CN113822734A

CN113822734A - Method and apparatus for generating information

Info

Publication number: CN113822734A
Application number: CN202110060339.0A
Authority: CN
Inventors: 张秀军
Original assignee: Beijing Jingdong Century Trading Co Ltd; Beijing Wodong Tianjun Information Technology Co Ltd
Current assignee: Beijing Jingdong Century Trading Co Ltd; Beijing Wodong Tianjun Information Technology Co Ltd
Priority date: 2021-01-18
Filing date: 2021-01-18
Publication date: 2021-12-21

Abstract

The embodiment of the disclosure discloses a method and a device for generating information. One embodiment of the method comprises: acquiring initial information to be pushed, wherein the initial information to be pushed comprises commodity information of initial pushed commodities which are arranged in sequence; determining at least one target sorting position from the initial information to be pushed based on the corresponding relation between the predetermined commodity information sorting positions and the user loss quantity; and inserting commodity information of a predetermined target pushed commodity into each target sequencing position to obtain target information to be pushed. The commodity information of the target push commodity with the exposure requirement can be inserted into the information to be pushed, and the adverse effect of the commodity information which is not interested by the user and is included in the generated target push information on the interest degree of the user is reduced.

Description

Method and apparatus for generating information

Technical Field

The embodiment of the disclosure relates to the technical field of computers, in particular to a method and a device for generating information.

Background

The e-commerce platform generally generates push information according to behavior data of a user (e.g., purchasing habits of the user), and then pushes information of commodities that the user may be interested in to the user, for example, the information to be pushed may be generated by using a CVR (Conversion Rate) or CTR (Click Through Rate) predictive model. In some practical scenarios, a business party may put forward an exposure appeal to an e-commerce platform, and designate to push commodity information of some commodities to a user, so as to improve the exposure degree of the commodities.

In the related art, a certain weight is usually given to the information of the commodities to be exposed, so as to promote the ranking of the part of commodities in the push information and further increase the exposure chance of the part of commodities.

Disclosure of Invention

Embodiments of the present disclosure propose methods and apparatuses for generating information.

In a first aspect, an embodiment of the present disclosure provides a method for generating information, the method including: acquiring initial information to be pushed, wherein the initial information to be pushed comprises commodity information of initial pushed commodities which are arranged in sequence; determining at least one target sorting position from the initial information to be pushed based on the corresponding relation between the predetermined commodity information sorting positions and the user loss quantity; and inserting commodity information of a predetermined target pushed commodity into each target sequencing position to obtain target information to be pushed.

In some embodiments, the target information to be pushed is generated via the following steps: receiving push demand information, wherein the push demand information comprises target push commodities and target push users; the method comprises the steps of obtaining commodity characteristic information of a target pushed commodity, historical sales information of the target pushed commodity within a first preset time length and user characteristic information of a target pushed user, wherein the user characteristic information comprises user behavior characteristics and/or user portrait; determining an initial sequencing characteristic value of the target pushed commodity based on the commodity characteristic information, the historical sales information and the user characteristic information; arranging the commodity information of the target pushed commodities according to the descending order of the initial sequencing characteristic values of the target pushed commodities; and respectively inserting the commodity information of each target pushed commodity into each target sequencing position in sequence according to the sequence of each target sequencing position from front to back to obtain target information to be pushed.

In some embodiments, the historical sales information comprises: the order quantity, the order quantity and the total volume of the transaction; and, the initial ranking characteristic value is determined via the steps of: coding the commodity feature information to generate a commodity feature vector; encoding the user characteristic information to generate a user characteristic vector; determining the distance between the commodity feature vector and the user feature vector as a user preference feature value of the target pushed commodity; respectively normalizing the order quantity, the order quantity and the total transaction quantity to obtain the normalized order quantity, the normalized order quantity and the normalized total transaction quantity; determining the weighted sum of the normalized order quantity, the normalized order quantity and the normalized transaction total quantity as a commodity quality characteristic value of the target pushed commodity; and determining the weighted sum of the user preference characteristic value and the commodity quality characteristic value as the initial sorting characteristic value of the target pushed commodity.

In a second aspect, an embodiment of the present disclosure provides a method for pushing information, where the method includes: pushing target information to be pushed, wherein the target information to be pushed is generated by the method for generating information; every interval of a preset recall cycle, the following off-line updating steps are executed on the target information to be pushed: the method comprises the steps of obtaining recall information of commodity information of each target pushed commodity in target information to be pushed in a last recall period, wherein the recall information comprises first preset type user behavior information; adopting a pre-trained recall model, determining a recall characteristic value of the target pushed commodity based on recall information, wherein the recall characteristic value represents the matching degree of the commodity information of the target pushed commodity and user behaviors; determining a recall ranking characteristic value of each target pushed commodity based on a recall characteristic value of the target pushed commodity and a predetermined grade weight value of the target pushed commodity; updating the ranking of the commodity information of each target pushed commodity in the target information to be pushed based on the recall ranking characteristic value of each target pushed commodity; the method further comprises the following steps: and pushing the updated information to be pushed of the target.

In some embodiments, the rank weight value is determined via: acquiring historical characteristic information and value attributes of the target pushed commodity within a second preset time length, wherein the historical characteristic information comprises second preset type user behavior information; determining the weighted sum of various types of user behavior information in the historical characteristic information as a user behavior characteristic value of the target pushed commodity; determining the grade of the target pushed commodity based on a first difference value and a second difference value of the target pushed commodity, wherein the first difference value is the difference between the user behavior characteristic value of the target pushed commodity and the mean value of the user behavior characteristic values of all the target pushed commodities, and the second difference value is the difference between the value attribute of the target pushed commodity and the mean value of the value attribute of all the target pushed commodities; and determining the grade weight value of the target pushed commodity based on the corresponding relation between the preset grade and the grade weight value.

In some embodiments, determining the rank of the targeted pushed good based on the first difference and the second difference of the targeted pushed good comprises: establishing a coordinate system by taking the mean value of the user behavior characteristic values and the mean value of the value attributes of the target pushed commodities as origin coordinates and the user behavior characteristic values and the value attributes of the target pushed commodities as coordinate axes; determining the position of each target pushed commodity in the coordinate system based on the first difference value and the second difference value of each target pushed commodity; and determining the grade of each target pushed commodity based on the corresponding relation between the preset coordinate system quadrant and the grade.

In some embodiments, the method further comprises the following step of updating the target information to be pushed in real time: in response to the fact that the exposure times of the commodity information of the target pushed commodity reach a preset exposure time threshold, taking the product of the initial sorting characteristic value of the target to-be-pushed commodity and a preset exposure weight value as an updated sorting characteristic value, wherein the exposure weight value is negatively related to the exposure times; and updating the ordering of the commodity information of each target pushed commodity in the target information to be pushed based on the updated ordering characteristic value of the commodity information of each target pushed commodity.

In a third aspect, an embodiment of the present disclosure provides an apparatus for generating information, the apparatus including: the system comprises an initial information acquisition unit, a display unit and a display unit, wherein the initial information acquisition unit is configured to acquire initial information to be pushed, and the initial information to be pushed comprises commodity information of initial pushed commodities which are arranged in sequence; the sorting position determining unit is configured to determine at least one target sorting position from the initial information to be pushed based on the corresponding relation between the predetermined commodity information sorting positions and the user loss quantity; and the target information generating unit is configured to insert commodity information of a predetermined target pushed commodity into each target sequencing position to obtain target information to be pushed.

