Recommendation information generation method, device, electronic device and readable storage medium

Abstract

The embodiment of the present application provides a method, device, electronic device and readable storage medium for generating recommendation information. The method includes: according to the correlation between the user identifier, the object identifier and the context attribute parameter, determining multiple paths including a first node including the context attribute parameter, a second node including the user identifier, and a third node including the object identifier, grouping the multiple paths according to the sub-parameter values ​​and the third node included in the first node on the multiple paths to obtain multiple groups of paths, generating recommendation information of the object to be recommended according to the path attribute values ​​of each of the multiple groups of paths, using the method of the embodiment of the present application, the process of generating the information to be recommended combines the context attribute parameters and path attribute values ​​in the historical data of the field to which the recommendation information belongs, so that the recommendation information for the object to be recommended is richer and more referenceable, which can bring more surprises and attraction to users.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of data processing technology, and in particular to a method, device, electronic device and readable storage medium for generating recommendation information. Background Art

[0002] With the development of Internet technology, users choose items online more frequently. Usually, the system will recommend an item to the user to help him choose. When recommending an item, the system will also give the reason for recommending the item, i.e., recommendation information, in order to enhance the user's trust in the recommended item and improve the user experience. Existing recommendation information generation methods can be roughly divided into the following categories:

[0003] Based on user reviews. Mining high-quality review text from a large number of user reviews of the item, and then recommending it to the user. For example, "Tom Yum Goong soup is good and tasty."

[0004] Based on social relationships. Based on the current user's social network and behavior, the current user's social network's preference for the item is mined. For example, "6 of your friends like this store."

[0005] Based on user similarity. Recommendations are made based on the preferences of other users who are similar to the current user. For example, "You might like this store because users similar to you like this store."

[0006] Based on item similarity. Recommendations are made based on the item set that the current user prefers and the similarity of the current item. For example, "You may like this store because you have been to similar stores many times before."

[0007] Among them, recommendation information based on user comments is more direct and more common. The comments of other users have a greater reference value for the current user's decision-making, and users tend to comprehensively compare the comments of multiple users. However, in the recommendation reason scenario, a single user comment often has limited information, and the sense of surprise and highlights are not prominent enough. The problem with these three types of recommendation information based on social relationships, user similarity, and item similarity is that the descriptions are too simple and their frequent appearance will make users feel tired.

[0008] Therefore, the recommendation information generation method in the prior art needs to be improved. Summary of the invention

[0009] The embodiments of the present application provide a method, device, electronic device and readable storage medium for generating recommendation information, aiming to provide users with rich, reference-worthy and attractive recommendation information.

[0010] A first aspect of an embodiment of the present application provides a method for generating recommendation information, the method comprising:

[0011] According to the correlation relationship between the user identifier, the object identifier and the context attribute parameter, a plurality of paths including a first node including the context attribute parameter, a second node including the user identifier, and a third node including the object identifier are determined;

[0012] According to each sub-parameter value included in the first node on the multiple paths and the third node, the multiple paths are grouped to obtain multiple groups of paths, wherein each sub-path included in a group of paths corresponds to the same sub-parameter value and the same object identifier;

[0013] According to the path attribute values ​​of each of the plurality of groups of paths, recommendation information of the object to be recommended is generated.

[0014] Optionally, the path attribute value includes a traffic parameter value; the traffic parameter value of any group of paths in the multiple groups of paths is determined according to the following steps:

[0015] Determine the flow parameter value of each sub-path included in the group of paths, wherein the flow parameter value of a sub-path is determined according to the flow transfer value between every two nodes on the path, wherein the flow transfer value between two nodes is determined according to the out-degree of the first node of the two nodes;

[0016] The flow parameter value of the group of paths is determined according to the flow parameter value of each sub-path included in the group of paths.

[0017] Optionally, the path attribute value includes a preference; the preference of any group of paths in the plurality of groups of paths is determined according to the following steps:

[0018] Determining the preference of the group of paths according to the flow parameter values ​​of the multiple groups of paths corresponding to the same object identifier as the group of paths, the number of the multiple groups of paths corresponding to the same object identifier as the group of paths, and the flow parameter values ​​of the group of paths; or

[0019] The preference of the group of paths is determined according to the number of user identifiers on the group of paths and the total number of user identifiers on the plurality of groups of paths corresponding to the same object identifier as the group of paths.

[0020] Optionally, generating recommendation information of the object to be recommended according to the path attribute values ​​of each of the plurality of groups of paths includes:

[0021] Determine a target group of paths whose path attribute values ​​are greater than a preset threshold from the multiple groups of paths corresponding to the same object identifier;

[0022] Determine the object represented by the same object identifier corresponding to each sub-path in the target group path as the object to be recommended;

[0023] The recommendation information of the object to be recommended is generated according to the same sub-parameter value corresponding to each sub-path in the target group path.

[0024] Optionally, a single path among the multiple paths further includes an intermediate node of other parameters; generating recommendation information of the object to be recommended according to the same sub-parameter value corresponding to each sub-path in the target group path includes:

[0025] Determine a target node label according to a node label of at least one intermediate node on each sub-path in the target group path;

[0026] Generate recommendation information carrying labels for the object to be recommended, where the labels carried by the recommendation information are the target node label and / or the object label of the object to be recommended.

[0027] Optionally, according to the correlation relationship between the user identifier, the object identifier and the context attribute parameter, a plurality of paths including a first node including the context attribute parameter, a second node including the user identifier, and a third node including the object identifier are determined, including:

[0028] Determine a first node, a second node, and a third node from a preset knowledge graph;

[0029] According to the first node, the second node and the third node, the preset knowledge graph is traversed to obtain the multiple paths.

[0030] Optionally, the association relationship between the user identifier, the object identifier and the context attribute parameter is determined according to the following steps:

[0031] Determine a preset operation relationship between the user identifier and the object identifier according to the user's historical behavior data, wherein the preset operation is at least one of an order placement operation, a browsing operation, and a click operation;

[0032] Determine a first correspondence between a user identifier and a context attribute parameter according to the user attribute information, or determine a second correspondence between an object identifier and a context attribute parameter according to the object attribute information;

[0033] The preset operation relationship and the first corresponding relationship form the association relationship, or the preset operation relationship and the second corresponding relationship form the association relationship;

[0034] The context attribute parameters include at least one of the following: a location parameter, a time parameter, and a tag parameter.

