Online social media topic propagation collaborative loop detection method and device

Abstract

The present invention provides a method and device for detecting collaborative loops in online social media topic propagation. The method comprises: steps 1 to 6. The present invention constructs a cascade graph for the message forwarding process of online social media topic propagation, and uses the number of node hops and the number of forwarding as weights to search for the top-K diffusion paths in the deep direction of topic propagation and the top-K diffusion paths in the broad direction of topic propagation; accordingly, using the number of forwarding as weight, extracts the top-N loops in the top-K diffusion paths in the deep direction of topic propagation and the top-N loops in the forwarding number of the loop path and the top-K diffusion paths in the broad direction of topic propagation, and takes the union to obtain the collaborative loop topology structure in the topic propagation process. The present invention can effectively detect the key topological structure of path dependence and node collaborative loop feedback in the process of online social media topic propagation, thereby further assisting in identifying manipulation of public opinion topics.

Description

Technical Field

[0001] The embodiments of the present invention relate to the technical field of online social media data mining, and in particular to a method and device for detecting online social media topic propagation collaborative loops. Background Art

[0002] With the advent of the internet age, the efficiency and convenience of online social media have become the mainstream medium for information dissemination in the internet era. Effective detection of online social media topic propagation loops is of great significance for discovering online social media public opinion manipulation and managing the online public opinion environment. Related technologies have proposed detection based on social robot node accounts, and detection based on similar node forwarding behaviors, joint mentions / forwards / likes, and similar posting behaviors. These detections mainly target node account characteristics and propagation behavior characteristics, and are still unable to effectively detect the loop topology generated by node collaboration during topic propagation. Therefore, the development of an online social media topic propagation collaborative loop detection method and equipment that can effectively overcome the shortcomings of the above-mentioned related technologies has become a technical problem that needs to be solved urgently in the industry. Summary of the Invention

[0003] In response to the above-mentioned problems existing in the prior art, embodiments of the present invention provide a method and device for detecting online social media topic propagation collaborative loops.

[0004] In the first aspect, an embodiment of the present invention provides a method for detecting collaborative loops of topic propagation on online social media, including: step 1: constructing a topic propagation forwarding network cascade graph based on the information forwarding relationship between node accounts in the online social media network under the target topic, and calculating the number of hops and forwarding amount of each node; step 2: starting from each source account, using the number of node hops as a weight, searching for the deep diffusion path of the topic propagation on the cascade graph, and extracting the top-K deep diffusion paths of the propagation path depth; step 3: starting from each source account, using the number of node forwarding as a weight, searching for the broad diffusion path of the topic propagation, and extracting the top-K broad diffusion paths of the propagation path forwarding number; step 4: based on the obtained top-K deep diffusion path of the topic propagation, using the number of node forwarding as a weight, detecting the loop on the path, i.e., the head and tail nodes. Step 5: According to the obtained Top-K broad-directional diffusion path of topic propagation, the number of node forwarding is used as the weight to detect the loop on the path, that is, the path connected by the head and tail nodes, and the Top-N loops in the broad-directional loop path of topic propagation are extracted; Step 6: According to the extracted Top-N loops in the deep-directional loop path of topic propagation and the Top-N loops in the broad-directional loop path, the union of the two is taken as the collaborative loop topology structure of topic propagation; wherein, the deep direction of propagation is the propagation direction with a first predetermined number of node forwarding hops from the source node in the propagation process; the broad direction of propagation is the propagation direction with a second predetermined number of node forwarding hops in the propagation process; K is the first constant given by the upper-layer application; N is the second constant given by the upper-layer application.

[0005] Based on the content of the above method embodiment, the online social media topic propagation collaborative loop detection method provided in the embodiment of the present invention specifically includes the following steps: Step 1.1: Based on the online social media network forwarding relationship dataset of the target topic and the account nodes, a node set V of the cascade graph is formed, and the forwarding relationship between the nodes constitutes a directed edge set E of the cascade graph, thereby constructing a topic propagation cascade graph G(V,E); Step 1.2: Extracting the user node that publishes the original blog post in the cascade graph, and determining that the user node is the source account of the message propagated by the target topic; starting from the source account, the number of hops from the node to the source account includes: obtaining the number of hops forwarded by the node; and calculating the number of forwarding of the node according to the out-degree of the node in the cascade graph.

[0006] Based on the content of the above method embodiment, the online social media topic propagation collaborative loop detection method provided in the embodiment of the present invention, wherein the forwarding number of a node is calculated based on the out-degree of the node in the cascade graph, includes:

[0007] fw(v)=OD(v),v∈V

[0008] Wherein, node v is any node in the node set V, fw(v) is the forwarding number of node v, and OD(v) is the node out-degree.

[0009] Based on the content of the above method embodiment, the online social media topic propagation collaborative loop detection method provided in the embodiment of the present invention specifically includes the following steps: Step 2.1: Starting from any source account, visit this node, and start from the neighboring nodes that have not been visited by the source account in turn, using the number of hops of the neighboring nodes as weights, and perform depth-first search according to the weights from large to small until the search of K paths is completed, or all nodes in the cascade graph that have paths connected to the source account are visited; during the search process, the depth of the path is calculated and saved, that is, the number of hops from the current search path to the source account; Step 2.2: If there is still a source account that has not been visited at this time, select another source account that has not been visited as the starting point and repeat the above search process; Step 2.3: If the number of search paths in step 2.1 reaches K, and the depth of the new search path in step 2.2 is greater than the depth of the search path in step 2.1, then replace it and retain the path with the top-K depth in the current steps 2.1 and 2.2; the specific calculation method is:

