CN117932087A

CN117932087A - User portrait generation method and device, electronic equipment and storage medium

Info

Publication number: CN117932087A
Application number: CN202410115288.0A
Authority: CN
Inventors: 许楠楠; 李翛然; 赵守宣
Original assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Current assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Priority date: 2024-01-26
Filing date: 2024-01-26
Publication date: 2024-04-26

Abstract

The disclosure provides a user portrait generation method, a user portrait generation device, electronic equipment and a storage medium, relates to the technical field of computers, and particularly relates to the technical field of artificial intelligence such as knowledge graph, data processing and the like. The specific implementation scheme is as follows: acquiring commodity sharing behaviors among users, first interaction behaviors between each user and a first commodity, second interaction behaviors between each user and a commodity anchor, and second commodities associated with the commodity anchor; generating a first knowledge graph according to commodity sharing behaviors; generating a second knowledge graph corresponding to each user according to the first interaction behavior; generating a third knowledge graph corresponding to each user by the second interaction behavior and the second commodities associated with the commodity anchor; and analyzing at least one of the first knowledge graph, the second knowledge graph and the third knowledge graph to obtain a target user portrait corresponding to each user.

Description

User portrait generation method and device, electronic equipment and storage medium

Technical Field

The disclosure relates to the technical field of computers, in particular to the technical fields of knowledge maps, data processing and the like, and specifically relates to a user portrait generation method, a user portrait generation device, electronic equipment and a storage medium.

Background

User portrayal, namely user information tagging (such as gender, age, etc.), is that after collecting and analyzing data of main information such as user static attribute, social attribute, behavior attribute, etc., a basic mode that user's overall view is used for supporting big data applications such as personalized recommendation is abstracted. The user portraits have wider application prospect, especially in the field of electronic commerce, the user population and the equivalent value information of user demands can be rapidly and accurately positioned based on the user portraits, so that the user portraits have important roles in commodity searching, commodity recommending, advertising, accurate marketing and the like.

Disclosure of Invention

The disclosure provides a user portrait generation method, a user portrait generation device, electronic equipment and a storage medium.

According to a first aspect of the present disclosure, there is provided a user portrait generating method, including:

Acquiring commodity sharing behaviors among users, first interaction behaviors between each user and a first commodity, second interaction behaviors between each user and a commodity anchor, and second commodities associated with the commodity anchor;

generating a first knowledge graph according to the commodity sharing behavior;

generating a second knowledge graph corresponding to each user according to the first interaction behavior;

generating a third knowledge graph corresponding to each user according to the second interaction behavior and the second commodity associated with the commodity anchor;

And analyzing at least one of the first knowledge graph, the second knowledge graph and the third knowledge graph to obtain a target user portrait corresponding to each user.

According to a second aspect of the present disclosure, there is provided a user portrait generating apparatus, including:

The first acquisition module is used for acquiring commodity sharing behaviors between users, first interaction behaviors between each user and a first commodity, second interaction behaviors between each user and a commodity anchor and second commodities associated with the commodity anchor;

The first generation module is used for generating a first knowledge graph according to the commodity sharing behavior;

the second generation module is used for generating a second knowledge graph corresponding to each user according to the first interaction behavior;

the third generation module is used for generating a third knowledge graph corresponding to each user according to the second interaction behavior and the second commodities associated with the commodity anchor;

And the second acquisition module is used for analyzing at least one of the first knowledge graph, the second knowledge graph and the third knowledge graph so as to acquire a target user portrait corresponding to each user.

According to a third aspect of the present disclosure, there is provided an electronic device comprising:

At least one processor; and

A memory communicatively coupled to the at least one processor; wherein,

The memory stores instructions executable by the at least one processor to enable the at least one processor to perform the user representation generation method of the first aspect.

According to a fourth aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions for causing the computer to perform the user portrait generation method according to the first aspect.

According to a fifth aspect of the present disclosure, there is provided a computer program product comprising computer instructions which, when executed by a processor, implement the steps of the user representation generation method of the first aspect.

The user portrait generation method, device, electronic equipment and storage medium provided by the disclosure have the following beneficial effects:

In the embodiment of the disclosure, a first knowledge graph is generated according to commodity sharing behaviors between users, a second knowledge graph corresponding to each user is generated according to first interaction behaviors between the users and the first commodities, a third knowledge graph corresponding to each user is generated according to second interaction behaviors between each user and a commodity anchor and the second commodities associated with the commodity anchor, and at least one knowledge graph among the first knowledge graph, the second knowledge graph and the third knowledge graph is analyzed to obtain a target user portrait corresponding to each user. Therefore, a plurality of knowledge maps associated with the user can be generated according to the interactive behavior of the user on the commodity, and then the knowledge maps are analyzed, so that the portrait of the user can be quickly and accurately generated.

It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the disclosure, nor is it intended to be used to limit the scope of the disclosure. Other features of the present disclosure will become apparent from the following specification.