[0035] Optionally, according to the correlation relationship between the user identifier, the object identifier and the context attribute parameter, a plurality of paths including a first node including the context attribute parameter, a second node including the user identifier, and a third node including the object identifier are determined, including:

[0036] According to a preset tag, determining a set of identifiers that satisfy the preset tag, where the identifiers in the set of identifiers are user identifiers or object identifiers;

[0037] According to the association relationship, determining a plurality of paths including a first node including a context attribute parameter, a second node including a user identifier, and a third node including an object identifier;

[0038] Generating recommendation information of the object to be recommended according to the path attribute values ​​of each of the plurality of groups of paths, including:

[0039] According to the path attribute values ​​of each of the multiple groups of paths, recommendation information carrying the preset tag of the object to be recommended is generated.

[0040] Optionally, after generating the recommendation information of the object to be recommended, the method further includes:

[0041] receiving an object recommendation request sent by a terminal;

[0042] The object to be recommended and the recommendation information of the object to be recommended are returned to the terminal.

[0043] A second aspect of an embodiment of the present application provides a recommendation information generating device, the device comprising:

[0044] A first determination module, configured to determine, according to the correlation relationship between the user identifier, the object identifier and the context attribute parameters, a plurality of paths including a first node including the context attribute parameters, a second node including the user identifier, and a third node including the object identifier;

[0045] A grouping module, configured to group the plurality of paths according to the respective sub-parameter values ​​included in the first nodes on the plurality of paths and the third node, to obtain a plurality of groups of paths, wherein the respective sub-paths included in a group of paths correspond to the same sub-parameter value and the same object identifier;

[0046] The generating module is used to generate the recommendation information of the object to be recommended according to the path attribute values ​​of each of the plurality of groups of paths.

[0047] Optionally, the path attribute value includes a traffic parameter value; and the generating module includes:

[0048] A first determination submodule is used to determine the flow parameter value of each subpath included in the group of paths, wherein the flow parameter value of a subpath is determined according to the flow transfer value between every two nodes on the path, wherein the flow transfer value between two nodes is determined according to the out-degree of the first node of the two nodes;

[0049] The second determining submodule is used to determine the flow parameter value of the group of paths according to the flow parameter value of each sub-path included in the group of paths.

[0050] Optionally, the path attribute value includes a preference; and the generating module includes:

[0051] A third determination submodule, configured to determine the preference of the group of paths according to the flow parameter values ​​of the plurality of groups of paths corresponding to the same object identifier as the group of paths, the number of the plurality of groups of paths corresponding to the same object identifier as the group of paths, and the flow parameter values ​​of the group of paths;

[0052] The fourth determination submodule is used to determine the preference of the group of paths according to the number of user identifiers on the group of paths and the total number of user identifiers on the multiple groups of paths corresponding to the same object identifier as the group of paths.

[0053] Optionally, the generating module includes:

[0054] A fifth determination submodule, configured to determine, from the plurality of groups of paths corresponding to the same object identifier, a target group of paths whose path attribute values ​​are greater than a preset threshold;

[0055] A sixth determination submodule, configured to determine objects represented by the same object identifier corresponding to each subpath in the target group path as objects to be recommended;

[0056] The first generating submodule is used to generate the recommendation information of the to-be-recommended object according to the same sub-parameter value corresponding to each sub-path in the target group path.

[0057] Optionally, a single path among the multiple paths further includes intermediate nodes of other parameters; and the first generating submodule includes:

[0058] A determination subunit, configured to determine a target node label according to a node label of at least one intermediate node on each subpath in the target group path;

[0059] The generating subunit is used to generate recommendation information carrying labels for the object to be recommended, where the labels carried by the recommendation information are the target node label and / or the object label of the object to be recommended.

[0060] Optionally, the first determining module includes:

[0061] A seventh determination submodule is used to determine a first node, a second node, and a third node from a preset knowledge graph;

[0062] The acquisition submodule is used to traverse the preset knowledge graph according to the first node, the second node and the third node to obtain the multiple paths.

[0063] Optionally, the first determining module includes:

[0064] An eighth determination submodule, used to determine a preset operation relationship between the user identifier and the object identifier according to the user historical behavior data, wherein the preset operation is at least one of an order placement operation, a browsing operation, and a click operation;

[0065] A ninth determination submodule, configured to determine a first correspondence between a user identifier and a context attribute parameter according to the user attribute information, or to determine a second correspondence between an object identifier and a context attribute parameter according to the object attribute information;

[0066] Wherein, the preset operation relationship and the first corresponding relationship form the association relationship, or the preset operation relationship and the second corresponding relationship form the association relationship;

[0067] The context attribute parameters include at least one of the following: a location parameter, a time parameter, and a tag parameter.

[0068] Optionally, the first determining module includes:

[0069] A tenth determination submodule, configured to determine, according to a preset tag, a set of identifiers satisfying the preset tag, wherein the identifiers in the set of identifiers are user identifiers or object identifiers;

[0070] an eleventh determination module, configured to determine, according to the association relationship, a plurality of paths including a first node including a context attribute parameter, a second node including a user identifier, and a third node including an object identifier;

[0071] The generating module comprises:

[0072] The second generating submodule is used to generate recommendation information carrying the preset tag of the object to be recommended according to the path attribute values ​​of each of the multiple groups of paths.

[0073] Optionally, the device further comprises:

[0074] A receiving module, used for receiving an object recommendation request sent by a terminal;

[0075] The returning module is used to return the object to be recommended and the recommendation information of the object to be recommended to the terminal.

[0076] A third aspect of an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the method described in the first aspect of the present application.

[0077] A fourth aspect of an embodiment of the present application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executed, implements the steps of the method described in the first aspect of the present application.

