An information recommendation method, device and readable storage medium based on a knowledge graph

Abstract

The present application provides a method, device, and readable storage medium for information recommendation based on a knowledge graph. In response to an information search request, the method determines the optimal search path for the termination search node indicated by the information search request in the fully associated knowledge graph. When the historical search traffic of the optimal search path exceeds the traffic tolerance of any alternative search path in the optimal search path, the method determines the alternative search path set used to share the search traffic of the alternative search path through Nash equilibrium, and determines the search access traffic of the alternative search path and the alternative search path set. According to the search access traffic, the method recommends the optimal search path and the recommendation information corresponding to each intermediate search node in the candidate search path set. In this way, the intermediate search nodes related to the termination search node can be expanded to enrich the recommendation information recommended to the user.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a method, device and readable storage medium for information recommendation based on knowledge graphs. Background Art

[0002] Existing data asset management products generally use knowledge graphs to store the structural relationships between assets, and can continuously expand and enrich these knowledge graphs through backend maintenance, mining, analysis, and user behavior intervention. To enrich and diversify user choices, when a user has a search request, in addition to displaying the target information, other highly relevant information will also be recommended to the user.

[0003] The construction of knowledge graphs has greatly enriched the search and recommendation functions of data asset management products. However, since the structure in the knowledge graph usually does not change, the recommendation results are relatively fixed. That is, when repeatedly searching for the same target in a short period of time, the recommended content to users is relatively fixed or repeated. Summary of the Invention

[0004] In view of this, the purpose of this application is to provide an information recommendation method, device and readable storage medium based on knowledge graph to solve the problem that the recommended content to users is relatively fixed.

[0005] This embodiment of the present application provides an information recommendation method based on a knowledge graph, the information recommendation method comprising:

[0006] In response to an information search request, determining an optimal search path for obtaining a termination search node indicated by the information search request by searching the fully associated knowledge graph;

[0007] When the historical search traffic of the optimal search path exceeds the traffic tolerance of any search edge in the optimal search path, a Nash equilibrium algorithm is used to determine, from the fully associative knowledge graph, a set of alternative search paths for sharing the search traffic of the optimal search path, as well as the search access traffic of each alternative search path in the set of alternative search paths;

[0008] According to the search access flow, the optimal search path and recommendation information corresponding to each intermediate search node involved in each candidate search path in the candidate search path set are sequentially recommended to the user.

[0009] In a possible implementation, the recommending to the user, in sequence according to the search access traffic, the optimal search path and the recommendation information corresponding to each intermediate search node involved in each alternative search path in the alternative search path set, includes:

[0010] According to the search access traffic, determining a first search path with a first recommendation priority from the optimal search path and the set of alternative search paths;

[0011] Starting from the search node at the last position in the first search path, recommending the recommendation information corresponding to each intermediate search node in the first search path to the user one by one;

[0012] After completing the recommendation of each intermediate search node in the first search path, determining a second search path with a second recommendation priority according to the search access traffic;

[0013] Starting from the search node at the last position in the second search path, recommending recommendation information corresponding to each intermediate search node that is not recommended in the second search path one by one;

[0014] This process is deduced in this way until the recommendation information corresponding to each intermediate search node in each search path is recommended.

[0015] In one possible implementation, constructing the fully associative knowledge graph includes:

[0016] Obtain a pre-built initial knowledge graph;

[0017] Connecting two adjacent edge nodes in the initial knowledge graph to obtain a fully associative knowledge graph to be configured, so that there is a direct or indirect connection relationship between all nodes in the fully associative knowledge graph;

[0018] In response to the traffic parameter configuration operation, at least one congestion parameter involved in the fully associative knowledge graph to be configured and the search weight of each search edge are configured to obtain a constructed fully associative knowledge graph.

[0019] In one possible implementation, in response to the information search request, determining the optimal search path for obtaining the termination search node indicated by the information search request by searching in the fully associated knowledge graph includes:

[0020] In response to an information search request, determining a start search node and an end search node for the information search from the information search request;

[0021] By using the Nash equilibrium algorithm and combining it with the search weight of each search edge in the fully associative knowledge graph, the optimal search path starting from the starting search node and searching for the ending search node is determined; wherein the search weight of each search edge is positively correlated with the degree of association between the two search nodes connected by the search edge.

[0022] In one possible implementation, a Nash equilibrium algorithm is used to determine, from the fully associative knowledge graph, a set of alternative search paths for sharing the search traffic of the optimal search path, and the search access traffic of each alternative search path in the set of alternative search paths, including:

[0023] By utilizing the Nash equilibrium algorithm and combining the traffic tolerance and search weight of each search edge, the historical search traffic is distributed during an iterative calculation process to determine the set of alternative search paths and the search access traffic of each alternative search path in the set of alternative search paths.

[0024] In a possible implementation, the information recommendation method further includes:

[0025] In response to an information selection operation, determining a target search node selected by a user;

[0026] The search flow of the path nodes on the target sub-path for obtaining the target search node by searching the candidate search path set is adjusted.