Drawings

The drawings are for a better understanding of the present solution and are not to be construed as limiting the present disclosure. Wherein:

FIG. 1 is a flow diagram of a method for generating a user representation according to an embodiment of the present disclosure;

FIG. 2 is a flow chart of a method for generating a user representation according to yet another embodiment of the present disclosure;

FIG. 3 is a flow chart of a method for generating a user representation according to yet another embodiment of the present disclosure;

FIG. 4 is a schematic diagram of a user portrait generating device according to an embodiment of the present disclosure;

FIG. 5 is a block diagram of an electronic device for implementing a user representation generation method of an embodiment of the present disclosure.

Detailed Description

Exemplary embodiments of the present disclosure are described below in conjunction with the accompanying drawings, which include various details of the embodiments of the present disclosure to facilitate understanding, and should be considered as merely exemplary. Accordingly, one of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

In the technical scheme of the disclosure, the related processes of collecting, storing, using, processing, transmitting, providing, disclosing and the like of the personal information of the user accord with the regulations of related laws and regulations, and the public order colloquial is not violated.

The following describes a user portrait generation method, apparatus, electronic device, and storage medium according to an embodiment of the present disclosure with reference to the accompanying drawings.

The execution body of the user portrait generating method according to this embodiment is a user portrait generating device, which may be implemented in software and/or hardware, and the device may be configured in an electronic apparatus, and the electronic apparatus may include, but is not limited to, a terminal, a server, and the like.

Fig. 1 is a flowchart of a method for generating a user portrait according to an embodiment of the present disclosure.

As shown in fig. 1, the user portrait generation method includes:

S101, acquiring commodity sharing behaviors among users, first interaction behaviors between each user and a first commodity, second interaction behaviors between each user and a commodity anchor, and second commodities associated with the commodity anchor.

The commodity sharing behavior between users may include that the user a shares a link of a commodity to the user B, and the user a shares the commodity to the user B in a text, voice, video, or the like.

The first interaction behavior may include a browsing behavior, a purchasing behavior, a comment behavior, a shopping cart adding behavior, a collecting behavior, and the like of the user on the first commodity. The present disclosure is not limited in this regard.

The second interaction behavior may include the user focusing on the merchandise anchor, watching live content of the merchandise anchor, the user clicking on a merchandise link of the merchandise anchor, the user purchasing merchandise through the merchandise anchor, and so on. The present disclosure is not limited in this regard.

Wherein the second merchandise associated with the merchandise host may be merchandise sold by the merchandise host.

S102, generating a first knowledge graph according to commodity sharing behaviors.

Wherein, a node in the first knowledge graph corresponds to a user. The edges between two nodes represent that commodity sharing behavior exists between the corresponding two users.

In some embodiments, the edges in the first knowledge-graph may be directed edges or undirected edges. If the commodity is a directed edge, when the user A shares the commodity to the user B, the direction of the first edge between the user A and the user B is that the user A points to the user B.

In some embodiments, edges in the first knowledge-graph may have corresponding first weights.

Specifically, the number of times that the user a shares the commodity with the user B may be counted according to the commodity sharing behavior, and then, according to the number of times that the commodity shares, a first weight corresponding to the directed edge of the user a pointing to the user B is determined.

Or counting the total commodity sharing times between the user A and the user B according to the commodity sharing behaviors, and then determining the first weight corresponding to the undirected edge between the user A and the user B according to the commodity sharing times.

It should be noted that, the more the commodity sharing times are, the larger the corresponding first weight is.

And S103, generating a second knowledge graph corresponding to each user according to the first interaction behavior.

It should be noted that, each user corresponds to a second knowledge graph, and each second knowledge graph corresponding to each user includes a second node corresponding to each user and at least one first commodity node corresponding to a first commodity, where when a user browses, clicks, purchases, architecture and reviews any first commodity, a second edge exists between the second node corresponding to the user and any first commodity.

In some embodiments, a second node corresponding to a user and a first commodity node corresponding to a first commodity are generated in the second knowledge graph, then a second edge between the second node and the first commodity node is generated according to the first interaction behavior, an interaction type corresponding to the first interaction behavior and a first interaction number corresponding to each interaction type are determined, and finally a second weight corresponding to the second edge is determined according to a preset weight and a third interaction number corresponding to each interaction type. Therefore, the second weight corresponding to each second side can be determined according to the interaction type and the first interaction times corresponding to each interaction type, and the favorite degree of the user on each first commodity can be represented in the second knowledge graph through the weights of the sides.

The interaction types may include browsing, purchasing, etc.

In the embodiment of the present disclosure, a corresponding weight may be set in advance for each interaction type. In some embodiments, the preset weight corresponding to the purchase type is greater than the preset weight corresponding to the purchase type, and the preset weight corresponding to the purchase type is greater than the preset weight corresponding to the browse type.

For example, the number of times the user browses the first commodity a is 5, the number of times the user purchases the first commodity a is 1, the number of times he purchases the first commodity a is 3, the preset weight corresponding to the purchase type is P1, the preset weight corresponding to the purchase type is P2, and the preset weight corresponding to the browse type is P3. The second weight of the edge between the second node corresponding to the user and the node of the second commodity a is: p1+3×p2+5×p3.