[0078] By adopting the recommendation information generation method provided in the embodiment of the present application, the process of generating the information to be recommended combines the contextual attribute parameters and path attribute values ​​in the historical data of the field to which the recommendation information belongs, fully explores the various characteristics of the object to be recommended, and generates the recommendation information using the various characteristics of the object to be recommended, so that the recommendation information for the object to be recommended is richer and more referenceable, which can bring more surprises and appeal to users. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments of the present application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0080] Figure 1 is a flow chart of a method for generating recommendation information proposed in an embodiment of the present application;

[0081] Figure 2 It is a schematic diagram of the path between "situation attribute parameter-user identification-object identification";

[0082] Figure 3 is a flow chart of a method for generating recommendation information proposed in an embodiment of the present application;

[0083] Figure 4 is a schematic diagram of the effect of recommendation information in an embodiment of the present application;

[0084] Figure 5 It is a schematic diagram of a recommendation information generating device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0085] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0086] refer to Figure 1 , Figure 1 : is a flow chart of a method for generating recommendation information proposed in an embodiment of the present application. Figure 1 As shown, the method comprises the following steps:

[0087] Step S11, according to the correlation relationship among the user identifier, the object identifier and the context attribute parameters, a plurality of paths including a first node including the context attribute parameters, a second node including the user identifier and a third node including the object identifier are determined.

[0088] In this embodiment, the user identifier is used to uniquely identify a user, wherein the user is a user existing in the historical data of the field to which the recommended information belongs. The object identifier is used to uniquely identify an object, wherein the object is an object existing in the historical data of the field to which the recommended information belongs. The object may specifically be a merchant, a commodity, a store, etc. The context attribute parameters may include location parameters, time parameters, and label parameters. The location parameters may specifically be parameters that can represent a location, such as a province, Shanghai, Sichuan, a commercial street, or a subway entrance. The time parameters may specifically be parameters that can represent a time node, such as the most recent month or the most recent week. The label parameters may specifically be parameters that can represent a feature label, such as 5-star, new store, store with an average price of 50-100 yuan, etc.

[0089] For example, refer to Figure 2 , Figure 2 It is a schematic diagram of the path between "situation attribute parameter-user ID-object ID". Figure 2 In, the context attribute parameters are Shanghai and Sichuan, etc., the user identifiers are user134, user232, user479 and user645, etc., and the object identifiers are shop113, shop200 (shop200 is not shown in the figure) and shop245, etc. Among them, the association relationship between user134 and user232 and Shanghai can be hometown, that is, the hometown of the user identified by user134 and the user identified by user232 is Shanghai, and the association relationship between user479 and user645 and Sichuan can also be hometown. Similarly, the hometown of the user identified by user479 and the user identified by user645 is Sichuan. The association relationship between user134 and shop113 and shop200 is ordering, that is, the user identified by user134 has placed orders in the store identified by shop113 and the store identified by shop200. Similarly, the user identified by user479 and the user identified by user645 have placed orders in the store identified by shop245. At this time, refer to Figure 2, the following multiple paths can be determined: Shanghai-user134-shop113, Shanghai-user134-shop200 (shop200 is not shown in the figure), Shanghai-user232-shop200, Sichuan-user479-shop200, Sichuan-user479-shop245, Sichuan-user645-shop245.

[0090] It should be noted that the first node, the second node and the third node in this embodiment do not represent the order of node arrangement, that is, the first node including the situational attribute parameters can be the starting node, the intermediate node or the ending node. Similarly, the second node including the user identifier can be the starting node, the intermediate node or the ending node. The third node including the object identifier can be the starting node, the intermediate node or the ending node.

[0091] Figure 2 In the illustrated embodiment, multiple paths determined by context attribute parameters as the starting node, user identification as the intermediate node and object identification as the terminating node are exemplified. In an optional implementation, multiple paths determined by object identification as the starting node, user identification as the intermediate node and context attribute parameters as the terminating node may also be used, or multiple paths determined by user identification as the starting node, object identification as the intermediate node and context attribute parameters as the terminating node. In the present application, the arrangement relationship among user identification, object identification and context attribute parameters is not specifically limited.

[0092] In one implementation, the association relationship between the user identifier, the object identifier, and the context attribute parameters may be determined by:

[0093] According to the user's historical behavior data, a preset operation relationship between the user identifier and the object identifier is determined, and the preset operation is at least one of an ordering operation, a browsing operation, and a clicking operation.

[0094] According to the user attribute information, a first corresponding relationship between the user identifier and the context attribute parameter is determined, or, according to the object attribute information, a second corresponding relationship between the object identifier and the context attribute parameter is determined. The preset operation relationship and the first corresponding relationship form the association relationship, or the preset operation relationship and the second corresponding relationship form the association relationship. In this embodiment, after a user performs operations such as placing an order, browsing, and clicking on a certain object, the historical behavior data of each user on each object is recorded. In this way, the preset operation relationship between the user identifier and the object identifier can be obtained. For example, the user identified by user479 has placed an order in the store identified by the object shop245, and the association relationship between user479 and shop245 is placing an order.

[0095] In addition, each user usually registers their identity and fills in their attribute information. For users who do not fill in their attribute information, their attribute information is mined through user profiling technology. For example, the attribute information filled in by the user identified by user479 is that their registered permanent residence is Sichuan. In this way, the first corresponding relationship between the user identification and the context attribute parameter can be obtained, that is, the association relationship between user479 and Sichuan is their hometown.

[0096] Similarly, each object is usually registered first, and the object attribute information is filled in; or the attribute information of the object is mined. For example, the attribute information filled in for the shop identified by shop113 is that the cuisine is Shanghai cuisine and the business district is Wangjing. In this way, the second correspondence between the object identifier and the context attribute parameters can be obtained, that is, the association between shop113 and Shanghai cuisine is the cuisine, and the association between shop113 and Wangjing is the business district. In addition, each object can also be marked with attribute information by the user. For example, after other users place an order at the shop identified by shop113, they can mark the attribute information for shop113 according to the characteristics of shop113, such as the price range of 50-100 yuan. In this way, the second correspondence between the object identifier and the context attribute parameters can also be obtained, that is, the association between shop113 and 50-100 yuan is the price range.

[0097] In practical applications, a domain knowledge graph including class, entity, relation, etc. can be constructed through historical data of the domain to which the recommended information belongs. Through the knowledge graph, user identification, object identification, context attribute parameters and the relationship between the three can be obtained quickly, intuitively and accurately. At this time, when determining multiple paths including a first node including context attribute parameters, a second node including a user identification, and a third node including an object identification based on the relationship between the user identification, object identification and context attribute parameters, the following steps can be included:

[0098] Step S111, determining a first node, a second node, and a third node from a preset knowledge graph;

[0099] Step S112: traverse the preset knowledge graph according to the first node, the second node, and the third node to obtain the multiple paths.