[0027] In a possible implementation, the information recommendation method further includes:

[0028] In response to a traffic parameter adjustment operation, at least one congestion parameter involved in the fully associative knowledge graph and / or a traffic tolerance on a target search edge selected by the traffic parameter adjustment operation are adjusted.

[0029] The present application also provides an information recommendation device based on a knowledge graph, the information recommendation device comprising:

[0030] An optimal path determination module, configured to determine, in response to an information search request, an optimal search path for obtaining a termination search node indicated by the information search request by searching the fully associated knowledge graph;

[0031] a traffic distribution module, configured to, when the historical search traffic of the optimal search path exceeds the traffic tolerance of any search edge in the optimal search path, use a Nash equilibrium algorithm to determine, from the fully associative knowledge graph, a set of alternative search paths for sharing the search traffic of the optimal search path, and the search access traffic of each alternative search path in the set of alternative search paths;

[0032] The information recommendation module is used to recommend the optimal search path and the recommendation information corresponding to each intermediate search node involved in each alternative search path in the alternative search path set to the user in sequence according to the search access traffic.

[0033] In one embodiment, when the information recommendation module is used to recommend the optimal search path and the recommended information corresponding to each intermediate search node involved in each alternative search path in the alternative search path set to the user in sequence according to the search access traffic, the information recommendation module is used to:

[0034] According to the search access traffic, determining a first search path with a first recommendation priority from the optimal search path and the set of alternative search paths;

[0035] Starting from the search node at the last position in the first search path, recommending the recommendation information corresponding to each intermediate search node in the first search path to the user one by one;

[0036] After completing the recommendation of each intermediate search node in the first search path, determining a second search path with a second recommendation priority according to the search access traffic;

[0037] Starting from the search node at the last position in the second search path, recommending recommendation information corresponding to each intermediate search node that is not recommended in the second search path one by one;

[0038] This process is deduced in this way until the recommendation information corresponding to each intermediate search node in each search path is recommended.

[0039] In a possible implementation, the information recommendation device further includes a graph construction module, which is configured to construct the fully-associative knowledge graph through the following steps:

[0040] Obtain a pre-built initial knowledge graph;

[0041] Connecting two adjacent edge nodes in the initial knowledge graph to obtain a fully associative knowledge graph to be configured, so that there is a direct or indirect connection relationship between all nodes in the fully associative knowledge graph;

[0042] In response to the traffic parameter configuration operation, at least one congestion parameter involved in the fully associative knowledge graph to be configured and the search weight of each search edge are configured to obtain a constructed fully associative knowledge graph.

[0043] In a possible implementation, when the optimal path determination module is used to determine, in response to an information search request, the optimal search path for obtaining the termination search node indicated by the information search request by searching in the fully associated knowledge graph, the optimal path determination module is used to:

[0044] In response to an information search request, determining a start search node and an end search node for the information search from the information search request;

[0045] By using the Nash equilibrium algorithm and combining it with the search weight of each search edge in the fully associative knowledge graph, the optimal search path starting from the starting search node and searching for the ending search node is determined; wherein the search weight of each search edge is positively correlated with the degree of association between the two search nodes connected by the search edge.

[0046] In one possible implementation, the traffic distribution module is configured to use a Nash equilibrium algorithm to determine, from the fully associative knowledge graph, a set of alternative search paths for sharing the search traffic of the optimal search path, and the search access traffic of each alternative search path in the set of alternative search paths. The traffic distribution module is configured to:

[0047] By utilizing the Nash equilibrium algorithm and combining the traffic tolerance and search weight of each search edge, the historical search traffic is distributed during an iterative calculation process to determine the set of alternative search paths and the search access traffic of each alternative search path in the set of alternative search paths.

[0048] In a possible implementation, the information recommendation device further includes a traffic tolerance adjustment module, and the traffic tolerance adjustment module is configured to:

[0049] In response to an information selection operation, determining a target search node selected by a user;

[0050] The search flow of the path nodes on the target sub-path for obtaining the target search node by searching the candidate search path set is adjusted.

[0051] In a possible implementation, the information recommendation device further includes a traffic parameter adjustment module, and the traffic parameter adjustment module is configured to:

[0052] In response to a traffic parameter adjustment operation, at least one congestion parameter involved in the fully associative knowledge graph and / or a traffic tolerance on a target search edge selected by the traffic parameter adjustment operation are adjusted.

[0053] An embodiment of the present application also provides an electronic device, comprising: a processor, a memory, and a bus, wherein the memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor and the memory communicate through the bus, and when the machine-readable instructions are executed by the processor, the steps of the above-mentioned knowledge graph-based information recommendation method are performed.

[0054] An embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned knowledge graph-based information recommendation method are executed.