In some embodiments, the second edge may also have no corresponding second weight. The present disclosure is not limited in this regard.

And S104, generating a third knowledge graph corresponding to each user according to the second interaction behavior and the second commodities associated with the commodity anchor.

In some embodiments, the third knowledge graph may include a fourth node corresponding to the user, a third node corresponding to the product anchor, and a node corresponding to the second product, and the fourth edge between the fourth node corresponding to the user and the third node corresponding to the product anchor may be determined according to the second interaction behavior of each user on each product anchor, and then the edge between the third node corresponding to the product anchor and the node corresponding to the second product may be determined according to the association relationship between the product anchor and the second product.

In some embodiments, a third commodity corresponding to the second interaction behavior is determined, then a second commodity identical to the third commodity is determined, a target commodity associated with a commodity anchor is determined, a fourth node corresponding to a user and included in a third knowledge graph is generated, a third node corresponding to the commodity anchor, a second commodity node corresponding to the target commodity are generated, and then a third edge between the fourth node and the third node, a fourth edge between the third node and the target commodity, and a fifth edge between the third node and the third node are respectively generated according to the second interaction behavior, the association relationship between the commodity anchor and the target commodity, and the type of the commodity anchor. Therefore, the target commodity interacted with the user can be screened out, and the third knowledge graph is constructed by combining the target commodity, so that the relevance between the target commodity in the second knowledge graph and the user can be improved.

The third commodity corresponding to the second interaction behavior can be obtained by identifying the commodity related to the second interaction behavior. For example, the user clicks a link corresponding to a third item associated with the item anchor, the third item is included in a bullet screen that the user publishes at the live broadcast room, and the user purchases the third item through the item anchor.

Wherein, the type of the commodity anchor can be determined according to the category of the commodity sold by the commodity anchor. For example, if the commodity is a clothing, the type of the commodity is a clothing. In the embodiment of the disclosure, the second nodes corresponding to the same type of commodity anchor can be connected.

In some embodiments, the second interaction times between the user and the product anchor and the third interaction times between the user and the target product through the product anchor may be determined according to the second interaction behaviors, further, the third weight corresponding to the third side may be determined according to the second interaction times, and finally, the fourth weight corresponding to the fourth side may be determined according to the third interaction times.

The third weight of the third side can represent the favorites of the user on the commodity anchor. The greater the third weight of the third side, the higher the user's preference for the merchandise host.

The fourth weight of the fourth side can represent the interest degree of the user on the target commodity sold by the commodity anchor. The greater the fourth weight of the fourth side, the greater the degree of interest to the target commodity by the user.

Therefore, the third weight of the third side is determined according to the second interaction times between the user and the commodity anchor, and the fourth weight corresponding to the fourth side is determined according to the third interaction times between the user and the target commodity through the commodity anchor, so that the favorite degree of the user on the commodity anchor and the interested degree of each target commodity of the user can be represented through the weights of the sides in the third knowledge graph.

In some embodiments, a second merchandise that is the same as the first merchandise may also be determined to be a target merchandise associated with the merchandise host.

The first commodity is a commodity which is browsed, purchased by a user. Thus, a second item that is the same as the first item is identified as a target item associated with the item anchor, identifying that the user may also purchase the item through the item anchor. Therefore, the association between the user and the target commodity through the commodity anchor in the third knowledge graph can be perfected more.

In some embodiments, the nodes corresponding to the same type of commodity anchor can also be connected

In some embodiments, the attributes associated with the node to which the user corresponds may include the user's gender, age, location, etc. The attributes associated with the node to which the commodity corresponds may also include the category of the commodity, the commodity click rate, the brand of the commodity, and the like.

S105, analyzing at least one of the first knowledge graph, the second knowledge graph and the third knowledge graph to obtain a target user portrait corresponding to each user.

In some embodiments, a graph attention network may be employed to analyze at least one of the first, second, and third knowledge-graphs.

Specifically, a graph attention network may be adopted to learn the feature representation corresponding to each first node in the first knowledge graph; learning a characteristic representation of a second node corresponding to the user in the second knowledge graph; and learning the characteristic representation of the third node corresponding to each commodity anchor in the third knowledge graph, and further generating a target user portrait corresponding to the user according to the characteristic representation corresponding to the first node, and/or the characteristic representation corresponding to the second node, and/or the characteristic representation of the third node.

In some embodiments, a complex network analysis method may be further used to analyze at least one of the first, second, and third knowledge-maps.

In some embodiments, if the first knowledge-graph, the second knowledge-graph and the third knowledge-graph are analyzed at the same time to obtain the target user portrait corresponding to each user, a more complete user portrait can be generated.

as shown in fig. 2, the user portrait generation method includes:

S201, acquiring commodity sharing behaviors among users, first interaction behaviors between each user and a first commodity, second interaction behaviors between each user and a commodity anchor, and second commodities associated with the commodity anchor.