[0100] In this embodiment, in a preset knowledge graph, that is, in a domain knowledge graph including type (class), entity (entity), relation (relation), etc., constructed by historical data of the domain to which the recommended information belongs, a first node including context attribute parameters, a second node including a user identifier, and a third node including an object identifier can constitute a meta-path, corresponding to the path composed of types as nodes and relationships as edges in the meta-path in the knowledge graph, for example, province-[hometown]-user-[order]-merchant, is a meta-path. After determining the meta-path, multiple paths can be obtained by traversing the knowledge graph.

[0101] It should be noted that the association between user identification, object identification and context attribute parameters is not limited to the preset operation relationship between user identification and object identification, the first correspondence between user identification and context attribute parameters, and the second correspondence between object identification and context attribute parameters, but includes all relationships and attributes in the knowledge graph.

[0102] In one implementation, there is already a ready-made knowledge graph for recommendation information in certain fields, so the ready-made knowledge graph can be directly used, making it more convenient to obtain user identification, object identification and context attribute parameters and the relationship between the three.

[0103] Step S12: grouping the multiple paths according to the sub-parameter values ​​and the third nodes included in the first nodes on the multiple paths to obtain multiple groups of paths.

[0104] Each sub-path included in a group of paths corresponds to the same sub-parameter value and the same object identifier.

[0105] In this embodiment, the sub-parameter value refers to the next level parameter included in the context attribute parameter. For example, when the context attribute parameter is a location parameter, the sub-parameter may be parameters such as Shanghai, Sichuan, etc. When the context attribute parameter is a time parameter, the sub-parameter may be parameters such as the last month or the last week. When the context attribute parameter is a label parameter, the sub-parameter may be parameters such as 5-star, new store, store with an average price of 50-100 yuan, etc.

[0106] For example, continue to refer to Figure 2 , Sichuan-user479-shop245 and Sichuan-user645-shop245 have the same sub-parameters and the same object identifier, so Sichuan-user479-shop245 and Sichuan-user645-shop245 are the same group of paths.

[0107] It should be understood that Figure 2The path diagram shown only shows a small part of the path data. In the remaining unshown part of the path data, there may also be paths such as Shanghai-user134-shop200 (shop200 is not shown in the figure) and Shanghai-user232-shop200. Similarly, multiple paths are grouped according to the above grouping method. If there are paths of Shanghai-user134-shop200 and Shanghai-user232-shop200, then Shanghai-user134-shop200 and Shanghai-user232-shop200 are also in the same group of paths.

[0108] Step S13: generating recommendation information of the object to be recommended according to the path attribute values ​​of each of the plurality of groups of paths.

[0109] In this embodiment, the path attribute value may include two types: traffic parameter value and preference degree.

[0110] When the path attribute value is a traffic parameter value, the traffic parameter values ​​corresponding to the multiple groups of paths can be determined by the following steps:

[0111] Step S131, determining the flow parameter value of each sub-path included in the group of paths, the flow parameter value of a sub-path is determined based on the flow transfer value between every two nodes on the path, wherein the flow transfer value between two nodes is determined based on the out-degree of the first node of the two nodes.

[0112] Step S132: determining the flow parameter value of the group of paths according to the flow parameter value of each sub-path included in the group of paths.

[0113] In this embodiment, the out-degree refers to the number of edges from this node to the next type of node, see Figure 2 , the out-degree of node Sichuan is 2, the out-degree of node user479 is 2, and the out-degree of node user645 is 1.

[0114] Each sub-path in the group of paths has the context attribute parameter as the starting node, the user identifier as the intermediate node, and the object identifier as the ending node. Therefore, when calculating the flow transfer value between the starting node and the intermediate node, the starting node is the first node of the intermediate node, so the calculation is based on the out-degree of the starting node. When calculating the flow transfer value between the intermediate node and the ending node, the intermediate node is the first node of the ending node, so the calculation is based on the out-degree of the intermediate node. For example, the flow parameter value of each sub-path in the group of paths is equal to the product of the flow transfer values ​​between the two nodes in the sub-path, and the flow transfer value between the two nodes is equal to the inverse of the out-degree of the first node of the two nodes, that is, 1 / the out-degree of the first node.

[0115] Therefore, the flow parameter value of each sub-path included in the group of paths can be calculated first, and then the flow parameter values ​​of each sub-path can be added together to determine the flow parameter value of the group of paths. Figure 2 , taking the calculation of the flow parameter value of the path group Sichuan-user479-shop245 and Sichuan-user645-shop245 as an example, first calculate the flow parameter values ​​of the sub-path Sichuan-user479-shop245 and the sub-path Sichuan-user645-shop245 included in the path group. The flow parameter value of the sub-path Sichuan-user479-shop245 is 1 / 2*1 / 2=1 / 4, the flow parameter value of the sub-path Sichuan-user645-shop245 is 1 / 2*1=1 / 2, and the flow parameter value of the path group is 1 / 4+1 / 2=3 / 4.

[0116] When the path attribute value is a preference, the preferences corresponding to the multiple groups of paths can be determined by the following steps:

[0117] Step S133, determining the preference of the group of paths according to the respective traffic parameter values ​​of the multiple groups of paths corresponding to the same object identifier as the group of paths, the number of the multiple groups of paths corresponding to the same object identifier as the group of paths, and the traffic parameter values ​​of the group of paths; or determining the preference of the group of paths according to the number of user identifiers on the group of paths and the total number of user identifiers on the multiple groups of paths corresponding to the same object identifier as the group of paths.

[0118] In this embodiment, there are two specific calculation methods for preference calculation. One is based on the calculation of the flow parameter value of the group of paths and the flow parameter values ​​of the plurality of groups of paths corresponding to the same object identifier as the group of paths. Specifically, the preference for the object identifier to be recommended can be calculated according to the following formula:

[0119] p k =(D k -avg(D1+D2+…+D n )) / avg(D1+D2+…+D n )

[0120] Among them, p k represents the preference of the Kth group of paths after the flow parameter values ​​of the group of paths and the flow parameter values ​​of the multiple groups of paths corresponding to the same object identifier as the group of paths are arranged in reverse order, D kIt represents the flow parameter value of the Kth group of paths after the flow parameter values ​​of the group of paths and the flow parameter values ​​of the multiple groups of paths corresponding to the same object identifier as the group of paths are arranged in reverse order, and n represents the sum of the number of groups of paths and the multiple groups of paths corresponding to the same object identifier as the group of paths. A positive preference indicates a positive preference, and a negative preference indicates a negative preference.