[0055] The information recommendation method, device and readable storage medium based on the knowledge graph provided by the embodiment of the present application, in response to an information search request, determine the optimal search path for the termination search node indicated by the information search request in the fully associated knowledge graph; when the historical search traffic of the optimal search path exceeds the traffic tolerance of any search edge in the optimal search path, the Nash equilibrium algorithm is used to determine the set of alternative search paths for sharing the search traffic of the optimal search path, as well as the search access traffic of each alternative search path in the alternative search path set from the fully associated knowledge graph; according to the search access traffic, the recommended information corresponding to each intermediate search node involved in each alternative search path in the alternative search path set is recommended to the user in turn. In this way, under the restriction of traffic tolerance, by determining the candidate alternative search paths that can share the search traffic of the alternative search paths, the intermediate search nodes related to the termination search node requested by the user can be expanded, and the recommended information corresponding to the intermediate search nodes can be recommended to the user; at the same time, the phenomenon of delayed search result feedback caused by searching using the optimal search path can be avoided.

[0056] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without creative work.

[0058] Figure 1 A flowchart of a knowledge graph-based information recommendation method provided in an embodiment of the present application;

[0059] Figure 2 A schematic diagram of the structure of a fully associative knowledge graph provided in an embodiment of the present application;

[0060] Figure 3 This is one of the structural diagrams of an information recommendation device based on a knowledge graph provided in an embodiment of the present application;

[0061] Figure 4 This is a second structural diagram of an information recommendation device based on a knowledge graph provided in an embodiment of the present application;

[0062] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0063] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application for which protection is claimed, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, each other embodiment obtained by those skilled in the art without making creative work falls within the scope of protection of the present application.

[0064] Research has found that the construction of knowledge graphs has greatly enriched the search and recommendation functions of data asset management products. However, since the structure in the knowledge graph usually does not change, the recommendation results are relatively fixed. That is, when repeatedly searching for the same target in a short period of time, the recommended content recommended to users is relatively fixed or repeated.

[0065] Based on this, an embodiment of the present application provides an information recommendation method based on a knowledge graph, which can, under the limitation of traffic tolerance, expand the intermediate search nodes related to the terminal search node requested by the user by determining the candidate alternative search paths that can share the search traffic of the optimal search path, and recommend the recommended information corresponding to the intermediate search nodes to the user.

[0066] See also Figure 1 , Figure 1 This is a flow chart of an information recommendation method based on knowledge graph provided in an embodiment of the present application. Figure 1 As shown in , the information recommendation method based on knowledge graph provided by the embodiment of the present application includes:

[0067] S101. In response to an information search request, determine an optimal search path for obtaining a termination search node indicated by the information search request by searching in a fully associated knowledge graph.

[0068] S102. When the historical search traffic of the optimal search path exceeds the traffic tolerance of any search edge in the optimal search path, the Nash equilibrium algorithm is used to determine from the fully associative knowledge graph a set of alternative search paths for sharing the search traffic of the optimal search path and the search access traffic of each alternative search path in the alternative search path set.

[0069] S103: Recommend the optimal search path and the recommendation information corresponding to each intermediate search node involved in each alternative search path in the alternative search path set to the user in sequence according to the search access traffic.

[0070] An information recommendation method based on a knowledge graph provided in an embodiment of the present application determines the optimal search path for the terminal search node indicated by the information search request in the fully associated knowledge graph when a user has an information search request. When the historical search traffic of the optimal search path exceeds the traffic tolerance of any alternative search path in the optimal search path, the alternative search path is selected to share the search traffic of any alternative search path to avoid the phenomenon of delayed search result feedback when searching using the optimal search path. At the same time, the candidate alternative search paths can be further used to expand the intermediate search nodes related to the terminal search node requested by the user, and recommend the recommended information corresponding to the intermediate search nodes to the user, thereby achieving the purpose of enriching the recommended information.

[0071] In step S101, in response to a received information search request, a search is performed in a pre-built fully associative knowledge graph to determine an optimal search path in the fully associative knowledge graph to obtain a termination search node indicated by the information search request.

[0072] Among them, the optimal search path refers to the search path that can search for the terminating search node fastest; therefore, the optimal search path for searching the terminating search node can be determined from multiple search paths that can search for the terminating search node with reference to the length of the search path.

[0073] In one embodiment, step S101 includes:

[0074] S1011. In response to an information search request, determine a start search node and an end search node for the information search from the information search request.

[0075] In this step, in response to receiving the information search request, the information search request is identified to determine the starting search node and the ending search node of this information search request; wherein, the starting search node is the node that starts this information search; the ending search node refers to the node that ends this search request, that is, the search target of this search request.

[0076] For example, taking the information search request as a route search, the starting search node is the user's departure point; the ending search node is the user's destination.

[0077] S1012. Utilize the Nash equilibrium algorithm, combined with the search weight of each search edge in the fully associative knowledge graph, to determine the optimal search path starting from the starting search node and searching for the ending search node.

[0078] In this step, in order to search for the terminating search node as quickly as possible, the Nash equilibrium algorithm can be used, combined with the search weight of each search edge in the fully associative knowledge graph, to determine the optimal search path from the starting search node to the terminating search node.

[0079] Here, the search weight of each search edge is pre-set, and the search weight of each search edge is positively correlated with the degree of association between the two search nodes connected by the search edge.