S202, generating a first knowledge graph according to commodity sharing behaviors.

And S203, generating a second knowledge graph corresponding to each user according to the first interaction behavior.

S204, generating a third knowledge graph corresponding to each user according to the second interaction behavior and the second commodities associated with the commodity anchor.

The specific implementation manner of step S201 to step S204 may refer to the detailed description in other embodiments in the disclosure, and will not be described in detail herein.

S205, determining a target liveness label corresponding to each user according to the first attribute of the first node corresponding to each user in the first knowledge graph.

In some embodiments, the plurality of first attributes are ranked in order from large to small to obtain a target ranking result, then the activity level corresponding to each user is determined according to the position of the first attribute corresponding to each user in the target ranking result, and finally the activity level corresponding to each user is determined as a target activity label corresponding to each user. Thus, the first attribute of each user can be combined

In some embodiments, the first attribute may be a first degree of centrality or a first cluster coefficient corresponding to the first node. In some embodiments, the first centrality or the first clustering coefficient corresponding to each first node may be further calculated in combination with the first weight corresponding to each first edge in the first knowledge-graph.

The first degree centrality and the first clustering coefficient can represent the importance degree of the user in the first knowledge graph, and further represent the activity degree of the user. The larger the first attribute, the more important the characterization user is in the first knowledge-graph.

Therefore, in the embodiment of the present disclosure, the plurality of first attributes are ranked in order from large to small, and then, according to the position of the first attribute corresponding to each user in the target ranking result, the activity level corresponding to each user is determined. The more forward the location, the higher the liveness level.

In some embodiments, multiple users may also be classified into multiple liveness levels based on the target ranking result. For example, the three classes of high activity, medium activity and low activity are classified.

In the embodiment of the disclosure, each user may be ranked by the first attribute, so as to determine the activity level of each user in all users, and further make the determined target activity label more accurate.

In some embodiments, the first attribute corresponding to each first node is updated according to the first attributes corresponding to other first nodes to which each first node is connected.

It should be noted that, when the user a shares the commodity to the user B, if the first attribute of the user B is higher, the user B may also share the commodity to other users, so that when the commodity is pushed to the user a, the sales promotion range of the commodity may be increased. Therefore, in the embodiment of the present disclosure, the first attribute of the first node may be updated according to the first attribute corresponding to the other first nodes connected to the first node, so as to obtain a more accurate first attribute.

In some embodiments, the first attributes corresponding to other first nodes may be added to obtain a sum of the attributes, and then the attributes and the first attributes corresponding to the first nodes are fused according to a preset proportion to obtain updated first attributes of the first nodes.

In some embodiments, according to commodity sharing behavior between every two users, determining the number of commodity sharing between every two users, according to the number of commodity sharing, determining a first weight corresponding to a first edge between every two first nodes in a first knowledge graph, then determining a correction value corresponding to each first node according to first attributes corresponding to other first nodes connected with each first node and the first weight, and finally updating the first attributes corresponding to each first node based on the correction value.

Specifically, according to a first weight corresponding to a first edge between a first node and other first nodes, first attributes corresponding to other first nodes are weighted and summed to obtain a correction value, and then the correction value is fused with the first attributes corresponding to the first nodes to obtain updated first attributes.

Therefore, under the condition that the first side in the second knowledge graph has the corresponding first weight, the correction value can be determined by combining the first weight, so that the corrected first attribute is more accurate.

S206, extracting sub-knowledge maps corresponding to each commodity category from the second knowledge maps according to the commodity category to which the first commodity belongs.

The categories of goods may include apparel, electrical appliances, fresh, household, snack, and the like, among others. Further, each category may be further divided in detail. For example, apparel may also be divided into coats, pants, skirts, shoes, bags, hats, hair accessories, and the like. The present disclosure is not limited to a particular classification of merchandise.

And each sub-knowledge graph comprises a second node corresponding to the user and a node corresponding to the commodity in the same category.

S207, determining target preference labels of the users for each commodity category according to second attributes of the second nodes corresponding to the users in each sub-knowledge graph.

In some embodiments, the second attribute may be a second degree of centrality corresponding to the second node. Optionally, a second centrality corresponding to the second node may also be calculated by combining a second weight corresponding to each second edge in the sub-knowledge graph.

In some embodiments, the second attribute corresponding to each sub-knowledge graph may be directly determined as a target preference label of the user for each commodity category.

In some embodiments, the plurality of second attributes may be further ranked in order from large to small, so as to obtain a first ranking result, then according to the position of the user in the first ranking result, corresponding to the second attribute in each sub-knowledge graph, a preference level of the user for each commodity category is determined, and finally, the preference level of the user for each commodity is determined as a target preference label of each user for each commodity category.

Wherein, the larger the second attribute corresponding to the commodity category, the higher the preference level of the user.

In some embodiments, the preference degree corresponding to each commodity category by the user may be divided into a plurality of preference levels according to the first sorting result. For example, three levels are classified as very preferred, generally preferred, and not preferred.