[0121] The idea of ​​this method is a relative calculation method, that is, calculation is performed in proportion, which is equivalent to each position parameter having one vote, and the vote of each position parameter is split into multiple votes. If the position parameter corresponds to two users, each user obtains 1 / 2 of the vote of the position parameter, which is the inverse of the out-degree of the position parameter. Similarly, each user has one vote, and the vote of each user is split into multiple votes. If the user has been to three stores, each store obtains 1 / 3 of the user's vote, which is the inverse of the user's out-degree.

[0122] For example, assume that there are three groups of paths with the sub-parameter of the location parameter as the starting node, user as the intermediate node to the terminal node shop113, namely along Shanghai-user-shop113, the flow parameter value is 0.3, along Sichuan-user-shop113, the flow parameter value is 0.1, and along Guizhou-user-shop113, the flow parameter value is 0.08.

[0123] Then shop113's preference for Shanghai is:

[0124] (0.3-avg(0.3+0.1+0.08)) / avg(0.3+0.1+0.08)=0.875.

[0125] Then shop113's preference for Sichuan is:

[0126] (0.1-avg(0.3+0.1+0.08)) / avg(0.3+0.1+0.08)=-0.375.

[0127] Then shop113's preference for Guizhou is:

[0128] (0.08-avg(0.3+0.1+0.08)) / avg(0.3+0.1+0.08)=-0.5.

[0129] The other is to directly determine the preference of the group of paths according to the number of user identifiers on the group of paths and the total number of user identifiers on the multiple groups of paths corresponding to the same object identifier as the group of paths. Specifically, the preference for the identifier of the object to be recommended can be calculated according to the following formula:

[0130] p k =(D k-avg(D1+D2+…+D n )) / avg(D1+D2+…+D n )

[0131] Among them, p k represents the preference of the kth group of paths, D k represents the number of user identifiers of the kth group of paths, avg(D1+D2+…+D n ) represents the average number of user IDs of all group paths.

[0132] The idea of ​​this method is an absolute calculation method, that is, calculation according to the actual number, which is equivalent to each position parameter corresponding to several users, and the position parameter casts one vote for each of these users. If the position parameter corresponds to 70 users, these 70 users each get 1 vote for the position parameter. Similarly, each user corresponds to several stores, and the user casts one vote for each of these stores. If the user corresponds to 10 stores, each store gets 1 vote from the user.

[0133] For example, assume that there are three groups of paths with the sub-parameter of the position parameter as the starting node, user as the intermediate node to the terminal node shop113, namely along Shanghai-user-shop113, including 70 user identifiers, along Sichuan-user-shop113, including 40 user identifiers (i.e., the intermediate node user), and along Guizhou-user-shop113, including 10 user identifiers.

[0134] Then shop113’s preference for Shanghai is: (70-avg(70+40+10)) / avg(70+40+10)=0.75.

[0135] Then shop113's preference for Sichuan is: (40--avg(70+40+10)) / -avg(70+40+10)=0.

[0136] Then shop113's preference for Guizhou is: (10--avg(70+40+10)) / -avg(70+40+10)=-0.75.

[0137] In this embodiment, each group of paths has a corresponding path attribute value. After multiple paths are grouped to obtain multiple groups of paths, the embodiment of the present invention can further generate recommendation information of the object to be recommended according to the path attribute values ​​of each of the multiple groups of paths.

[0138] In a preferred embodiment of the present invention, step S13 may further include the following steps:

[0139] Step S134: determining a target group of paths whose path attribute values ​​are greater than a preset threshold from the multiple groups of paths corresponding to the same object identifier.

[0140] Step S135: determine the objects represented by the same object identifier corresponding to each sub-path in the target group path as objects to be recommended.

[0141] Step S136: Generate recommendation information of the object to be recommended according to the same sub-parameter value corresponding to each sub-path in the target group path.

[0142] In this embodiment, in order to facilitate the accurate determination of the object to be recommended and the recommendation information of the object to be recommended, it is possible to select a target group path to be determined from the multiple groups of paths corresponding to the same object identifier. After determining the target group path, the object represented by the same object identifier corresponding to each sub-path in the target group path can be determined as the object to be recommended. At the same time, based on the same sub-parameter value corresponding to each sub-path in the target group path, the recommendation information of the object to be recommended is generated.

[0143] For example, there are three groups of paths, namely Shanghai-user-shop113, Guizhou-user-shop113 and Sichuan-user-shop113. The selected target group path is Shanghai-user-shop113. Since each sub-path included in a group of paths corresponds to the same sub-parameter value and the same object identifier, each sub-path included in the target group path Shanghai-user-shop113 corresponds to the same sub-parameter value Shanghai and the same object identifier shop113. At this time, the object to be recommended can be shop113, and the recommendation information for shop113 can be the store that Shanghai people like to shop most.

[0144] In this embodiment, the target group path can be selected by setting a preset threshold, so that the number of selected target group paths can be controlled. For example, in some cases, the multiple groups of paths corresponding to the same object identifier can be arranged in reverse order according to the path attribute value, and the fourth path attribute value is selected as the preset threshold. In this way, when generating recommendation information for the object to be recommended, three groups of target group paths can be obtained, namely Shanghai-user-shop113, Guizhou-user-shop113 and Sichuan-user-shop113. Since the sub-paths of these three target group paths are for the same object identifier, three recommendation information can be generated for the same object to be recommended. For example, for shop113, recommendation information can be generated, one of the stores that Shanghai people like to consume the most, one of the stores that Sichuan people like to consume the most, and one of the stores that Guizhou people like to consume the most.