In some embodiments, the target information generating unit includes: a demand information receiving module configured to receive push demand information, the push demand information including a target push commodity and a target push user; the characteristic information acquisition module is configured to acquire commodity characteristic information of the target pushed commodity, historical sales information of the target pushed commodity within a first preset time length and user characteristic information of a target pushing user, wherein the user characteristic information comprises user behavior characteristics and/or a user portrait; the initial sorting characteristic value determining module is configured to determine an initial sorting characteristic value of the target pushed commodity based on the commodity characteristic information, the historical sales amount information and the user characteristic information; the information sorting module is configured to sort the commodity information of the target pushed commodities according to the descending order of the initial sorting characteristic values of the target pushed commodities; and the information insertion module is configured to insert the commodity information of each target pushed commodity into each target sorting position in sequence according to the sequence of the target sorting positions from front to back to obtain the target information to be pushed.

In some embodiments, the historical sales information comprises: the order quantity, the order quantity and the total volume of the transaction; and the initial ranking feature value determination module is further configured to: coding the commodity feature information to generate a commodity feature vector; encoding the user characteristic information to generate a user characteristic vector; determining the distance between the commodity feature vector and the user feature vector as a user preference feature value of the target pushed commodity; respectively normalizing the order quantity, the order quantity and the total transaction quantity to obtain the normalized order quantity, the normalized order quantity and the normalized total transaction quantity; determining the weighted sum of the normalized order quantity, the normalized order quantity and the normalized transaction total quantity as a commodity quality characteristic value of the target pushed commodity; and determining the weighted sum of the user preference characteristic value and the commodity quality characteristic value as the initial sorting characteristic value of the target pushed commodity.

In a fourth aspect, an embodiment of the present disclosure further provides an apparatus for pushing information, where the apparatus includes: the information pushing unit is configured to push target information to be pushed, and the target information to be pushed is generated by the information generating method; the information updating unit is configured to execute the following offline updating steps on the target information to be pushed at intervals of a preset recall cycle: the method comprises the steps of obtaining recall information of each target pushed commodity in target information to be pushed in a last recall period, wherein the recall information comprises first preset type user behavior information; adopting a pre-trained recall model, determining a recall characteristic value of a target pushed commodity based on recall information of the target pushed commodity, wherein the recall characteristic value represents the matching degree of commodity information of the target pushed commodity and user behaviors; determining a recall ranking characteristic value of each target pushed commodity based on a recall characteristic value of the target pushed commodity and a predetermined grade weight value of the target pushed commodity; updating the ranking of each target pushed commodity in the target information to be pushed based on the recall ranking characteristic value of each target pushed commodity; the information pushing unit is further configured to: and pushing the updated information to be pushed of the target.

In some embodiments, the apparatus further comprises: the historical information acquisition unit is configured to acquire historical characteristic information and value attributes of each target pushed commodity within a second preset time length, wherein the historical characteristic information comprises second preset types of user behavior information; the behavior characteristic determining unit is configured to determine the weighted sum of various types of user behavior information in the historical characteristic information as a user behavior characteristic value of the target pushed commodity; a grade determining unit configured to determine a grade of the target pushed commodity based on a first difference value and a second difference value of the target pushed commodity, wherein the first difference value is a difference between a user behavior feature value of the target pushed commodity and a mean value of the user behavior feature values of the target pushed commodities, and the second difference value is a difference between a value attribute of the target pushed commodity and a mean value of the value attributes of the target pushed commodities; and the grade weight determining unit is configured to determine a grade weight value of the target pushed commodity based on a corresponding relation between a preset grade and the grade weight value.

In some embodiments, the rank determination unit further comprises: the coordinate system building module is configured to build a coordinate system by taking the mean value of the user behavior characteristic values and the mean value of the value attributes of the target pushed commodities as origin coordinates and taking the user behavior characteristic values and the value attributes of the target pushed commodities as coordinate axes; a position determination module configured to determine a position of each target pushed commodity in the coordinate system based on the first difference value and the second difference value of each target pushed commodity; and the grade determining module is configured to determine the grade of each target pushed commodity based on the corresponding relation between the preset coordinate system quadrant and the grade.

In some embodiments, the apparatus further comprises a real-time update unit configured to: in response to the fact that the exposure times of the commodity information of the target pushed commodity reach a preset exposure time threshold, taking the product of the initial sorting characteristic value of the target to-be-pushed commodity and a preset exposure weight value as an updated sorting characteristic value, wherein the exposure weight value is negatively related to the exposure times; and updating the ordering of the commodity information of each target pushed commodity in the target information to be pushed based on the updated ordering characteristic value of the commodity information of each target pushed commodity.

According to the method and the device for generating information, the target sorting position is determined from the initial information to be pushed based on the corresponding relation between the commodity information sorting position and the user loss quantity, the commodity information of the target pushed commodity is inserted into the target sorting position, the commodity information of the target pushed commodity with exposure requirements can be inserted into the initial information to be pushed, and the adverse effect of the commodity information which is not interested by the user and is included in the generated target pushed information on the user interest degree is reduced.

According to the method and the device for pushing the information, the target information to be pushed is updated in an off-line mode every other preset recall period, so that the sequence of the commodity information of the target pushed commodity in the target information to be pushed is updated according to the matching degree of the commodity information of each target pushed commodity in the last recall period and the user behavior and the preset grade weight, and the pertinence of the information pushed to the user can be improved.

Drawings

Other features, objects and advantages of the disclosure will become more apparent upon reading of the following detailed description of non-limiting embodiments thereof, made with reference to the accompanying drawings in which:

FIG. 1 is an exemplary system architecture diagram in which some embodiments of the present disclosure may be applied;

FIG. 2 is a flow diagram for one embodiment of a method for generating information, according to the present disclosure;

FIG. 3 is a flow diagram of yet another embodiment of a method for generating information according to the present disclosure;

FIG. 4 is a flow diagram for one embodiment of a method for pushing information, according to the present disclosure;

FIG. 5 is a flow diagram of yet another embodiment of a method for pushing information according to the present disclosure;

FIG. 6 is a schematic block diagram illustrating one embodiment of an apparatus for generating information according to the present disclosure;

FIG. 7 is a schematic block diagram illustrating one embodiment of an apparatus for pushing information according to the present disclosure;

FIG. 8 is a schematic structural diagram of an electronic device suitable for use in implementing embodiments of the present disclosure.

Detailed Description

The present disclosure is described in further detail below with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the relevant invention and not restrictive of the invention. It should be noted that, for convenience of description, only the portions related to the related invention are shown in the drawings.

It should be noted that, in the present disclosure, the embodiments and features of the embodiments may be combined with each other without conflict. The present disclosure will be described in detail below with reference to the accompanying drawings in conjunction with embodiments.

Fig. 1 illustrates an exemplary system architecture 100 of a method for pushing information or an apparatus for pushing information to which embodiments of the present disclosure may be applied.

As shown in fig. 1, the system architecture 100 may include

terminal devices

101, 102, 103, a network 104, and a server 105. The network 104 serves as a medium for providing communication links between the

terminal devices

101, 102, 103 and the server 105. Network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, to name a few.

The

terminal devices

101, 102, 103 interact with the server 105 through the network 104 to receive or send messages and the like, for example, commodity information of a target pushed commodity with an exposure demand may be sent to the server, and target to-be-pushed information may be received from the server.