And S208, determining a target interaction strength label of the user for each commodity anchor according to the third attribute of the third node corresponding to each commodity anchor in the third knowledge graph.

In some embodiments, the third attribute of the third node may be a mediator centrality corresponding to the third node. The intermediation centrality of the third node corresponding to the commodity anchor can represent the probability that the user purchases the commodity through the commodity anchor. The higher the center of the intermediary corresponding to the commodity anchor, the higher the interaction strength between the user and the commodity anchor, and the higher the probability of purchasing the commodity through the commodity anchor.

In some embodiments, the third attribute of the third node may be determined directly as a target interaction strength tag for the user to the anchor of each item.

In some embodiments, the third attributes may be further ranked in order from large to small to obtain a second ranking result, then according to the position of the third attribute in the second ranking result, the interaction strength level of the user on each of the product anchors is determined, and finally the interaction strength level of the user on each of the product anchors is determined as the target interaction strength tag of each of the user on each of the product anchors.

In some embodiments, the interaction strength corresponding to each of the product anchor by the user may be divided into a plurality of interaction strength levels according to the second sorting result. For example, the method is divided into three grades with high interaction strength, general interaction strength and low interaction strength.

S209, generating a target user portrait according to at least one of the target liveness label, the target preference label and the target interaction strength label.

Specifically, at least one of a target liveness label, a target preference label and a target interaction strength label is determined as a target user portrait.

In the embodiment of the disclosure, after a first knowledge graph, a second knowledge graph and a third knowledge graph are generated, determining a target activity label corresponding to a user according to a first attribute of a first node corresponding to each user in the first knowledge graph, extracting a sub-knowledge graph corresponding to each commodity category from the second knowledge graph according to the commodity category to which the first commodity belongs, determining a target preference label of the user for each commodity category according to a second attribute of a second node corresponding to the user in each sub-knowledge graph, determining a target interaction strength label of the user for each commodity according to a third attribute of a third node corresponding to each commodity anchor in the third knowledge graph, and finally generating a target user portrait according to at least one of the target activity label, the target preference label and the target interaction strength label. Therefore, the first knowledge graph, the second knowledge graph and the third knowledge graph can be respectively analyzed to acquire the attribute information corresponding to the nodes related to the user in the whole knowledge graph, and the target user portrait can be further accurately generated.

as shown in fig. 3, the user portrait generation method includes:

S301, acquiring commodity sharing behaviors among users, first interaction behaviors between each user and a first commodity, second interaction behaviors between each user and a commodity anchor, and second commodities associated with the commodity anchor.

S302, generating a first knowledge graph according to commodity sharing behaviors.

S303, generating a second knowledge graph corresponding to each user according to the first interaction behavior.

S304, generating a third knowledge graph corresponding to each user according to the second interaction behavior and the second commodities associated with the commodity anchor.

S305, determining a target liveness label corresponding to each user according to the first attribute of the first node corresponding to each user in the first knowledge graph.

S306, extracting sub-knowledge maps corresponding to each commodity category from the second knowledge maps according to the commodity category to which the first commodity belongs.

S307, determining target preference labels of the users for each commodity category according to the second attributes of the second nodes corresponding to the users in each sub-knowledge graph.

And S308, determining a target interaction strength label of the user for each commodity anchor according to the third attribute of the third node corresponding to each commodity anchor in the third knowledge graph.

S309, generating a target user portrait according to at least one of the target liveness label, the target preference label and the target interaction strength label.

The specific implementation manner of step S301 to step S309 may refer to the detailed descriptions in other embodiments in the disclosure, and will not be described in detail herein.

S310, acquiring an initial user portrait corresponding to each user generated in the previous period.

The user portrait of the user can be updated in real time based on the preset period, so that the accuracy of the user is ensured, and therefore, the initial user portrait of the user generated in the previous period can be acquired.

The specific implementation form of the initial user portrait corresponding to the user generated in the previous period can refer to the implementation form of the target user portrait of the user generated in the application.

The initial user portrait comprises an initial activity label, an initial preference label and an initial interaction strength label of the user.

S311, determining the liveness change trend according to the initial liveness label corresponding to the initial user portrait and the target liveness label.

The activity change trend may include that the activity is strong and the activity is weak, and when the activity of the user is strong, the user can be actively pushed the favorite goods of the user.

S312, determining a preference change trend according to the initial preference label corresponding to the initial user image and the target preference label.

The preference change trend may include that the preference of the user for a certain category of goods is weakened and the preference of the user for the certain category of goods is weakened, and when the preference of the user for the certain category of goods is strengthened, the category of goods may be actively pushed to the user.

S313, determining the change trend of the interaction strength according to the initial interaction strength label corresponding to the initial user portrait and the target interaction strength label.