[0145] In one implementation, for example, the multiple groups of paths corresponding to the same object identifier can also be arranged in reverse order according to the path attribute value, and the second path attribute value is selected as the preset threshold. In this way, when generating the recommendation information of the object to be recommended, a set of target group paths can be obtained, that is, the path with the largest path attribute value, such as Shanghai-user-shop113. At this time, the object to be recommended can be shop113, and the recommendation information for shop113 can be the store that Shanghai people love to consume the most.

[0146] It should be noted that the path attribute value in this embodiment can be selected solely by the traffic parameter value, or the preference, or by the traffic parameter value and the preference at the same time. Selecting the traffic parameter value alone or the preference can reduce the processing steps and improve the efficiency of generating recommended information, but the accuracy is relatively low. Although selecting the traffic parameter value and the preference at the same time reduces the efficiency of generating recommended information, it can improve the accuracy of the recommended information. Therefore, technical personnel in this field can select the parameters included in the path attribute value according to actual needs.

[0147] In the embodiment of the present invention, the process of generating the information to be recommended combines the context attribute parameters and path attribute values ​​in the historical data of the field to which the recommended information belongs, so that the recommended information for the object to be recommended is richer and more referenceable, which can bring more surprises and appeal to users.

[0148] In order to further enrich the content of the recommendation information, as an optional implementation, a single path in the multiple paths also includes intermediate nodes of other parameters. In this case, when generating the recommendation information of the to-be-recommended object according to the same sub-parameter value corresponding to each sub-path in the target group path, the following steps may be included:

[0149] Step S136-1, determining a target node label according to a node label of at least one intermediate node on each sub-path in the target group path.

[0150] Step S136 - 2 , generating recommendation information carrying labels for the object to be recommended, where the labels carried in the recommendation information are the target node label and / or the object label of the object to be recommended.

[0151] In this embodiment, on the one hand, the content of the recommendation information can be enriched by adding other intermediate nodes. For example, the intermediate node can include not only the user identification but also the taste node. The recommendation information generated in this way can include taste. For example, Shanghai people like to consume hot and sour dishes in this restaurant the most.

[0152] On the other hand, the content of the recommended information can be enriched by directly adding the target node label and / or the object label of the object to be recommended to the recommended information. If the intermediate node of the user identifier has a node label, the node label of the user identifier can be directly used. If other intermediate nodes, such as the taste node, have a node label, the node label of the intermediate node of the taste can also be directly used. Still taking the recommended information as: Shanghai people like to consume hot and sour dishes in this store as an example, the corresponding target group path is Shanghai-user-sour and spicy-shop113. At this time, the intermediate nodes are user and sour and spicy, wherein the intermediate node user has a node label white-collar crowd, the intermediate node sour and spicy has a node label hot dishes, and the object to be recommended shop113 has an object label new store. At this time, the target node label and / or the object label of the object to be recommended can be directly added to the recommended information, that is, the generated recommended information is: Shanghai white-collar people like to consume hot and sour dishes in a new store. It can be seen that through the method of this embodiment, the content of the recommended information is further enriched, which can bring more surprises and attractions to users. At the same time, this embodiment directly uses the node labels of the intermediate nodes and the object labels of the objects to be recommended, which is simple to operate.

[0153] In order to further enrich the content of the recommendation information, as another optional implementation, when determining multiple paths including a first node including the context attribute parameter, a second node including the user identifier, and a third node including the object identifier according to the association relationship between the user identifier, the object identifier, and the context attribute parameter, the following steps may be included:

[0154] Step S113: determining, according to a preset tag, a set of identifiers satisfying the preset tag, where the identifiers in the set of identifiers are user identifiers or object identifiers.

[0155] Step S114: determining, according to the association relationship, a plurality of paths including a first node including a context attribute parameter, a second node including a user identifier, and a third node including an object identifier.

[0156] At this time, when generating recommendation information of the to-be-recommended object according to the path attribute values ​​of each of the plurality of groups of paths, the following steps may be included:

[0157] Step S137: Generate recommendation information of the object to be recommended that carries the preset tag according to the path attribute values ​​of each of the multiple groups of paths.

[0158] In this embodiment, the preset label refers to a label predetermined according to the attributes and type of the user or object. Each user identifier or object identifier has its own attribute or type. For example, for user identifiers, user134, user135, user136, and user137 have attributes after 90, user140 and user141 have attributes after 80, and user150 and user151 have attributes after 70. In this case, the preset label can be after 90, so that a set of identifiers that meet the label after 90 can be obtained, namely user134, user135, user136, and user137. The user identifiers in the identifier set are a subset of all user identifiers.

[0159] Similarly, for object identifiers, shop113, shop114, shop115 and shop116 have the attribute Sichuan restaurant, shop121 and shop122 have the attribute 5-star rated store, and shop131 and shop132 have the attribute 10-year-old store. In this case, the preset label can be Sichuan restaurant, so the identifier set that satisfies the label Sichuan restaurant can be obtained, namely shop113, shop114, shop115 and shop116. The object identifiers in the identifier set are subsets of all object identifiers.

[0160] At this time, multiple paths including a first node including a context attribute parameter, a second node including a user identifier, and a third node including an object identifier can be determined according to the association relationship. For example, if Shanghai, user134, and shop113 have an association relationship, the path Shanghai-user134-shop113 can be obtained. For example, if Shanghai, user135, and shop113 have an association relationship, Shanghai-user135-shop113 can be obtained. After obtaining multiple paths, the grouping method of step S12 can be referred to to group the multiple paths to obtain multiple groups of paths, and then the method of steps S131-step S136 can be referred to to generate the recommended object and the recommendation information of the recommended object according to the path attribute values ​​of each of the multiple groups of paths. The difference is that the recommendation information in this embodiment carries a preset label. For example, when the identification set includes the user identification, the recommendation information generated in this embodiment is the store that Shanghai post-90s users like to consume the most. When the identification set includes the object identification, the recommendation information generated in this embodiment is the 10-year-old store that Shanghai users like to consume the most.

[0161] In this embodiment, by determining the preset tags to limit the user identifier or the object identifier, richer recommendation information can be obtained, which further brings more surprises and attractions to the user.

[0162] refer to Figure 3 , Figure 3: is a flow chart of a method for generating recommendation information proposed in an embodiment of the present application. Figure 3 As shown, the method further comprises the following steps:

[0163] Step S14: receiving an object recommendation request sent by a terminal.