[0080] Among them, the Nash equilibrium algorithm may be a linear constraint solving algorithm such as the Frank-Wolf algorithm.

[0081] Here, in order to avoid the problem of high search traffic on the path, which may cause overload of the search node and delay in search result feedback, the search traffic in the alternative search path can be controlled by configuring a traffic tolerance for each alternative search path to avoid the problem of "congestion" in the search path.

[0082] Here, the knowledge graph constructed under normal circumstances is a "non-closed" graph, that is, there are unconnected edge nodes in the knowledge graph. In this case, we cannot guarantee that each alternative search path can find an alternative search path that shares the search traffic. Therefore, before searching for alternative search paths, the initial knowledge graph needs to be improved to ensure that there are multiple reachable search paths for each terminating search node.

[0083] In one embodiment, the information recommendation method also includes: obtaining a pre-constructed initial knowledge graph; connecting two adjacent edge nodes in the initial knowledge graph to obtain a fully associative knowledge graph to be configured, so that there is a direct or indirect connection relationship between all nodes in the fully associative knowledge graph; in response to a traffic parameter configuration operation, configuring at least one congestion parameter involved in the fully associative knowledge graph to be configured and the search weight of each search edge to obtain a constructed fully associative knowledge graph.

[0084] In this step, for the pre-built initial knowledge graph, adjacent and unconnected edge nodes in the initial knowledge graph are connected to obtain a fully-connected knowledge graph to be configured with traffic parameters to be configured, so that all nodes in the fully-connected knowledge graph have direct or indirect connections; that is, any two edge nodes in the knowledge graph are reachable, and a fully-connected knowledge graph is constructed;

[0085] Moreover, while constructing the fully associative knowledge graph to be configured, the user can configure the various traffic parameters involved in the fully associative knowledge graph to be configured through the traffic parameter configuration operation; specifically, in response to the traffic parameter configuration operation applied by the user on the traffic parameter, at least one congestion parameter (α and / or β) involved in the fully associative knowledge graph to be configured and the search weight of each search edge are configured to obtain the constructed fully associative knowledge graph.

[0086] In addition, the fully associative knowledge graph constructed by initialization can be made so that the initial value of the historical search traffic of each search node in the fully associative knowledge graph is 0, and the search weights of each search edge in the fully associative knowledge graph are made consistent.

[0087] See also Figure 2 , Figure 2 This is a schematic diagram of the structure of a fully associative knowledge graph provided in the embodiment of this application. Figure 2 As shown, a pre-built initial knowledge graph is obtained, and for the adjacent edge nodes 2a-1, edge node 2a-2, edge node 2a-3 and edge node 2a-4 that are not connected in the initial knowledge graph, two adjacent edge nodes are connected, that is, edge node 2a-1 and edge node 2a-2, edge node 2a-2 and edge node 2a-3, and edge node 2a-3 and edge node 2a-4 are connected to obtain a fully associated knowledge graph 2b.

[0088] In step S102, after the optimal search path is determined, the historical search traffic of the optimal search path is obtained; here, the historical search traffic can represent the number of visits to the optimal search path. The greater the number of visits, the greater the possibility of "congestion". At this time, it is necessary to compare the historical search traffic with the traffic tolerance of each alternative search path in the optimal search path to determine whether "congestion" will occur in the optimal search path.

[0089] Specifically, the historical search traffic of the optimal search path is compared with the traffic tolerance of each search edge in the optimal search path one by one. When the historical search traffic of the optimal search path exceeds the traffic tolerance of any search edge in the optimal search path, it means that there is an overloaded search edge in the optimal search path. At this time, it is necessary to adopt a "diversion" method to disperse the search traffic of any search edge to other alternative search paths that can search for the termination search node; at this time, the Nash equilibrium algorithm can be used to search the fully associated knowledge graph again to determine other alternative search paths that can search for the termination search node, so as to obtain a set of alternative search paths in combination.

[0090] For example, Figure 2As shown in the fully associative knowledge graph 2b, taking the end search node as edge node 2a-2 as an example, the optimal search path of edge node 2a-2 is "Aa-2". If the search edge "a-2" has a load overload problem, it is necessary to determine other search edges that share the search traffic of the search edge "a-2". Figure 2 It can be seen that in the fully associative knowledge graph 2b, the search edge "a-1-2" can also search the terminal search node (i.e., edge node 2a-2). Therefore, the alternative search path "Aa-1-2" can be determined as the alternative search path in the alternative search path set.

[0091] At the same time, the Nash equilibrium algorithm is used to determine the search access flow of each alternative search path by allocating the historical search flow of the optimal search path on the partial search edges involved in each alternative search path.

[0092] In one embodiment, step S102 includes: using the Nash equilibrium algorithm, combining the traffic tolerance and search weight of each search edge, and in an iterative calculation process, distributing the historical search traffic to determine the alternative search path set and the search access traffic of each alternative search path in the alternative search path set.