The interaction strength change trend may include that the interaction strength of the user to the commercial anchor is stronger and the interaction strength of the user to the commercial anchor is weaker, and when the interaction strength of the user to a commercial anchor is stronger, the commercial anchor can be actively pushed to the user, for example, when the commercial anchor is live, live broadcast information can be pushed to the user.

And S314, updating the target user portrait according to at least one of the activity change trend, the preference change trend and the interaction strength change trend.

Specifically, at least one of the activity change trend, the preference change trend, and the interaction strength change trend may be added to the target user portrait of the user.

In the embodiment of the disclosure, after generating the target user portrait according to at least one of the target liveness tag, the target preference tag and the target interaction strength tag, an initial user portrait corresponding to each user generated in the previous period may be obtained, then an liveness change trend is determined according to the initial liveness tag corresponding to the initial user portrait and the target liveness tag, a preference change trend is determined according to the initial preference tag corresponding to the initial user portrait and the target preference tag, an interaction strength change trend is determined according to the initial interaction strength tag corresponding to the initial user portrait and the target interaction strength tag, and finally the target user portrait is updated according to at least one of the liveness change trend, the preference change trend and the interaction strength change trend. Therefore, the method can analyze the change trend of the activity, the change trend of the preference and the change trend of the interaction strength of the user according to the initial user portrait of the previous period and the target user portrait of the current period, and update the target user portrait according to the change trend, so that the generated target user portrait can reflect the change of the user and the target user portrait is more perfect.

In some embodiments, the user's basic information (e.g., age, height, weight, gender, etc.) may also be added to the user's target user representation.

In some embodiments, emotion analysis may also be performed on comment text of the user, etc. to obtain the emotion tendencies of the user to the merchandise, and the emotion tendencies are also added to the target user portraits of the user.

In some embodiments, the search text of the user may be analyzed to obtain keywords searched by the user and a search frequency of each keyword, so as to add the keywords and the search frequency of each keyword into the target user portrait of the user.

Thus, the target user portrait of the user can be further improved.

FIG. 4 is a schematic diagram of a user portrait generating device according to an embodiment of the present disclosure

As shown in fig. 4, the user portrait generating apparatus 400 includes:

a first obtaining module 401, configured to obtain a commodity sharing behavior between users, a first interaction behavior between each user and a first commodity, a second interaction behavior between each user and a commodity anchor, and a second commodity associated with the commodity anchor;

A first generation module 402, configured to generate a first knowledge graph according to a commodity sharing behavior;

a second generating module 403, configured to generate a second knowledge graph corresponding to each user according to the first interaction behavior;

a third generating module 404, configured to generate a third knowledge graph corresponding to each user according to the second interaction behavior and the second merchandise associated with the merchandise host;

The second obtaining module 405 is configured to analyze at least one of the first knowledge-graph, the second knowledge-graph, and the third knowledge-graph, so as to obtain a target user representation corresponding to each user.

In some embodiments of the present disclosure, the second obtaining module 405 is configured to:

determining a target liveness label corresponding to each user according to a first attribute of a first node corresponding to each user in the first knowledge graph;

extracting a sub-knowledge graph corresponding to each commodity category from the second knowledge graph according to the commodity category to which the first commodity belongs;

Determining target preference labels of the users for each commodity category according to second attributes of the second nodes corresponding to the users in each sub-knowledge graph;

Determining a target interaction strength label of the user on each commodity anchor according to a third attribute of a third node corresponding to each commodity anchor in the third knowledge graph;

And generating a target user portrait according to at least one of the target liveness label, the target preference label and the target interaction strength label.

Sequencing the first attributes according to the sequence from big to small to obtain a target sequencing result;

Determining the activity level corresponding to each user according to the position of the first attribute corresponding to each user in the target sequencing result;

and determining the activity level corresponding to each user as a target activity label corresponding to each user.

And updating the first attribute corresponding to each first node according to the first attributes corresponding to other first nodes connected with each first node.

According to commodity sharing behaviors between every two users, determining commodity sharing times between every two users;

Determining a first weight corresponding to a first edge between every two first nodes in the first knowledge graph according to the commodity sharing times;

Determining a correction value corresponding to each first node according to first attributes and first weights corresponding to other first nodes connected with each first node;

and updating the first attribute corresponding to each first node based on the correction value.

In some embodiments of the present disclosure, wherein,

The first attribute is a first centrality or a first clustering coefficient corresponding to the first node;

The second attribute is a second centrality corresponding to the second node;

the third attribute is the intermediacy center corresponding to the third node.

In some embodiments of the present disclosure, the method further includes an update module configured to:

Acquiring an initial user portrait corresponding to each user generated in the previous period;

determining an activity change trend according to the initial activity label corresponding to the initial user portrait and the target activity label;

determining a preference change trend according to the initial preference label corresponding to the initial user image and the target preference label;

Determining the change trend of the interaction strength according to the initial interaction strength label corresponding to the initial user portrait and the target interaction strength label;

And updating the target user portrait according to at least one of the activity change trend, the preference change trend and the interaction strength change trend.