[0164] Step S15: returning the object to be recommended and the recommendation information of the object to be recommended to the terminal.

[0165] In this embodiment, the terminal can be a terminal used by the user. For example, after the user logs in to the e-commerce platform page using the terminal, he can send an object recommendation request to the server. After the server receives the object recommendation request, it can return the object to be recommended and the recommendation information of the object to be recommended to the terminal.

[0166] refer to Figure 4 , Figure 4 It is a schematic diagram of the effect of recommendation information in one embodiment of the present application. Figure 4 In the example, after logging into the e-commerce page, the user clicks on the food column. The request to click on the food column can be regarded as an object recommendation request. Then the server will return various stores under the food column to the terminal according to the preset rules. Among them, the returned store "Kuan Bandeng Lao Zao Hot Pot" carries the recommendation information generated by the embodiment of the present application "near Palm Springs Life Plaza, a store that (petty bourgeoisie) users love to shop for", and the returned store "Iron Pot One-Bedroom" carries the recommendation information generated by the embodiment of the present application "Chaoyang Park / Tuanjie Lake (working people) are interested in noodle restaurants", and the returned store "Yoshinoya" carries the recommendation information based on user reviews "a cup of Coke, a bowl of rice, a portion of beef, a reasonable combination..." Obviously, the recommendation information of "Kuan Bandeng Lao Zao Hot Pot" and "Iron Pot One-Bedroom" can bring more surprises and appeal to users.

[0167] Based on the same inventive concept, an embodiment of the present application provides a device for generating recommendation information. Figure 5 , Figure 5 FIG. 5 is a schematic diagram of a recommendation information generating device 50 provided in an embodiment of the present application. Figure 5 As shown, the device comprises:

[0168] A first determination module 51, configured to determine, according to the correlation between the user identifier, the object identifier and the context attribute parameters, a plurality of paths including a first node including the context attribute parameters, a second node including the user identifier, and a third node including the object identifier;

[0169] A grouping module 52, configured to group the plurality of paths according to the respective sub-parameter values ​​included in the first nodes on the plurality of paths and the third node, to obtain a plurality of groups of paths, wherein the respective sub-paths included in a group of paths correspond to the same sub-parameter value and the same object identifier;

[0170] The generating module 53 is used to generate the recommendation information of the object to be recommended according to the path attribute values ​​of each of the plurality of groups of paths.

[0171] Optionally, the path attribute value includes a traffic parameter value; and the generating module includes:

[0172] A first determination submodule is used to determine the flow parameter value of each subpath included in the group of paths, wherein the flow parameter value of a subpath is determined according to the flow transfer value between every two nodes on the path, wherein the flow transfer value between two nodes is determined according to the out-degree of the first node of the two nodes;

[0173] The second determining submodule is used to determine the flow parameter value of the group of paths according to the flow parameter value of each sub-path included in the group of paths.

[0174] Optionally, the path attribute value includes a preference; and the generating module includes:

[0175] A third determination submodule, configured to determine the preference of the group of paths according to the flow parameter values ​​of the plurality of groups of paths corresponding to the same object identifier as the group of paths, the number of the plurality of groups of paths corresponding to the same object identifier as the group of paths, and the flow parameter values ​​of the group of paths;

[0176] The fourth determination submodule is used to determine the preference of the group of paths according to the number of user identifiers on the group of paths and the total number of user identifiers on the multiple groups of paths corresponding to the same object identifier as the group of paths.

[0177] Optionally, the generating module includes:

[0178] A fifth determination submodule, configured to determine, from the plurality of groups of paths corresponding to the same object identifier, a target group of paths whose path attribute values ​​are greater than a preset threshold;

[0179] A sixth determination submodule, configured to determine objects represented by the same object identifier corresponding to each subpath in the target group path as objects to be recommended;

[0180] The first generating submodule is used to generate the recommendation information of the to-be-recommended object according to the same sub-parameter value corresponding to each sub-path in the target group path.

[0181] Optionally, a single path among the multiple paths further includes intermediate nodes of other parameters; and the first generating submodule includes:

[0182] A determination subunit, configured to determine a target node label according to a node label of at least one intermediate node on each subpath in the target group path;

[0183] The generating subunit is used to generate recommendation information carrying labels for the object to be recommended, where the labels carried by the recommendation information are the target node label and / or the object label of the object to be recommended.

[0184] Optionally, the first determining module includes:

[0185] A seventh determination submodule is used to determine a first node, a second node, and a third node from a preset knowledge graph;

[0186] The acquisition submodule is used to traverse the preset knowledge graph according to the first node, the second node and the third node to obtain the multiple paths.

[0187] Optionally, the first determining module includes:

[0188] An eighth determination submodule, used to determine a preset operation relationship between the user identifier and the object identifier according to the user historical behavior data, wherein the preset operation is at least one of an order placement operation, a browsing operation, and a click operation;

[0189] A ninth determination submodule, configured to determine a first correspondence between a user identifier and a context attribute parameter according to the user attribute information, or to determine a second correspondence between an object identifier and a context attribute parameter according to the object attribute information;

[0190] Wherein, the preset operation relationship and the first corresponding relationship form the association relationship, or the preset operation relationship and the second corresponding relationship form the association relationship;

[0191] The context attribute parameters include at least one of the following: a location parameter, a time parameter, and a tag parameter.

[0192] Optionally, the first determining module includes:

[0193] A tenth determination submodule, configured to determine, according to a preset tag, a set of identifiers satisfying the preset tag, wherein the identifiers in the set of identifiers are user identifiers or object identifiers;

[0194] an eleventh determination module, configured to determine, according to the association relationship, a plurality of paths including a first node including a context attribute parameter, a second node including a user identifier, and a third node including an object identifier;

[0195] The generating module comprises:

[0196] The second generating submodule is used to generate recommendation information carrying the preset tag of the object to be recommended according to the path attribute values ​​of each of the multiple groups of paths.

[0197] Optionally, the device further comprises:

[0198] A receiving module, used for receiving an object recommendation request sent by a terminal;

[0199] The returning module is used to return the object to be recommended and the recommendation information of the object to be recommended to the terminal.

[0200] Based on the same inventive concept, another embodiment of the present application provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps in the method described in any of the above embodiments of the present application are implemented.