The

terminal apparatuses

101, 102, and 103 may be hardware or software. When the

terminal devices

101, 102, and 103 are hardware, they may be electronic devices with communication functions, including but not limited to smart phones, tablet computers, e-book readers, laptop portable computers, desktop computers, and the like. When the

terminal apparatuses

101, 102, 103 are software, they can be installed in the electronic apparatuses listed above. It may be implemented, for example, as multiple software or software modules to provide distributed services, or as a single software or software module. And is not particularly limited herein.

The server 105 may be a server providing various services, such as a background data server that processes the commodity information of the target pushed commodity uploaded by the

terminal devices

101, 102, 103 (for example, inserts the commodity information of the target pushed commodity into the initial information to be pushed). The background data server can determine a target sorting position from the initial information to be pushed, insert the commodity information of the target pushed commodity into the target sorting position, and feed back a processing result (for example, the target information to be pushed) to the terminal device.

It should be noted that the method for pushing information provided by the embodiment of the present disclosure may be executed by the

terminal devices

101, 102, and 103, or may be executed by the server 105. Accordingly, the means for pushing information may be provided in the

terminal devices

101, 102, 103, or in the server 105. And is not particularly limited herein.

The server may be hardware or software. When the server is hardware, it may be implemented as a distributed server cluster formed by multiple servers, or may be implemented as a single server. When the server is software, it may be implemented as multiple pieces of software or software modules, for example, to provide distributed services, or as a single piece of software or software module. And is not particularly limited herein.

It should be noted that the method for pushing information or the apparatus for pushing information of the present disclosure is also applicable to the exemplary system architecture 100 shown in fig. 1.

With continued reference to FIG. 2, a flow 200 of one embodiment of a method for generating information in accordance with the present disclosure is shown. The method for generating information comprises the following steps:

step 201, initial information to be pushed is obtained, and the initial information to be pushed includes commodity information of initial pushed commodities arranged in sequence.

In this embodiment, the ranking position of the commodity information of each initial pushed commodity in the initial information to be pushed represents the matching degree between the commodity information of the commodity and the behavior habits of the user, for example, the commodity information of the commodity which is most matched with the behavior habits of the user is ranked at the top of the initial information to be pushed. The execution subject (for example, the server 105 shown in fig. 1) may obtain the initial information to be pushed on the terminal device through the network, or the execution subject may generate the initial information to be pushed according to the behavior data of the user.

As an example, the executing body may obtain user behavior data of a full number of SKUs (Stock Keeping Unit, basic Unit for inventory in and out metering), for example, the user behavior data may include user order placing times, Click times, purchase quantity, and the like, then estimate a matching degree between the commodity information of each SKU and the user behavior by using a CVR (Conversion Rate) or CTR (Click Through Rate) estimation model, and then arrange the commodity information of each SKU in order of matching degree from high to low to obtain initial information to be pushed. The merchandise information may include information describing the characteristics of the SKU, such as the color, model, price, etc. of the SKU.

Step 202, determining at least one target sorting position from the initial information to be pushed based on the corresponding relation between the predetermined commodity information sorting positions and the user loss amount.

Generally, when a user browses pushed commodity information, the browsing interest of the user generally declines with the increase of the quantity of the browsed commodity information, and the corresponding relation between the sequencing position of the commodity information and the loss quantity of the user can be analyzed by adopting a statistical analysis method. As an example, taking 1000 users as a statistical sample, the number of users browsing a first item of merchandise information is 1000, the number of users browsing a second item of merchandise information is 800, and the number of users browsing a third item of merchandise information is 500, it may be determined that the number of users losing the second item of merchandise information is 200, and the number of users losing the third item of merchandise information is 300.

In this embodiment, the correspondence between the commodity information ranking position and the user loss amount represents the correspondence between the commodity information ranking position and the interest level of the user in browsing the commodity information, and a larger user loss amount indicates a lower browsing interest of the user. The target sorting position may be, for example, a sorting position where the number of lost commodities of the user is large, and inserting commodity information different from other initially pushed commodities into the target sorting position may increase browsing interest of the user, thereby increasing exposure probability of the commodity information.

In a specific example, the execution main body may determine that the user loss amount of the commodity information ranked at the 4 th position and the 7 th position is greater according to the corresponding relationship between the commodity information ranking position and the user loss amount, and then the execution main body may determine the 4 th position and the 7 th position in the initial information to be pushed acquired in step 201 as the target ranking position.

Step 203, inserting commodity information of a predetermined target pushed commodity into each target sequencing position to obtain target information to be pushed.

In this embodiment, the target pushed commodity represents a commodity with a forcible exposure demand, for example, commodity information for which a pushing demand is proposed by a service user. The execution main body may insert the predetermined commodity information of the target pushed commodity into the target sorting position determined in step 202 to obtain the target information to be pushed, so that the target information to be pushed includes both the commodity information of the initial pushed commodity determined according to the behavior habit of the user and the commodity information of the commodity with the forcible exposure requirement.

In a specific example, the target pushed commodities are basketball and football, and the initial information to be pushed acquired by the executing entity in step 201 includes commodity information of the following SKUs arranged in sequence: the mobile phone, the tablet, the skin care product, the wine, the stationery, the jacket, the shoes and the like, wherein the target sorting positions determined by the execution main body based on the corresponding relation between the commodity information sorting positions and the user loss quantity are 4 th and 7 th, the execution main body inserts the commodity information of the basketball and the football into the 4 th and 7 th positions of the initial information to be pushed respectively, and the obtained target information to be pushed comprises the commodity information of the following SKUs which are sequentially arranged: mobile phones, tablets, skin care products, basketballs, wine, stationery, football, coats, shoes and the like.

With continued reference to fig. 3, fig. 3 shows a flow 300 of generating target to-be-pushed information in an embodiment of the method for generating information of the present disclosure, the flow includes the following steps:

step 301, receiving push demand information, where the push demand information includes a target push commodity and a target push user.

In a specific scenario example, a Content Management System (CMS) may be loaded on a terminal device, and an operator may send push requirement information to a backend server (i.e., an execution main body of this embodiment) through a front-end operation page of the CMS. As an example, when the e-commerce platform needs to perform marketing support on wine commodities, an operator may select a target push commodity as a wine commodity within a marketing support range on a front-end operation page of the CMS, then may select a full number of users of the e-commerce platform as target push users, or may select member users of the e-commerce platform as target push users, and then the CMS sends the target push commodity and the target push user selected by the operator to the back-end server.

In some optional implementation manners of this embodiment, the push demand information may further include a push duration, where the push duration is used to represent a duration of pushing the commodity information of the target push commodity to the target push user. For example, the service user may put forward a push request from 10 months 1 to 10 months 5 days, and the operator may input a push duration from 10 months 1 to 10 months 5 days in the front-end operation page of the CMS. After receiving the push demand information, the execution main body pushes the target to-be-pushed information to the target push user in the period from 10 months 1 to 10 months 5.

Step 302, acquiring commodity characteristic information of a target pushed commodity, historical sales information of the target pushed commodity within a first preset time length and user characteristic information of a target pushed user, wherein the user characteristic information comprises user behavior characteristics and/or a user portrait.

In this embodiment, the commodity feature information may be an information set describing features of the commodity, and may include information such as a brand, a type, a place of origin, and the like of the commodity. The historical sales information includes: order quantity, order volume, and total volume of deals. The behavior characteristics of the user can include behavior record information of the user on the commodity, for example, the behavior record of the user clicking commodity information to enter a commodity detail page, the behavior record of a motor search result to enter the commodity detail page, the behavior record of adding the commodity into a shopping cart, ordering and the like; the user representation represents the user's own features, such as the user's gender, age, and location.