[0093] In this step, the difference between the optimal search path and each alternative search path in the alternative search path set is that the parts of the search edges involved in the process of searching for the termination search node are different; for different search edges, it is necessary to determine the search access traffic of different search edges on each alternative search path by allocating the historical search traffic of the optimal search path; and for the search edges shared by the two, the shared search edges can be considered as the search edges whose historical search traffic of the optimal search path does not exceed the traffic tolerance, indicating that the shared search edges are not overloaded, and the historical search traffic of the optimal search path can be determined as the search access traffic of the shared search edges; further, by combining the search access traffic of each search edge included in each alternative search path, the search access traffic of the alternative search path is determined.

[0094] Specifically, a linear constraint solving algorithm (for example, the Frank-Wolf algorithm) is used, combined with the search weight and traffic tolerance of each search edge in the fully associative knowledge graph, to iteratively solve a set of alternative search paths for sharing the search traffic of the optimal search path. By allocating the historical search traffic of the optimal search path, the alternative search path for terminating the search node is determined, and the search access traffic of each alternative search path is determined; in this way, a Nash equilibrium is achieved between the optimal search path and each alternative search path in the alternative search path set.

[0095] Corresponding to the above embodiment, taking the historical search traffic of the optimal search path "Aa-2" as 20 and the traffic tolerance of the search edge "a-2" as 10 as an example, the historical search traffic exceeds the traffic tolerance of the search edge "a-2". Therefore, it is necessary to use the search edge "a-1-2" to share the search traffic on the search edge "a-2". At this time, the Nash equilibrium algorithm is adopted, and with the help of the search weight of the search edge "a-2" and the path weight of the search edge "a-1-2", the historical search traffic is respectively allocated to the search edge "a-2" and the candidate search edge "a-1-2". For example, the search access traffic allocated to the search edge "a-2" is 13, and the search access traffic allocated to the search edge "a-1-2" is 7.

[0096] Here, since the search access traffic can indirectly reflect the degree of correlation between the search path and the terminal search node requested by the user, when recommending information to the user, the search access traffic of each alternative search path can be referred to, and the recommended information corresponding to the intermediate search nodes that are more relevant to the terminal search node can be recommended to the user first.

[0097] Normally, when recommending information to users, only the relevant information of the intermediate search nodes involved in the optimal search path will be recommended to the users. This will also make the recommended content each time limited and relatively fixed. Therefore, further considering the problem of relatively fixed recommended content, in order to enrich the content recommended to users, when recommending information to users, the intermediate search nodes involved in each alternative search path can be recommended to the users at the same time.

[0098] In step S103, according to the search access traffic of each alternative search path and the search access traffic of the optimal search path, the recommendation information corresponding to each intermediate search node involved in any alternative search path and the optimal search path is recommended to the user in turn, thereby enriching the information recommended to the user.

[0099] Here, considering that the intermediate search nodes with higher correlation with the terminal search node requested by the user should be recommended to the user first, therefore, before the recommendation, the recommendation order of the optimal search path and the candidate search path needs to be sorted according to the search access traffic, and the recommendation information corresponding to the intermediate search nodes involved in the search path is recommended to the user in the recommended order.

[0100] In one embodiment, step S103 includes:

[0101] S1031: Determine a first search path with a first recommendation priority from the optimal search path and the candidate search path set according to the search access traffic.

[0102] In this step, the optimal search path and each alternative search path in the alternative search path set are sorted with reference to the historical search traffic of the optimal search path and the search access traffic of each alternative search path, so as to determine the first search path with the first recommendation order from the optimal search path and the alternative search paths.

[0103] S1032: Starting from the search node at the last position in the first search path, recommend the recommendation information corresponding to each intermediate search node in the first search path to the user one by one.

[0104] In this step, starting from the search node at the last position in the first search path, and moving forward in reverse order, the recommended information corresponding to each intermediate search node in the first search path is recommended to the user one by one.

[0105] Corresponding to the above embodiment, taking the optimal search path "Aa-2" as the first search path as an example, first, the recommendation information corresponding to the terminal search node "2" requested by the user is fed back to the user; secondly, the recommendation information corresponding to the intermediate search node "a" is recommended to the user; finally, the recommendation information corresponding to the intermediate search node "A" is recommended to the user.

[0106] S1033: After completing the recommendation of each intermediate search node in the first search path, determine a second search path with a second recommendation priority according to the search access traffic.

[0107] In this step, after completing the recommendation of the recommendation information corresponding to each intermediate search node in the first search path, refer again to the search access traffic of any alternative search path and the search access traffic of each candidate alternative search path to determine the second search path at the second recommendation order from the optimal search path and the candidate search paths.

[0108] S1034: Starting from the search node at the last position in the second search path, recommend recommendation information corresponding to each intermediate search node that is not recommended in the second search path one by one.

[0109] In this step, starting from the search node at the last position in the second search path, the recommended information corresponding to each intermediate search node in the second search path that has not been recommended to the user is recommended to the user one by one in reverse order.