[0201] Based on the same inventive concept, another embodiment of the present application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executed, implements the steps of the method described in any of the above embodiments of the present application.

[0202] As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0203] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0204] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, devices, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0205] The embodiments of the present application are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0206] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0207] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0208] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present application.

[0209] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or terminal device including the elements.

[0210] The above is a detailed introduction to a recommendation information generation method, device, electronic device and readable storage medium provided by the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for general technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A method for generating recommendation information, characterized in that: include: According to the correlation relationship between the user identifier, the object identifier and the context attribute parameter, a plurality of paths including a first node including the context attribute parameter, a second node including the user identifier, and a third node including the object identifier are determined; According to each sub-parameter value included in the first node on the multiple paths and the third node, the multiple paths are grouped to obtain multiple groups of paths, wherein each sub-path included in a group of paths corresponds to the same sub-parameter value and the same object identifier; generating recommendation information of the object to be recommended according to the path attribute values ​​of each of the plurality of groups of paths; The path attribute value includes a flow parameter value; the flow parameter value of any group of paths in the plurality of groups of paths is determined according to the following steps: Determine the flow parameter value of each sub-path included in the group of paths, wherein the flow parameter value of a sub-path is determined according to the flow transfer value between every two nodes on the path, wherein the flow transfer value between two nodes is determined according to the out-degree of the first node of the two nodes; The flow parameter value of the group of paths is determined according to the flow parameter value of each sub-path included in the group of paths.

2. The method according to claim 1, characterized in that The path attribute value also includes a preference; the preference of any group of paths in the plurality of groups of paths is determined according to the following steps: Determine the preference of the group of paths according to the flow parameter values ​​of the multiple groups of paths corresponding to the same object identifier as the group of paths, the number of the multiple groups of paths corresponding to the same object identifier as the group of paths, and the flow parameter values ​​of the group of paths; Or the preference of the group of paths is determined according to the number of user identifiers on the group of paths and the total number of user identifiers on the multiple groups of paths corresponding to the same object identifier as the group of paths.

3. The method according to claim 1, characterized in that Generating recommendation information of the object to be recommended according to the path attribute values ​​of each of the plurality of groups of paths, including: Determine a target group of paths whose path attribute values ​​are greater than a preset threshold from the multiple groups of paths corresponding to the same object identifier; Determine the object represented by the same object identifier corresponding to each sub-path in the target group path as the object to be recommended; The recommendation information of the object to be recommended is generated according to the same sub-parameter value corresponding to each sub-path in the target group path.

4. The method according to claim 3, characterized in that A single path among the plurality of paths also includes intermediate nodes of other parameters; Generating recommendation information of the object to be recommended according to the same sub-parameter value corresponding to each sub-path in the target group path includes: Determine a target node label according to a node label of at least one intermediate node on each sub-path in the target group path; Generate recommendation information carrying labels for the object to be recommended, where the labels carried by the recommendation information are the target node label and / or the object label of the object to be recommended.

5. The method according to claim 1, characterized in that According to the correlation relationship between the user identifier, the object identifier and the context attribute parameter, a plurality of paths including a first node including the context attribute parameter, a second node including the user identifier, and a third node including the object identifier are determined, including: Determine a first node, a second node, and a third node from a preset knowledge graph; According to the first node, the second node and the third node, the preset knowledge graph is traversed to obtain the multiple paths.

6. The method according to claim 1, characterized in that The association between the user ID, object ID and context attribute parameters is determined according to the following steps: Determine a preset operation relationship between the user identifier and the object identifier according to the user's historical behavior data, wherein the preset operation is at least one of an order placement operation, a browsing operation, and a click operation; Determine a first correspondence between a user identifier and a context attribute parameter according to the user attribute information, or determine a second correspondence between an object identifier and a context attribute parameter according to the object attribute information; The preset operation relationship and the first corresponding relationship form the association relationship, or the preset operation relationship and the second corresponding relationship form the association relationship; The context attribute parameters include at least one of the following: a location parameter, a time parameter, and a tag parameter.

7. The method according to claim 1, characterized in that According to the correlation relationship between the user identifier, the object identifier and the context attribute parameter, a plurality of paths including a first node including the context attribute parameter, a second node including the user identifier, and a third node including the object identifier are determined, including: According to a preset tag, determining a set of identifiers that satisfy the preset tag, where the identifiers in the set of identifiers are user identifiers or object identifiers; According to the association relationship, determining a plurality of paths including a first node including a context attribute parameter, a second node including a user identifier, and a third node including an object identifier; Generating recommendation information of the object to be recommended according to the path attribute values ​​of each of the plurality of groups of paths, including: According to the path attribute values ​​of each of the multiple groups of paths, recommendation information carrying the preset tag of the object to be recommended is generated.

8. The method according to any one of claims 1 to 7, characterized in that: After generating the recommendation information of the object to be recommended, the method further includes: receiving an object recommendation request sent by a terminal; The object to be recommended and the recommendation information of the object to be recommended are returned to the terminal.

9. A device for generating recommendation information, characterized in that: The device comprises: A first determination module, configured to determine, according to the correlation relationship between the user identifier, the object identifier and the context attribute parameters, a plurality of paths including a first node including the context attribute parameters, a second node including the user identifier, and a third node including the object identifier; A grouping module, configured to group the plurality of paths according to the respective sub-parameter values ​​included in the first nodes on the plurality of paths and the third node, to obtain a plurality of groups of paths, wherein the respective sub-paths included in a group of paths correspond to the same sub-parameter value and the same object identifier; A generating module, used for generating recommendation information of the object to be recommended according to the path attribute values ​​of each of the plurality of groups of paths; The path attribute value includes a flow parameter value; the flow parameter value of any group of paths in the plurality of groups of paths is determined according to the following steps: Determine the flow parameter value of each sub-path included in the group of paths, wherein the flow parameter value of a sub-path is determined according to the flow transfer value between every two nodes on the path, wherein the flow transfer value between two nodes is determined according to the out-degree of the first node of the two nodes; The flow parameter value of the group of paths is determined according to the flow parameter value of each sub-path included in the group of paths.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps in any one of the methods according to claims 1-8 are implemented.

11. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When executed, the processor implements the steps of the method according to any one of claims 1 to 8.

Images