And 303, coding the commodity feature information to generate a commodity feature vector.

In this embodiment, the executing entity may arrange the commodity feature information obtained in step 302 in order based on a preset commodity coding policy to obtain a commodity feature vector of the target pushed commodity.

And step 304, encoding the user characteristic information to generate a user characteristic vector.

In this embodiment, the executing entity may arrange the user feature information obtained in step 302 in order based on a preset user feature encoding policy to obtain a user feature vector.

Step 305, determining the distance between the commodity feature vector and the user feature vector as a user preference feature value of the target pushed commodity.

In the implementation, the distance between the commodity feature vector and the user feature vector can represent the matching degree of the user features and the commodity features, and the user preference feature value can represent the preference degree of the user for the commodity.

Step 306, normalizing the order quantity, the order quantity and the total transaction quantity respectively to obtain the normalized order quantity, the normalized order quantity and the normalized total transaction quantity.

And 307, determining the weighted sum of the normalized order placing quantity, the normalized order quantity and the normalized transaction total quantity as a commodity quality characteristic value of the target pushed commodity.

In the present embodiment, the commodity quality characteristic value characterizes sales volume characteristics of the commodity.

And 308, determining the weighted sum of the user preference characteristic value and the commodity quality characteristic value as the initial sorting characteristic value of the target pushed commodity.

In this embodiment, the initial ranking characteristic value includes a user preference characteristic value and a commodity quality characteristic value, so that the user preference degree and the sales characteristic of the target pushed commodity can be characterized by the initial ranking characteristic value.

And 309, arranging the commodity information of the target pushed commodities according to the sequence of the initial sequencing characteristic values of the target pushed commodities from large to small.

In this embodiment, the executing agent ranks the item information of each target pushed item based on the initial ranking characteristic value of each target pushed item obtained in step 308, so that the user preference degree and sales characteristic of the target pushed item can be reflected in the ranking position of the target pushed item.

And 310, respectively inserting the commodity information of each target pushed commodity into each target sequencing position in sequence according to the sequence of each target sequencing position from front to back to obtain target information to be pushed.

In a specific example, the target pushed commodities are basketball and football, and the initial information to be pushed acquired by the execution main body includes commodity information of the following SKUs arranged in sequence: the execution main body determines target sequencing positions to be 4 th and 7 th positions based on the corresponding relation between the commodity information sequencing position and the user loss quantity. Then, the execution main body determines that the initial ranking characteristic value of the basketball is 80 and the initial ranking characteristic value of the football is 90 through step 310, then the execution main body inserts the commodity information of the basketball and the football into the 7 th and the 4 th of the initial information to be pushed respectively, and the obtained target information to be pushed comprises the commodity information of the SKUs which are arranged in sequence as follows: mobile phones, tablets, skin care products, football, wine, stationery, basketballs, coats, shoes and the like.

As can be seen from fig. 3, compared with the embodiment shown in fig. 2, the process for generating the target information to be pushed shown in fig. 3 highlights an initial sorting feature value for determining the target pushed commodities and a sorting of each target pushed commodity in the target information to be pushed based on the initial sorting feature value, and can determine a sorting position of the target pushed commodity in the target information to be pushed based on the user preference degree and the commodity sales volume feature of the target pushed commodity, so that the reduction speed of the user browsing interest can be reduced, which is helpful for improving the pertinence of the target information to be pushed, and further improving the exposure probability of the pushed commodity.

Referring next to fig. 4, a flow 400 of one embodiment of a method for pushing information is shown. The flow 400 of the method for pushing information comprises the following steps:

step 401, pushing target information to be pushed, where the target information to be pushed is generated by the above method for generating information.

In this embodiment, the executing subject may be the server 105 shown in fig. 1, and sends the target information to be pushed to the client (for example, the target information to be pushed is obtained by using the method for generating information shown in fig. 2 or fig. 3) through the network, where the target information to be pushed includes commodity information of an initial pushed commodity and commodity information of a target pushed commodity, where the target pushed commodity is a commodity with a forcible exposure requirement, for example, a commodity for which a business user proposes a pushing demand, and further, for example, a commodity supported by e-commerce platform marketing.

It is to be understood that the execution subject of the method for pushing information in this embodiment may also be the execution subject of the above method for generating information, and target information to be pushed is directly generated via the above method for generating information, or target information to be pushed generated by other electronic devices may also be received, which is not limited by this disclosure.

And 402, executing an offline updating step on the target information to be pushed every preset recall period.

In this embodiment, the recall cycle represents a cycle in which the execution main body performs offline update on the commodity information of each target pushed commodity in the target pushed information, and the execution main body may set an independent offline recall pool for the target with the commodity information of the target pushed commodity in the target information to be pushed, and perform offline recall on the commodity information of each target pushed commodity by using a recall algorithm.

In this embodiment, the offline updating step shown in step 402 specifically includes the following steps:

step 4021, retrieving information of the commodity information of each target pushed commodity in the target information to be pushed in the last retrieval period, wherein the retrieving information comprises user behavior information of a first preset type.

In this embodiment, the executing entity executes the offline recall algorithm on the target pushed commodity based on the recall information, for example, the first preset type of user behavior information may include behavior times of clicking and checking the target pushed commodity by the user, user order placing times, user collection times and other types of behavior times, and the executing entity may encode the first preset type of user behavior information of each target pushed commodity in the previous recall period based on a preset encoding policy to obtain the recall information of each target pushed commodity in the previous recall period, where the recall information may be in a vector form, for example.

Step 4022, adopting a pre-trained recall model, determining a recall characteristic value of the target pushed commodity based on the recall information, wherein the recall characteristic value represents the matching degree of the commodity information of the target pushed commodity and the user behavior.

In this embodiment, the recall model represents the correspondence between the recall information and the recall feature value. The recall model may be, for example, a DNN model (Deep Neural Networks), an RNN model (Recurrent Neural Networks), or a DSSM model (Deep Structured Semantic Models).

As an example, the executive body may adopt a Deep factor decomposition (Deep factor decomposition) model as a recall model, and the recall information of the commodity information of the target pushed commodity obtained in step 4021 in the last recall cycle is input into the Deep fm model trained in advance, so as to obtain a recall characteristic value of the target pushed commodity. The deep FM model effectively combines the advantages of a factorization machine and a neural network in feature learning, and can simultaneously extract low-order combined features and high-order combined features, so that the recall characteristic value can more accurately represent the matching degree of user behaviors and commodity information of target pushed commodities and the user behaviors.

Step 4023, determining a recall ranking characteristic value of each target pushed commodity based on the recall characteristic value of the target pushed commodity and a predetermined grade weight value of the target pushed commodity.

In this embodiment, each commodity may be graded in advance according to a preset grading strategy, and then a corresponding relationship between the grades and the grade weight values is established, so that the matching degree between the target push commodity and the preset grading strategy may be represented by the grade weight values.

As an example, each target pushed commodity can be divided into different grades according to the marketing support strength of the commodity, the higher the grade of the target pushed commodity with high marketing support strength is, and the larger the corresponding grade weight value is, the grade weight value can represent the marketing support strength of the target pushed commodity. For another example, each target pushed commodity can be divided into different grades according to the historical sales of the commodity, and considering the martensitic effect of the commodity in practical application, the sales of the commodity with high historical sales are higher and higher, and the sales of the commodity with low historical sales are lower and lower, so that a higher grade can be allocated to the target pushed commodity with low historical sales, and thus the target pushed commodity with low historical sales can obtain a higher grade weight value, and the grade weight value can represent the historical sales of the target pushed commodity.