In some embodiments of the present disclosure, the second generating module 403 is configured to:

Generating a second node corresponding to the user and a first commodity node corresponding to the first commodity, which are contained in the second knowledge graph;

generating a second edge between a second node and a first commodity node contained in a second knowledge graph according to the first interaction behavior;

Determining an interaction type corresponding to the first interaction behavior and a first interaction number corresponding to each interaction type;

and determining a second weight corresponding to the second side according to the preset weight corresponding to each interaction type and the third interaction times.

In some embodiments of the present disclosure, the third generating module 404 is configured to:

Determining a third commodity corresponding to the second interaction behavior;

Determining a target commodity associated with a commodity anchor from a second commodity identical to the third commodity;

generating a fourth node corresponding to the user and contained in the third knowledge graph, a third node corresponding to the commodity anchor and a second commodity node corresponding to the target commodity;

And respectively generating a third side between a fourth node and a third node and a fourth side between the third node and the target commodity, which are contained in the third knowledge graph, according to the second interaction behavior and the association relation between the commodity anchor and the target commodity.

Determining a second interaction time between the user and the commodity anchor and a third interaction time between the user and the target commodity through the commodity anchor according to the second interaction behavior;

determining a third weight corresponding to a third side according to the second interaction times;

and determining a fourth weight corresponding to the fourth side according to the third interaction times.

A second merchandise identical to the first merchandise is determined as a target merchandise associated with the merchandise host.

The explanation of the method for generating a user image is also applicable to the device for generating a user image of the present embodiment, and will not be repeated here.

According to the embodiment of the disclosure, a first knowledge graph is generated according to commodity sharing behaviors among users, a second knowledge graph corresponding to each user is generated according to first interaction behaviors between the users and the first commodities, a third knowledge graph corresponding to each user is generated according to second interaction behaviors between each user and a commodity anchor and the second commodities related to the commodity anchor, and at least one knowledge graph among the first knowledge graph, the second knowledge graph and the third knowledge graph is analyzed to obtain a target user portrait corresponding to each user. Therefore, a plurality of knowledge maps associated with the user can be generated according to the interactive behavior of the user on the commodity, and then the knowledge maps are analyzed, so that the portrait of the user can be quickly and accurately generated.

According to embodiments of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

Fig. 5 illustrates a schematic block diagram of an example electronic device 500 that may be used to implement embodiments of the present disclosure. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the disclosure described and/or claimed herein.

As shown in fig. 5, the apparatus 500 includes a computing unit 501 that can perform various suitable actions and processes according to a computer program stored in a Read Only Memory (ROM) 502 or a computer program loaded from a storage unit 508 into a Random Access Memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the device 500 can also be stored. The computing unit 501, ROM 502, and RAM 503 are connected to each other by a bus 504. An input/output (I/O) interface 505 is also connected to bus 504.

Various components in the device 500 are connected to the I/O interface 505, including: an input unit 506 such as a keyboard, a mouse, etc.; an output unit 507 such as various types of displays, speakers, and the like; a storage unit 508 such as a magnetic disk, an optical disk, or the like; and a communication unit 509 such as a network card, modem, wireless communication transceiver, etc. The communication unit 509 allows the device 500 to exchange information/data with other devices through a computer knowledge-graph such as the internet and/or various telecommunications knowledge-graphs.

The computing unit 501 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of computing unit 501 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various computing units running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 501 performs the respective methods and processes described above, for example, a user portrait generation method. For example, in some embodiments, the user representation generation method may be implemented as a computer software program tangibly embodied on a machine-readable medium, such as storage unit 508. In some embodiments, part or all of the computer program may be loaded and/or installed onto the device 500 via the ROM 502 and/or the communication unit 509. When the computer program is loaded into RAM 503 and executed by computing unit 501, one or more steps of the user portrait creation method described above may be performed. Alternatively, in other embodiments, the computing unit 501 may be configured to perform the user representation generation method in any other suitable way (e.g. by means of firmware).

Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuit systems, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems On Chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.

Program code for carrying out methods of the present disclosure may be written in any combination of one or more programming languages. These program code may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus such that the program code, when executed by the processor or controller, causes the functions/operations specified in the flowchart and/or block diagram to be implemented. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.

In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on 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.

To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and pointing device (e.g., a mouse or trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.

The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a knowledge graph browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication knowledge graph). Examples of communication knowledge maps include: local Area Network (LAN), wide Area Network (WAN), internet and blockchain knowledge graph.

The computer system may include a client and a server. The client and server are typically remote from each other and typically interact through a communication knowledge-graph. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so that the defects of high management difficulty and weak service expansibility in the traditional physical hosts and VPS service ("Virtual PRIVATE SERVER" or simply "VPS") are overcome. The server may also be a server of a distributed system or a server that incorporates a blockchain.

It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps recited in the present disclosure may be performed in parallel or sequentially or in a different order, provided that the desired results of the technical solutions of the present disclosure are achieved, and are not limited herein.