[0110] Corresponding to the above embodiment, taking the candidate search path "Aa-1-2" as the second search path as an example, after the recommendation of the terminal search node (ie, search node "2") has been completed, the recommendation information corresponding to the intermediate search node "1" is recommended to the user.

[0111] S1035 , and so on, until the recommendation information corresponding to each intermediate search node in each search path is recommended.

[0112] After the user completes the selection of the search node, in order to be able to timely update the search data of the fully associated knowledge graph according to the user's selection, at this time, it is necessary to adjust the search traffic involving the search node according to the user's selection.

[0113] In one embodiment, the information recommendation method further includes: determining a target search node selected by a user in response to an information selection operation; and adjusting search traffic of path nodes on a target sub-path of the target search node obtained by searching the target search node in the alternative search path set.

[0114] In this step, after the user completes the selection of recommended information corresponding to the search node through the information selection operation, the search flow of each search node in the fully associated knowledge graph can be adjusted according to the actual selection situation; specifically, in response to the information selection operation applied by the user, the target search node selected by the user through the information selection operation is determined;

[0115] Determine the target subpath for searching the target search node in the set of alternative search paths, and adjust the search traffic of the pathway nodes on the target subpath to achieve real-time update of the search data of the relevant search nodes in the fully associative knowledge graph.

[0116] Here, the administrator of the fully associative knowledge graph can also adjust the traffic tolerance of each alternative search path during operation.

[0117] In one embodiment, the information recommendation method further includes:

[0118] In response to a traffic parameter adjustment operation, at least one congestion parameter involved in the fully associative knowledge graph and / or a traffic tolerance on a target search edge selected by the traffic parameter adjustment operation are adjusted.

[0119] In this step, in response to the traffic parameter adjustment operation imposed by the administrator, the target search edge selected by the administrator through the process parameter adjustment operation is determined to achieve the adjustment of the traffic tolerance of the target search edge. In this way, the historical search traffic can be distributed to more search paths. By increasing the number of search paths involved, the number of intermediate search nodes recommended to users is enriched, and the diversity of recommendation results in the information recommendation process is indirectly improved.

[0120] During the adjustment process, the administrator can adjust the flow tolerance of the alternative search path selected by the operation by inputting parameters; the administrator can also adjust the flow tolerance of the alternative search path selected by the operation by selecting parameters.

[0121] At the same time, the administrator can also adjust the parameters through traffic parameters to adjust at least one congestion parameter (α and / or β) involved in the fully associated knowledge graph.

[0122] The information recommendation method based on the knowledge graph provided by the embodiment of the present application responds to an information search request and determines the optimal search path for the termination search node indicated by the information search request in the fully associated knowledge graph; when the historical search traffic of the optimal search path exceeds the traffic tolerance of any search edge in the optimal search path, the Nash equilibrium algorithm is used to determine the set of alternative search paths for sharing the search traffic of the optimal search path and the search access traffic of each alternative search path in the alternative search path set from the fully associated knowledge graph; according to the search access traffic, the recommended information corresponding to each intermediate search node involved in each alternative search path in the optimal search path set and each alternative search path in the alternative search path set is recommended to the user in turn. In this way, under the restriction of traffic tolerance, by determining the candidate alternative search paths that can share the search traffic of the alternative search paths, the intermediate search nodes related to the termination search node requested by the user can be expanded, and the recommended information corresponding to the intermediate search nodes can be recommended to the user. At the same time, it can also avoid the phenomenon of delayed search result feedback caused by searching using the optimal search path.

[0123] See also Figure 3 、 Figure 4 , Figure 3 This is one of the structural diagrams of an information recommendation device based on a knowledge graph provided in an embodiment of the present application. Figure 4 This is a second structural diagram of an information recommendation device based on knowledge graph provided in an embodiment of the present application. Figure 3 As shown in , the information recommendation device 300 includes:

[0124] An optimal path determination module 310 is configured to determine, in response to an information search request, an optimal search path for obtaining a termination search node indicated by the information search request by searching the fully associated knowledge graph;

[0125] The traffic distribution module 320 is configured to, when the historical search traffic of the optimal search path exceeds the traffic tolerance of any search edge in the optimal search path, use a Nash equilibrium algorithm to determine, from the fully associative knowledge graph, a set of alternative search paths for sharing the search traffic of the optimal search path, as well as the search access traffic of each alternative search path in the set of alternative search paths;

[0126] The information recommendation module 330 is configured to recommend to the user, in accordance with the search access traffic, the optimal search path and the recommended information corresponding to each intermediate search node involved in each candidate search path in the candidate search path set.

[0127] In one embodiment, when the information recommendation module 330 is used to recommend the optimal search path and the recommended information corresponding to each intermediate search node involved in each alternative search path in the alternative search path set to the user in sequence according to the search access traffic, the information recommendation module 330 is used to:

[0128] According to the search access traffic, determining a first search path with a first recommendation priority from the optimal search path and the set of alternative search paths;

[0129] Starting from the search node at the last position in the first search path, recommending the recommendation information corresponding to each intermediate search node in the first search path to the user one by one;

[0130] After completing the recommendation of each intermediate search node in the first search path, determining a second search path with a second recommendation priority according to the search access traffic;

[0131] Starting from the search node at the last position in the second search path, recommending recommendation information corresponding to each intermediate search node that is not recommended in the second search path one by one;

[0132] This process is deduced in this way until the recommendation information corresponding to each intermediate search node in each search path is recommended.