In this embodiment, the product of the recall feature value and the rank weight value may be used as the recall ranking feature value of the target pushed commodity.

Step 4024, updating the ranking of the commodity information of each target pushed commodity in the target information to be pushed based on the recall ranking characteristic value of each target pushed commodity.

As an example, the executive body may update the ranking of the product information of each target pushed product in the target information to be pushed based on the recall ranking characteristic value of each target pushed product obtained in step 4023 by the following steps: the method comprises the steps that firstly, an execution main body updates the ordering of commodity information of each target push commodity according to the sequence of recall ordering characteristic values from large to small, and then inserts each target ordering position in initial information to be pushed in sequence according to the updated ordering of the commodity information of each target push commodity so as to update the ordering of each target push commodity in the information to be pushed.

In a specific example, the target information to be pushed comprises commodity information of a mobile phone, a tablet, a skin care product, a basketball, wine, stationery, a football, a jacket, shoes and the like which are sequentially arranged, wherein the skin care product, the stationery and the jacket are target pushed commodities, and a sequencing position corresponding to each target pushed commodity is a target sequencing position. After the executing main body pushes the target information to be pushed to the client, the executing main body executes the offline recall step in step 402 every preset recall period, which may be, for example, one day, assuming that the recall characteristic value of the skin care product is 80 and the level weight value is 0.5; the recalling characteristic value of the stationery is 90, and the grade weight value is 0.4; the recall feature value for the jacket was 80 and the rank weight value was 0.6. The recall ranking characteristic value for skin care products is 40, the recall ranking characteristic value for stationery is 36, and the recall ranking characteristic value for the jacket is 48. And then the execution main body updates the target information to be pushed according to the recall sequencing characteristic value of each target pushed commodity, and the obtained updated target information to be pushed comprises commodity information such as mobile phones, flat plates, coats, basketballs, wine, skin care products, football, stationery, shoes and the like which are sequentially arranged.

And step 403, pushing the updated information to be pushed of the target.

In some optional implementations of this embodiment, the rank weight value is determined via: acquiring historical characteristic information and value attributes of the target pushed commodity within a second preset time length, wherein the historical characteristic information comprises second preset type user behavior information; determining the weighted sum of various types of user behavior information in the historical characteristic information as a user behavior characteristic value of the target pushed commodity; determining the grade of the target pushed commodity based on a first difference value and a second difference value of the target pushed commodity, wherein the first difference value is the difference between the user behavior characteristic value of the target pushed commodity and the mean value of the user behavior characteristic values of all the target pushed commodities, and the second difference value is the difference between the value attribute of the target pushed commodity and the mean value of the value attribute of all the target pushed commodities; and determining the grade weight value of the target pushed commodity based on the corresponding relation between the preset grade and the grade weight value.

In this implementation manner, the historical characteristic information may include incremental data generated by the user executing a second preset type of behavior on the target pushed commodity within a second preset time period, so that the degree of behavior precipitation of the user on the target pushed commodity may be represented by the historical characteristic information. For example, the second preset duration may be one month, the second preset type of user behavior may be ordering, joining a shopping cart, sharing a link or clicking, and the like, and the historical feature information may include the number of times that the user performs the above behavior on the target pushed commodity within one month. The first difference value may represent a relative position of the user behavior precipitation degree of the target pushed commodity in all the target pushed commodities, for example, if the first difference value is a large positive number, it indicates that the user behavior precipitation degree of the target pushed commodity is located at a front position in all the target pushed commodities; the first difference is a small negative number, indicating that the user behavior precipitation degree of the target pushed commodity is in a later position in all the target pushed commodities.

In this implementation, the value attribute may represent the profit margin of the target pushed commodity, so that the level weight value may reflect the profit margin of the target pushed commodity. The second difference value may characterize the relative position of the profit margin of the target pushed good among all the target pushed goods. For example, the second difference is a large positive number indicating that the profit margin for the target pushed good is at a more advanced position on all target pushes; the second difference is a small negative number indicating that the profit margin for the target pushed commodity is at a later position among all the target pushed commodities.

In this implementation manner, the grade of the target pushed commodity is determined based on the first difference and the second difference, so that the user behavior precipitation degree and the profit margin of the target pushed commodity can be considered for the grade weight value of the target pushed commodity. For example, if the first difference value of the target pushed commodity is a large positive number, and the second difference value is a small negative number, it indicates that the user behavior of the target pushed commodity is more precipitated, but the profit margin is lower, and the grade of the target pushed commodity can be determined to be higher; if the first difference value of the target pushed commodity is a larger positive number and the second difference value is also a larger positive number, the user behavior precipitation degree of the target pushed commodity is higher, meanwhile, the profit margin is also higher, and the grade of the target pushed commodity can be determined to be high; if the first difference value of the target pushed commodity is a small negative number and the second difference value is a small negative number, which indicates that the user behavior deposition degree and the profit margin of the target pushed commodity are both low, the grade of the target pushed commodity can be determined to be low.

As an example, the first difference value and the second difference value may be divided into a plurality of sections, respectively, and then the correspondence relationship between each section of the first difference value and each section of the second difference value and the rank may be established, whereby the execution main body may determine the rank of the target pushed commodity based on the first difference value and the second difference value of each target pushed commodity.

In a preferred embodiment of this implementation, determining the rank of the target pushed commodity based on the first difference and the second difference of the target pushed commodity comprises: establishing a coordinate system by taking the mean value of the user behavior characteristic values and the mean value of the value attributes of the target pushed commodities as origin coordinates and the user behavior characteristic values and the value attributes of the target pushed commodities as coordinates; determining the position of each target pushed commodity in the coordinate system based on the first difference value and the second difference value of each target pushed commodity; and determining the grade of each target pushed commodity based on the corresponding relation between the preset coordinate system quadrant and the grade.

In this implementation, each quadrant in the coordinate system may represent the user behavior deposition degree and the profit margin of the target pushed commodity at the same time. For example, if the execution subject uses the value attribute as the abscissa and the user behavior feature value as the ordinate, the first quadrant indicates that the profit margin of the target pushed commodity is high, the user behavior precipitation degree is high, and mapping in the actual application scenario is that the behavior of the user clicking to access the target pushed commodity is frequent and the e-commerce platform or the business user can obtain a large profit, so the level of the target pushed commodity in the quadrant can be determined as a high level to obtain a large level weight value.

The second quadrant indicates that the user behavior of the targeted push good is precipitated to a high degree, but the profit margin is low. The mapping in the actual application scene is that the behavior of clicking to access the target pushed commodity by the user is frequent but the profit obtained by the e-commerce platform or the business user is low, and the target pushed commodity in the quadrant can be used as a drainage commodity, so that the grade of the target pushed commodity in the quadrant can be determined to be a lower grade to obtain a lower grade weight value;

the third quadrant indicates that the user behavior of the targeted push good is low in both the degree of precipitation and the amount of profit. The mapping in the actual application scenario is that the behavior of the user for clicking to access the target push commodity is less, and the profit obtained by the e-commerce platform or the business user is lower. The level of the targeted pushed goods in this quadrant can thus be determined to be a low level to obtain a small level weight value.

The fourth quadrant indicates that the user behavior of the targeted push good is precipitated to a low degree, but the profit margin is high. In the practical application scenario, the user has less behavior of clicking to access the target pushed commodity but the profit obtained by the e-commerce platform or the service user is higher. Exposure opportunities may be increased to bring higher revenue, and therefore, the level of the target pushed goods in this quadrant may be determined to be a higher level to obtain a larger level weight value.