Furthermore, the terms "first," "second," and the like, are used for descriptive purposes only and are not to be construed as indicating or implying a relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include at least one such feature. In the description of the present disclosure, the meaning of "a plurality" is at least two, such as two, three, etc., unless explicitly specified otherwise. In the description of the present disclosure, the words "if" and "if" as used may be interpreted as "at … …" or "at … …" or "in response to a determination" or "in the … … case".

The above detailed description should not be taken as limiting the scope of the present disclosure. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present disclosure are intended to be included within the scope of the present disclosure.

Claims

1. A user portrait generation method includes:

generating a first knowledge graph according to the commodity sharing behavior;

2. The method of claim 1, wherein the analyzing at least one of the first, second, and third knowledge-maps to obtain a target user representation for each of the users comprises:

Determining a target preference label of the user for each commodity category according to a second attribute of a second node corresponding to the user in each sub-knowledge graph;

Determining a target interaction strength label of the user for each commodity anchor according to a third attribute of a third node corresponding to each commodity anchor in the third knowledge graph;

and generating the target user portrait according to at least one of the target liveness label, the target preference label and the target interaction strength label.

3. The method of claim 2, wherein the determining, according to the first attribute of the first node corresponding to each user in the first knowledge-graph, the target liveness label corresponding to the user includes:

sorting the first attributes according to the order from big to small to obtain a target sorting result;

And determining the activity level corresponding to each user as the target activity label corresponding to each user.

4. The method of claim 3, wherein before said sorting the plurality of the first attributes in order from large to small to obtain a target sorting result, further comprising:

and updating the first attribute corresponding to each first node according to the first attribute corresponding to other first nodes connected with each first node.

5. The method of claim 4, wherein the updating the first attribute corresponding to each of the first nodes according to the first attributes corresponding to other first nodes to which each of the first nodes is connected comprises:

determining a correction value corresponding to each first node according to first attributes corresponding to other first nodes connected with each first node and the first weights;

6. The method according to any one of claims 2 to 5, wherein,

the second attribute is a second centrality corresponding to the second node;

And the third attribute is the intermediation centrality corresponding to the third node.

7. The method of any of claims 2-5, further comprising:

Determining an interaction strength change trend according to the initial interaction strength label corresponding to the initial user portrait and the target interaction strength label;

and updating the target user portrait according to at least one of the liveness change trend, the preference change trend and the interaction strength change trend.

8. The method of claim 1, wherein the generating a second knowledge-graph corresponding to each user according to the first interaction behavior comprises:

generating a second edge between the second node and the first commodity node contained in the second knowledge-graph according to the first interaction behavior;

9. The method of claim 1, wherein the generating a third knowledge-graph corresponding to each user according to the second interaction behavior and the second merchandise associated with the merchandise host comprises:

determining a third commodity corresponding to the second interaction behavior;

determining a target commodity associated with the commodity anchor from a second commodity identical to the third commodity;

Generating a fourth node corresponding to the user and contained in the third knowledge graph, wherein the third node corresponds to the commodity anchor and the second commodity node corresponds to the target commodity;

And respectively generating a third edge between the fourth node and the third node, a fourth edge between the third node and the target commodity and a fifth edge between the third node and the third node, which are contained in the third knowledge graph, according to the second interaction behavior, the association relation between the commodity anchor and the target commodity and the type of the commodity anchor.

10. The method of claim 9, further comprising:

Determining a third weight corresponding to the third side according to the second interaction times;

11. The method of claim 9, further comprising:

and determining a second commodity which is the same as the first commodity as a target commodity associated with the commodity caster.

12. A user portrait generation apparatus comprising:

13. The apparatus of claim 12, wherein the second acquisition module is configured to:

14. The apparatus of claim 13, wherein the second acquisition module is configured to:

15. The apparatus of claim 14, wherein the second acquisition module is configured to:

16. The apparatus of claim 15, wherein the second acquisition module is configured to:

17. The device according to any one of claims 13-16, wherein,

the second attribute is a second centrality corresponding to the second node;

18. The apparatus of any of claims 13-16, further comprising an update module to:

and updating the target user portrait with the liveness change trend, the preference change trend and the interaction strength change trend.

19. The apparatus of claim 12, wherein the second generation module is configured to:

20. The apparatus of claim 12, wherein the third generation module is configured to:

determining a third commodity corresponding to the second interaction behavior;

and respectively generating a third side between the fourth node and the third node and a fourth side between the third node and the target commodity, which are contained in the third knowledge graph, according to the second interaction behavior and the association relation between the commodity anchor and the target commodity.

21. The apparatus of claim 20, wherein the third generation module is configured to:

22. The apparatus of claim 20, wherein the third generation module is configured to:

23. An electronic device, comprising:

At least one processor; and

A memory communicatively coupled to the at least one processor; wherein,

The memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-11.

24. A non-transitory computer readable storage medium storing computer instructions for causing the computer to perform the method of any one of claims 1-11.

25. A computer program product comprising computer instructions which, when executed by a processor, implement the steps of the method of any one of claims 1 to 11.