[0133] Furthermore, the information recommendation device 300 further includes a graph construction module 340, which is configured to construct the fully-associated knowledge graph through the following steps:

[0134] Obtain a pre-built initial knowledge graph;

[0135] Connecting two adjacent edge nodes in the initial knowledge graph to obtain a fully associative knowledge graph to be configured, so that there is a direct or indirect connection relationship between all nodes in the fully associative knowledge graph;

[0136] In response to the traffic parameter configuration operation, at least one congestion parameter involved in the fully associative knowledge graph to be configured and the search weight of each search edge are configured to obtain a constructed fully associative knowledge graph.

[0137] Furthermore, when the optimal path determination module 310 is used to determine, in response to an information search request, an optimal search path for obtaining a termination search node indicated by the information search request in the fully associated knowledge graph, the optimal path determination module 310 is used to:

[0138] In response to an information search request, determining a start search node and an end search node for the information search from the information search request;

[0139] By using the Nash equilibrium algorithm and combining it with the search weight of each search edge in the fully associative knowledge graph, the optimal search path starting from the starting search node and searching for the ending search node is determined; wherein the search weight of each search edge is positively correlated with the degree of association between the two search nodes connected by the search edge.

[0140] Furthermore, the traffic distribution module 320 is configured to use a Nash equilibrium algorithm to determine, from the fully associative knowledge graph, a set of alternative search paths for sharing the search traffic of the optimal search path, and the search access traffic of each alternative search path in the set of alternative search paths. The traffic distribution module 320 is configured to:

[0141] By utilizing the Nash equilibrium algorithm and combining the traffic tolerance and search weight of each search edge, the historical search traffic is distributed during an iterative calculation process to determine the set of alternative search paths and the search access traffic of each alternative search path in the set of alternative search paths.

[0142] Furthermore, the information recommendation device 300 further includes a flow tolerance adjustment module 350, and the flow tolerance adjustment module 350 is configured to:

[0143] In response to an information selection operation, determining a target search node selected by a user;

[0144] The search flow of the path nodes on the target sub-path for obtaining the target search node by searching the candidate search path set is adjusted.

[0145] Furthermore, the information recommendation device 300 further includes a traffic parameter adjustment module 360, which is configured to:

[0146] In response to a traffic parameter adjustment operation, at least one congestion parameter involved in the fully associative knowledge graph and / or a traffic tolerance on a target search edge selected by the traffic parameter adjustment operation are adjusted.

[0147] The information recommendation device based on the knowledge graph provided by the embodiment of the present application responds to an information search request and determines the optimal search path for the termination search node indicated by the information search request by searching in the fully-associated knowledge graph; when the historical search traffic of the optimal search path exceeds the traffic tolerance of any search edge in the optimal search path, a set of alternative search paths for sharing the search traffic of the optimal search path and the search access traffic of each alternative search path in the alternative search path set are determined from the fully-associated knowledge graph using a Nash equilibrium algorithm; according to the search access traffic, the recommended information corresponding to each intermediate search node involved in the optimal search path and each alternative search path in the alternative search path set is recommended to the user in turn. In this way, by determining the candidate alternative search paths that can share the search traffic of the alternative search paths under the restriction of the traffic tolerance, the intermediate search nodes related to the termination search node requested by the user can be expanded, and the recommended information corresponding to the intermediate search nodes can be recommended to the user. At the same time, the phenomenon of delayed search result feedback caused by searching using the optimal search path can be avoided.

[0148] See also Figure 5 , Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 5 As shown in FIG, the electronic device 500 includes a processor 510, a memory 520 and a bus 530.

[0149] The memory 520 stores machine-readable instructions executable by the processor 510. When the electronic device 500 is running, the processor 510 communicates with the memory 520 via the bus 530. When the machine-readable instructions are executed by the processor 510, the above-mentioned Figure 1 The steps of the knowledge graph-based information recommendation method in the method embodiment shown are specifically implemented in accordance with the method embodiment and will not be described in detail here.

[0150] The embodiment of the present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the computer program can execute the above-mentioned Figure 1 The steps of the knowledge graph-based information recommendation method in the method embodiment shown are specifically implemented in accordance with the method embodiment and will not be described in detail here.

[0151] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0152] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. There may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed may be through some communication interface, indirect coupling or communication connection of devices or units, which may be electrical, mechanical or other forms.