With further reference to fig. 5, a flow 500 of yet another embodiment of a method for pushing information of the present disclosure is shown, the flow comprising the steps of:

step 501, pushing the target information to be pushed, which corresponds to step 401, and is not described herein again.

Step 502, in response to determining that the exposure times of the commodity information of the target pushed commodity reach a preset exposure time threshold, taking the product of the initial sorting characteristic value of the target commodity to be pushed and a preset exposure weight value as an updated sorting characteristic value, wherein the exposure weight value is negatively related to the exposure times.

Generally, the more times the same piece of information is exposed, the lower the interest level of the user in the information, and thus in this embodiment, the exposure weight value may represent the influence of the exposure times of the target push goods on the interest level of the user.

Step 503, updating the ranking of the commodity information of each target pushed commodity in the target information to be pushed based on the updated ranking characteristic value of the commodity information of each target pushed commodity.

By way of example, the target information to be pushed comprises the commodity information of the following SKUs arranged in sequence: mobile phones, tablets, skin care products, basketballs, wine, stationery, football, coats and shoes. Wherein, basketball and football are target propelling movement commodity, and its initial sequencing eigenvalue is 90 and 80 respectively, if the exposure number of times of predetermineeing is 10, after the execution main part detects that the exposure number of times of basketball reaches 10, can confirm that the exposure weight value of basketball is 0.6 based on the predetermined corresponding relation of exposure number of times and exposure weight value, then the initial sequencing eigenvalue of basketball is updated to 54, and then the target waits that the propelling movement information is updated to be: mobile phones, tablets, skin care products, football, wine, stationery, basketballs, coats and shoes.

Step 504, executing an offline updating step on the target information to be pushed every preset recall period, which corresponds to the step 402 and is not described herein again.

And 505, pushing the updated information to be pushed of the target.

It should be noted that, in this embodiment, the real-time updating step represented by step 502 and step 503 is to update the target to-be-pushed information based on the initial ranking characteristic value of the target pushed commodity, and the offline updating step represented by step 504 is to update the target to-be-pushed information based on the recall ranking characteristic value of the target pushed commodity, which are independent from each other and have no sequential limitation.

Compared with the embodiment shown in fig. 4, the embodiment shown in fig. 5 for pushing information highlights that the ordering positions of the commodity information of each target push commodity in the target information to be pushed are updated in real time according to the exposure times, so that the situation that the browsing interest of a user is reduced due to excessive exposure times of the commodity information of the target push commodity can be avoided, and meanwhile, the pertinence of the target information to be pushed can be further improved by combining the real-time updating with the offline updating.

Referring next to fig. 6, as an implementation of the methods shown in the above-mentioned figures, the present disclosure provides an embodiment of an apparatus for generating information, which corresponds to the method embodiment shown in fig. 2, and which is particularly applicable in various electronic devices.

As shown in fig. 6, the apparatus 600 for generating information of the present embodiment includes: an initial information obtaining unit 601 configured to obtain initial information to be pushed, where the initial information to be pushed includes commodity information of initial pushed commodities arranged in sequence; a ranking position determining unit 602 configured to determine at least one target ranking position from the initial information to be pushed based on a correspondence between predetermined commodity information ranking positions and user loss amounts; the target information generating unit 603 is configured to insert commodity information of a predetermined target pushed commodity at each target sorting position, so as to obtain target information to be pushed.

In the present embodiment, the target information generation unit 603 includes: a demand information receiving module configured to receive push demand information, the push demand information including a target push commodity and a target push user; the characteristic information acquisition module is configured to acquire commodity characteristic information of the target pushed commodity, historical sales information of the target pushed commodity within a first preset time length and user characteristic information of a target pushing user, wherein the user characteristic information comprises user behavior characteristics and/or a user portrait; the initial sorting characteristic value determining module is configured to determine an initial sorting characteristic value of the target pushed commodity based on the commodity characteristic information, the historical sales amount information and the user characteristic information; the information sorting module is configured to sort the commodity information of the target pushed commodities according to the descending order of the initial sorting characteristic values of the target pushed commodities; and the information insertion module is configured to insert the commodity information of each target pushed commodity into each target sorting position in sequence according to the sequence of the target sorting positions from front to back to obtain the target information to be pushed.

In the present embodiment, the historical sales information includes: the order quantity, the order quantity and the total volume of the transaction; and the initial ranking feature value determination module is further configured to: coding the commodity feature information to generate a commodity feature vector; encoding the user characteristic information to generate a user characteristic vector; determining the distance between the commodity feature vector and the user feature vector as a user preference feature value of the target pushed commodity; respectively normalizing the order quantity, the order quantity and the total transaction quantity to obtain the normalized order quantity, the normalized order quantity and the normalized total transaction quantity; determining the weighted sum of the normalized order quantity, the normalized order quantity and the normalized transaction total quantity as a commodity quality characteristic value of the target pushed commodity; and determining the weighted sum of the user preference characteristic value and the commodity quality characteristic value as the initial sorting characteristic value of the target pushed commodity.

Referring next to fig. 7, as an implementation of the methods shown in the above-mentioned figures, the present disclosure provides an embodiment of an apparatus for pushing information, which corresponds to the method embodiment shown in fig. 4, and which is particularly applicable to various electronic devices.

As shown in fig. 7, the apparatus 700 for pushing information of the present embodiment includes: an information pushing unit 701 configured to push target information to be pushed, the target information to be pushed being generated by the above method for generating information; the information updating unit 702 is configured to perform the following offline updating steps on the target information to be pushed at preset recall intervals: the method comprises the steps of obtaining recall information of each target pushed commodity in target information to be pushed in a last recall period, wherein the recall information comprises first preset type user behavior information; adopting a pre-trained recall model, determining a recall characteristic value of a target pushed commodity based on recall information of the target pushed commodity, wherein the recall characteristic value represents the matching degree of commodity information of the target pushed commodity and user behaviors; determining a recall ranking characteristic value of each target pushed commodity based on a recall characteristic value of the target pushed commodity and a predetermined grade weight value of the target pushed commodity; updating the ranking of each target pushed commodity in the target information to be pushed based on the recall ranking characteristic value of each target pushed commodity; the information pushing unit 701 is further configured to: and pushing the updated information to be pushed of the target.

In this embodiment, the apparatus 700 further comprises: the historical information acquisition unit is configured to acquire historical characteristic information and value attributes of each target pushed commodity within a second preset time length, wherein the historical characteristic information comprises second preset types of user behavior information; the behavior characteristic determining unit is configured to determine the weighted sum of various types of user behavior information in the historical characteristic information as a user behavior characteristic value of the target pushed commodity; a grade determining unit configured to determine a grade of the target pushed commodity based on a first difference value and a second difference value of the target pushed commodity, wherein the first difference value is a difference between a user behavior feature value of the target pushed commodity and a mean value of the user behavior feature values of the target pushed commodities, and the second difference value is a difference between a value attribute of the target pushed commodity and a mean value of the value attributes of the target pushed commodities; and the grade weight determining unit is configured to determine a grade weight value of the target pushed commodity based on a corresponding relation between a preset grade and the grade weight value.

