push object in the target user's behavior data, the second predicted rating of the target user for each push object in the target object set is determined.

10. The method according to claim 3, characterized in that, Based on the first predicted score and the second predicted score of each target user in the target user set, the predicted score of each target user in the target user set is determined, including: Determine the rating adjustment factors for the first and second predicted scores; Based on the rating adjustment factor, the first predicted rating of each target object in the target user's target object set, and the second predicted rating of each target object in the target user's target object set, the target predicted rating of each target object in the target user's target object set is determined.

11. A device for pushing a target object, characterized in that, include: The acquisition module is used to acquire behavioral data of target users, behavioral data of each reference user in the reference user set, supply and demand data of each object to be pushed in the object to be pushed set, and multi-dimensional environmental data. The calculation module is used to determine the target prediction score of the target user for each object in the target object set based on the target user's behavior data, the behavior data of each reference user in the reference user set, the supply and demand data of each object to be pushed in the target object set, and multi-dimensional environmental data. The determination module is used to determine at least one object to be pushed to the target user based on the target prediction score of each object in the target user's set of objects to be pushed; A push module is used to push the at least one object to the target user.

12. An electronic device comprising a processor and a memory for storing processor-executable instructions, characterized in that, When the processor executes the instructions, it implements the steps of the method according to any one of claims 1 to 10.

13. A computer-readable storage medium storing computer instructions thereon, characterized in that, When the instructions are executed by the processor, they implement the steps of the method according to any one of claims 1 to 10.

Images