[0153] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0154] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0155] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0156] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The scope of protection of the present application is not limited thereto. Although the present application has been described in detail with reference to the above-mentioned embodiments, those skilled in the art should understand that any person skilled in the art can modify or easily conceive of changes to the technical solutions described in the above-mentioned embodiments within the technical scope disclosed in the present application, or perform equivalent replacements for some of the technical features thereof. These modifications, changes, or replacements do not deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. An information recommendation method based on knowledge graph, characterized in that: The information recommendation method includes: In response to an information search request, determining an optimal search path for obtaining a termination search node indicated by the information search request by searching the fully associated knowledge graph; When the historical search traffic of the optimal search path exceeds the traffic tolerance of any search edge in the optimal search path, a Nash equilibrium algorithm is used to determine, from the fully associative knowledge graph, a set of alternative search paths for sharing the search traffic of the optimal search path, as well as the search access traffic of each alternative search path in the set of alternative search paths; According to the search access traffic, the recommendation order of the optimal search path and the alternative search paths is sorted, and the recommendation information corresponding to each intermediate search node involved in each alternative search path in the optimal search path and the alternative search path set is recommended to the user in the recommendation order.

2. The information recommendation method according to claim 1, characterized in that The step of sorting the optimal search path and the alternative search paths in a recommended order according to the search access traffic, and recommending to the user, in order of recommendation, the optimal search path and the recommendation information corresponding to each intermediate search node involved in each alternative search path in the set of alternative search paths, including: According to the search access traffic, determining a first search path with a first recommendation priority from the optimal search path and the set of alternative search paths; Starting from the search node at the last position in the first search path, recommending the recommendation information corresponding to each intermediate search node in the first search path to the user one by one; After completing the recommendation of each intermediate search node in the first search path, determining a second search path with a second recommendation priority according to the search access traffic; Starting from the search node at the last position in the second search path, recommending recommendation information corresponding to each intermediate search node that is not recommended in the second search path one by one; This process is deduced in this way until the recommendation information corresponding to each intermediate search node in each search path is recommended.

3. The information recommendation method according to claim 1, wherein: The fully associative knowledge graph is constructed by the following steps, including: Obtain a pre-built initial knowledge graph; Connecting two adjacent edge nodes in the initial knowledge graph to obtain a fully associative knowledge graph to be configured, so that there is a direct or indirect connection relationship between all nodes in the fully associative knowledge graph; In response to the traffic parameter configuration operation, at least one congestion parameter involved in the fully associative knowledge graph to be configured and the search weight of each search edge are configured to obtain a constructed fully associative knowledge graph.

4. The information recommendation method according to claim 1, characterized in that The step of determining, in response to the information search request, an optimal search path for obtaining a termination search node indicated by the information search request by searching in the fully associated knowledge graph comprises: In response to an information search request, determining a start search node and an end search node for the information search from the information search request; By using the Nash equilibrium algorithm and combining it with the search weight of each search edge in the fully associative knowledge graph, the optimal search path starting from the starting search node and searching for the ending search node is determined; wherein the search weight of each search edge is positively correlated with the degree of association between the two search nodes connected by the search edge.

5. The information recommendation method according to claim 1, characterized in that: Using a Nash equilibrium algorithm, determining a set of alternative search paths for sharing the search traffic of the optimal search path from the fully associative knowledge graph, and the search access traffic of each alternative search path in the set of alternative search paths, including: By utilizing the Nash equilibrium algorithm and combining the traffic tolerance and search weight of each search edge, the historical search traffic is distributed during an iterative calculation process to determine the set of alternative search paths and the search access traffic of each alternative search path in the set of alternative search paths.

6. The information recommendation method according to claim 1, characterized in that: The information recommendation method further includes: In response to an information selection operation, determining a target search node selected by a user; The search flow of the path nodes on the target sub-path for obtaining the target search node by searching the candidate search path set is adjusted.

7. The information recommendation method according to claim 1, characterized in that: The information recommendation method further includes: In response to a traffic parameter adjustment operation, at least one congestion parameter involved in the fully associative knowledge graph and / or a traffic tolerance on a target search edge selected by the traffic parameter adjustment operation are adjusted.

8. An information recommendation device based on knowledge graph, characterized in that: The information recommendation device includes: An optimal path determination module, configured to determine, in response to an information search request, an optimal search path for obtaining a termination search node indicated by the information search request by searching the fully associated knowledge graph; a traffic distribution module, configured to, when the historical search traffic of the optimal search path exceeds the traffic tolerance of any search edge in the optimal search path, use a Nash equilibrium algorithm to determine, from the fully associative knowledge graph, a set of alternative search paths for sharing the search traffic of the optimal search path, and the search access traffic of each alternative search path in the set of alternative search paths; An information recommendation module is used to sort the recommended order of the optimal search path and the alternative search path according to the search access traffic, and recommend to the user the recommended information corresponding to each intermediate search node involved in the optimal search path and each alternative search path in the alternative search path set in the recommended order.

9. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor and the memory communicate through the bus, and the machine-readable instructions are executed by the processor to execute the steps of the information recommendation method based on the knowledge graph as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which, when executed by a processor, executes the steps of the information recommendation method based on a knowledge graph as described in any one of claims 1 to 7.

Images