In this embodiment, the rank determination unit further includes: the coordinate system building module is configured to build a coordinate system by taking the mean value of the user behavior characteristic values and the mean value of the value attributes of the target pushed commodities as origin coordinates and taking the user behavior characteristic values and the value attributes of the target pushed commodities as coordinate axes; a position determination module configured to determine a position of each target pushed commodity in the coordinate system based on the first difference value and the second difference value of each target pushed commodity; and the grade determining module is configured to determine the grade of each target pushed commodity based on the corresponding relation between the preset coordinate system quadrant and the grade.

In this embodiment, the apparatus 700 further comprises a real-time updating unit configured to: in response to the fact that the exposure times of the commodity information of the target pushed commodity reach a preset exposure time threshold, taking the product of the initial sorting characteristic value of the target to-be-pushed commodity and a preset exposure weight value as an updated sorting characteristic value, wherein the exposure weight value is negatively related to the exposure times; and updating the ordering of the commodity information of each target pushed commodity in the target information to be pushed based on the updated ordering characteristic value of the commodity information of each target pushed commodity.

Referring now to fig. 8, a schematic diagram of an electronic device (e.g., a server or terminal device of fig. 1) 800 suitable for use in implementing embodiments of the present disclosure is shown. The terminal device in the embodiments of the present disclosure may include, but is not limited to, a mobile terminal such as a mobile phone, a notebook computer, a digital broadcast receiver, a PDA (personal digital assistant), a PAD (tablet computer), and the like, and a fixed terminal such as a digital TV, a desktop computer, and the like. The terminal device shown in fig. 8 is only an example, and should not bring any limitation to the functions and the use range of the embodiments of the present disclosure.

As shown in fig. 8, an electronic device 800 may include a processing means (e.g., central processing unit, graphics processor, etc.) 801 that may perform various appropriate actions and processes in accordance with a program stored in a Read Only Memory (ROM)802 or a program loaded from a storage means 808 into a Random Access Memory (RAM) 803. In the RAM803, various programs and data necessary for the operation of the electronic apparatus 800 are also stored. The processing apparatus 801, the ROM 802, and the RAM803 are connected to each other by a bus 804. An input/output (I/O) interface 805 is also connected to bus 804.

Generally, the following devices may be connected to the I/O interface 805: input devices 806 including, for example, a touch screen, touch pad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; output devices 807 including, for example, a Liquid Crystal Display (LCD), speakers, vibrators, and the like; storage 808 including, for example, magnetic tape, hard disk, etc.; and a communication device 809. The communication means 809 may allow the electronic device 800 to communicate wirelessly or by wire with other devices to exchange data. While fig. 8 illustrates an electronic device 800 having various means, it is to be understood that not all illustrated means are required to be implemented or provided. More or fewer devices may alternatively be implemented or provided. Each block shown in fig. 8 may represent one device or may represent multiple devices as desired.

In particular, according to an embodiment of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method illustrated in the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network through the communication means 809, or installed from the storage means 808, or installed from the ROM 802. The computer program, when executed by the processing apparatus 801, performs the above-described functions defined in the methods of the embodiments of the present disclosure. It should be noted that the computer readable medium described in the embodiments of the present disclosure may be a computer readable signal medium or a computer readable storage medium or any combination of the two. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In embodiments of the disclosure, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In embodiments of the present disclosure, however, a computer readable signal medium may comprise a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, optical cables, RF (radio frequency), etc., or any suitable combination of the foregoing.

The computer readable medium may be embodied in the electronic device; or may exist separately without being assembled into the electronic device. The computer readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to: acquiring initial information to be pushed, wherein the initial information to be pushed comprises commodity information of initial pushed commodities which are arranged in sequence; determining at least one target sorting position from the initial information to be pushed based on the corresponding relation between the predetermined commodity information sorting positions and the user loss quantity; and inserting commodity information of a predetermined target pushed commodity into each target sequencing position to obtain target information to be pushed. And/or the computer readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to: pushing target information to be pushed, wherein the target information to be pushed is generated by the method for generating information; every interval of a preset recall cycle, the following off-line updating steps are executed on the target information to be pushed: the method comprises the steps of obtaining recall information of commodity information of each target pushed commodity in target information to be pushed in a last recall period, wherein the recall information comprises first preset type user behavior information; adopting a pre-trained recall model, determining a recall characteristic value of the target pushed commodity based on recall information, wherein the recall characteristic value represents the matching degree of the commodity information of the target pushed commodity and user behaviors; determining a recall ranking characteristic value of each target pushed commodity based on a recall characteristic value of the target pushed commodity and a predetermined grade weight value of the target pushed commodity; updating the ranking of the commodity information of each target pushed commodity in the target information to be pushed based on the recall ranking characteristic value of each target pushed commodity; the method further comprises the following steps: and pushing the updated information to be pushed of the target.

Computer program code for carrying out operations for embodiments of the present disclosure may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C + +, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).

The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

The units described in the embodiments of the present disclosure may be implemented by software or hardware. The described units may also be provided in a processor, and may be described as: a processor includes an initial information acquisition member, a ranking position determination member, and a target information generation member. The names of these units do not in some cases constitute a limitation on the unit itself, and for example, the initial information generation unit may also be described as a "unit that acquires initial information to be pushed".

The foregoing description is only exemplary of the preferred embodiments of the disclosure and is illustrative of the principles of the technology employed. It will be appreciated by those skilled in the art that the scope of the invention in the embodiments of the present disclosure is not limited to the specific combination of the above-mentioned features, but also encompasses other embodiments in which any combination of the above-mentioned features or their equivalents is made without departing from the inventive concept as defined above. For example, the above features and (but not limited to) technical features with similar functions disclosed in the embodiments of the present disclosure are mutually replaced to form the technical solution.

Claims

1. A method of generating information, comprising:

acquiring initial information to be pushed, wherein the initial information to be pushed comprises commodity information of initial pushed commodities which are arranged in sequence;

determining at least one target sorting position from the initial information to be pushed based on the corresponding relation between the predetermined commodity information sorting positions and the user loss quantity;

and inserting commodity information of a predetermined target pushed commodity into each target sequencing position to obtain target information to be pushed.

2. The method of claim 1, wherein the target information to be pushed is generated via:

receiving push demand information, wherein the push demand information comprises a target push commodity and a target push user;

acquiring commodity characteristic information of the target pushed commodity, historical sales information of the target pushed commodity within a first preset time length and user characteristic information of the target pushed user, wherein the user characteristic information comprises user behavior characteristics and/or user portrait;

determining an initial sequencing characteristic value of the target pushed commodity based on the commodity characteristic information, the historical sales information and the user characteristic information;

arranging the commodity information of each target pushed commodity according to the descending order of the initial sequencing characteristic value of each target pushed commodity;

and respectively inserting the commodity information of each target pushed commodity into each target sequencing position in sequence according to the sequence of each target sequencing position from front to back to obtain the target information to be pushed.

3. The method of claim 2, wherein the historical sales information comprises: the order quantity, the order quantity and the total volume of the transaction; and, the initial ranking characteristic value is determined via the steps of:

coding the commodity feature information to generate a commodity feature vector;

encoding the user characteristic information to generate a user characteristic vector;

determining the distance between the commodity feature vector and the user feature vector as a user preference feature value of the target pushed commodity;

normalizing the order quantity, the order quantity and the total transaction quantity respectively to obtain normalized order quantity, normalized order quantity and normalized total transaction quantity;

determining the weighted sum of the normalized order quantity, the normalized order quantity and the normalized transaction total quantity as a commodity quality characteristic value of the target pushed commodity;

and determining the weighted sum of the user preference characteristic value and the commodity quality characteristic value as the initial sorting characteristic value of the target pushed commodity.

4. A method for pushing information, comprising:

pushing target information to be pushed, the target information to be pushed being generated via the method of one of claims 1 to 3;

every interval of a preset recall cycle, executing the following off-line updating steps on the target information to be pushed: obtaining recall information of commodity information of each target pushed commodity in the target information to be pushed in a last recall period, wherein the recall information comprises first preset type user behavior information; determining a recall characteristic value of the target pushed commodity by adopting a pre-trained recall model based on the recall information, wherein the recall characteristic value represents the matching degree of the commodity information of the target pushed commodity and the user behavior; determining a recall ranking characteristic value of each target pushed commodity based on the recall characteristic value of the target pushed commodity and a predetermined grade weight value of the target pushed commodity; updating the ranking of the commodity information of each target pushed commodity in the target information to be pushed based on the recall ranking characteristic value of each target pushed commodity;

the method further comprises the following steps: and pushing the updated information to be pushed of the target.

5. The method of claim 4, wherein the rank weight value is determined via:

acquiring historical characteristic information and value attributes of the target pushed commodity within a second preset time length, wherein the historical characteristic information comprises second preset type user behavior information;

determining the weighted sum of various types of user behavior information in the historical characteristic information as the user behavior characteristic value of the target pushed commodity;

determining the grade of the target pushed commodity based on a first difference value and a second difference value of the target pushed commodity, wherein the first difference value is the difference between the user behavior characteristic value of the target pushed commodity and the mean value of the user behavior characteristic values of the target pushed commodities, and the second difference value is the difference between the value attribute of the target pushed commodity and the mean value of the value attribute of the target pushed commodity;

and determining the grade weight value of the target pushed commodity based on the corresponding relation between the preset grade and the grade weight value.

6. The method of claim 5, wherein determining the rank of the targeted push good based on the first difference and the second difference of the targeted push good comprises:

establishing a coordinate system by taking the mean value of the user behavior characteristic values and the mean value of the value attributes of the target pushed commodities as origin coordinates and the user behavior characteristic values and the value attributes of the target pushed commodities as coordinate axes;

determining the position of each target pushed commodity in the coordinate system based on the first difference value and the second difference value of each target pushed commodity;

and determining the grade of each target pushed commodity based on the corresponding relation between the preset coordinate system quadrant and the grade.

7. The method according to one of claims 4 to 6, wherein the method further comprises the step of updating the target information to be pushed in real time as follows:

in response to the fact that the exposure times of the commodity information of the target pushed commodity reach a preset exposure time threshold, taking the product of the initial sorting characteristic value of the target commodity to be pushed and a preset exposure weight value as an updated sorting characteristic value, wherein the exposure weight value is negatively related to the exposure times;

updating the ranking of the commodity information of each target pushed commodity in the target information to be pushed based on the updated ranking characteristic value of the commodity information of each target pushed commodity.

8. An apparatus for pushing information, comprising:

an initial information acquisition unit configured to acquire initial information to be pushed, the initial information to be pushed including commodity information of initial pushed commodities arranged in sequence;

the sorting position determining unit is configured to determine at least one target sorting position from the initial information to be pushed based on a corresponding relation between predetermined commodity information sorting positions and user loss quantity;

and the target information generating unit is configured to insert commodity information of a predetermined target pushed commodity into each target sequencing position to obtain target information to be pushed.

9. The apparatus of claim 8, wherein the target information generating unit comprises:

a demand information receiving module configured to receive push demand information including a target push commodity and a target push user;

the characteristic information acquisition module is configured to acquire commodity characteristic information of the target pushed commodity, historical sales information of the target pushed commodity within a first preset time length and user characteristic information of the target pushed user, wherein the user characteristic information comprises user behavior characteristics and/or a user portrait;

an initial ranking characteristic value determination module configured to determine an initial ranking characteristic value of the target pushed commodity based on the commodity characteristic information, the historical sales amount information, and the user characteristic information;

the information sorting module is configured to sort the commodity information of each target pushed commodity according to the descending order of the initial sorting characteristic value of each target pushed commodity;

the information insertion module is configured to insert the commodity information of each target pushed commodity into each target sorting position in sequence according to the sequence of each target sorting position from front to back so as to obtain the target information to be pushed.

10. The apparatus of claim 9, wherein the historical sales information comprises: the order quantity, the order quantity and the total volume of the transaction; and the initial ranking feature value determination module is further configured to:

11. An apparatus for pushing information, comprising:

an information pushing unit configured to push target information to be pushed, the target information to be pushed being generated via the method of one of claims 1 to 3;

the information updating unit is configured to execute the following offline updating steps on the target information to be pushed at intervals of a preset recall cycle: obtaining recall information of commodity information of each target pushed commodity in the target information to be pushed in a last recall period, wherein the recall information comprises first preset type user behavior information; determining a recall characteristic value of the target pushed commodity based on recall information of the target pushed commodity by adopting a pre-trained recall model, wherein the recall characteristic value represents the matching degree of the target pushed commodity and user behaviors; determining a recall ranking characteristic value of each target pushed commodity based on the recall characteristic value of the target pushed commodity and a predetermined grade weight value of the target pushed commodity; updating the ranking of each target pushed commodity in the target information to be pushed based on the recall ranking characteristic value of each target pushed commodity;

the information pushing unit is further configured to: and pushing the updated information to be pushed of the target.

12. The apparatus of claim 11, wherein the apparatus further comprises:

the historical information acquisition unit is configured to acquire historical characteristic information and value attributes of each target pushed commodity within a second preset time length, wherein the historical characteristic information comprises second preset type of user behavior information;

a behavior feature determination unit configured to determine a weighted sum of various types of user behavior information in the history feature information as a user behavior feature value of the target pushed commodity;

a grade determining unit configured to determine a grade of the target pushed commodity based on a first difference value and a second difference value of the target pushed commodity, wherein the first difference value is a difference between a user behavior feature value of the target pushed commodity and a mean value of the user behavior feature values of the target pushed commodities, and the second difference value is a difference between a value attribute of the target pushed commodity and a mean value of the value attributes of the target pushed commodities;

and the grade weight determining unit is configured to determine a grade weight value of the target pushed commodity based on a corresponding relation between a preset grade and the grade weight value.

13. The apparatus of claim 12, wherein the rank determination unit further comprises:

the coordinate system building module is configured to build a coordinate system by taking the mean value of the user behavior characteristic values and the mean value of the value attributes of the target pushed commodities as origin coordinates and taking the user behavior characteristic values and the value attributes of the target pushed commodities as coordinate axes;

a position determination module configured to determine a position of each of the target pushed commodities in the coordinate system based on a first difference value and a second difference value of each of the target pushed commodities;

the grade determining module is configured to determine the grade of each target pushed commodity based on the corresponding relation between preset coordinate system quadrants and the grades.

14. The apparatus according to one of claims 11 to 13, wherein the apparatus further comprises a real-time update unit configured to:

15. An electronic device, comprising:

one or more processors;

a storage device having one or more programs stored thereon,

when executed by the one or more processors, cause the one or more processors to implement the method of any one of claims 1-3.

16. An electronic device, comprising:

one or more processors;

a storage device having one or more programs stored thereon,

when executed by the one or more processors, cause the one or more processors to implement the method of any one of claims 4-7.

17. A computer-readable medium, on which a computer program is stored, wherein the program, when executed by a processor, implements the method of any one of claims 1-3.

18. A computer-readable medium, on which a computer program is stored, wherein the program, when executed by a processor, implements the method of any one of claims 4-7.