[0010] min(p∈P d )=p new ,if|P d |==K

[0011] Among them, K is the first constant given by the upper application, P d is the depth path set currently searched by step 2.1, p new is the new path found in step 2.2, |P d |Get the potential of the deep direction path set, that is, the number of paths; min(p∈P d ) Calculate the path with the minimum weight, i.e., the minimum path depth, in the current depth-direction path set; Step 2.4: If the number of search paths in step 2.1 does not reach K, then the new search path in step 2.2 is added to the search path set. The specific calculation method is:

[0012] P d =P d ∪{p new},if|P d | <K

[0013] Among them, K is the first constant given by the upper application, P d is the depth path set currently searched by step 2.1, p new is the new path searched in step 2.2, and |Pd| is the potential of the deep direction path set, that is, the number of paths.

[0014] Based on the content of the above method embodiment, the online social media topic propagation collaborative loop detection method provided in the embodiment of the present invention specifically includes the following steps: Step 3.1: Starting from a source account, access nodes, starting from the neighboring nodes that have not been visited by the source account in turn, using the forwarding number of the neighboring nodes as weights, and performing a depth-first search according to the weights from large to small until the search of N paths is completed, or all nodes in the cascade graph that have paths connected to the source account are visited; during the search process, the forwarding number of the path is calculated and saved, that is, the sum of the forwarding numbers of each node on the current search path; Step 3.2: If there is still a source account that has not been visited at this time, another source account that has not been visited is selected as the starting point, and the above search process is repeated; Step 3.3: If the number of search paths in step 3.1 reaches K, and the depth of the new search path in step 3.2 is greater than the depth of the search path in step 3.1, then a replacement is performed, and the diffusion path with the top-K forwarding numbers in the current steps 3.1 and 3.2 is retained; the specific calculation method is:

[0015] min(p∈P b )=p new1 ,if|P b |==K

[0016] Among them, K is the first constant given by the upper application, P b is the set of broad-direction paths currently searched in step 3.1, p new1 is the new path found in step 3.2, |P b |Get the potential of the broad path set, that is, the number of paths; min(p∈P b ) Calculate the path with the minimum weight in the current broad direction path set, i.e., the path with the minimum forwarding number; Step 3.4: If the number of search paths in step 3.1 does not reach K, then the new search path in step 3.2 is added to the search path set. The specific calculation method is: P b =P b ∪{p new1},if|P b | <K。

[0017] Based on the content of the above method embodiment, the online social media topic propagation collaborative loop detection method provided in the embodiment of the present invention, step 4 specifically includes: step 4.1: according to step 2, extracting a deep-direction diffusion path set consisting of the diffusion paths of the top-K in the deep direction; setting all path nodes to an unvisited state, starting from any unvisited node on any path, using the forwarding number of neighboring nodes as a weight, and performing a depth-first search from large to small according to the weight until the current search node has an edge with the visited node, then extracting the loop structure, that is, the searched end-to-end path structure, recording the node sequence of the loop path and the forwarding number of the loop path; or until all nodes on the current path are visited; step 4.2: if there are still paths in the deep-direction diffusion path set that have not been visited at this time, select another unvisited path for search, and repeat the above search process; step 4.3: extracting the loop structure information of the top-N before the total forwarding number of the loop path, including: the node sequence of the loop path and the forwarding number of the loop path.

[0018] Based on the content of the above method embodiment, the online social media topic propagation collaborative loop detection method provided in the embodiment of the present invention, step 5 specifically includes: step 5.1: according to step 3, extracting a broad-direction diffusion path set consisting of the top-K diffusion paths in the broad direction; setting all path nodes to an unvisited state, starting from any unvisited node on any path, using the forwarding number of neighboring nodes as a weight, and performing a depth-first search from large to small according to the weight until the current search node has an edge with the visited node, then extracting the loop structure, that is, the searched end-to-end path structure, recording the node sequence of the loop path and the forwarding number of the loop path; or until all nodes on the current path are visited; step 5.2: if there are still paths in the broad-direction diffusion path set that have not been visited at this time, select another unvisited path for search, and repeat the above search process; step 5.3: extracting the loop structure information of the top-N before the total forwarding number of the loop path, including: the nodes of the loop path and the total forwarding number of the loop path.

[0019] In the second aspect, an embodiment of the present invention provides an online social media topic propagation collaborative loop detection device, including: a first main module, which is used to realize the online social media network under the target topic, construct a topic propagation forwarding network cascade graph according to the information forwarding relationship between node accounts, and calculate the number of hops and forwarding amount forwarded by each node; a second main module, which is used to realize the search for the deep diffusion path of topic propagation on the cascade graph from each source account, and extract the deep diffusion path of the top-K deep diffusion path of the propagation path depth; a third main module, which is used to realize the search for the broad diffusion path of topic propagation from each source account, and extract the broad diffusion path of the top-K broad diffusion path of the propagation path forwarding number; a fourth main module, which is used to realize the detection of loops on the path based on the obtained deep diffusion path of topic propagation, that is, the path where the head and tail nodes are connected. path, extracting the top-N loops in the deep direction loop path forwarding number of topic propagation; the fifth main module is used to realize the detection of loops on the path according to the obtained broad direction diffusion path of topic propagation, that is, the path connecting the head and tail nodes, and extracting the top-N loops in the broad direction loop path forwarding number of topic propagation; the sixth main module is used to realize the extraction of the top-N loops in the deep direction loop path of topic propagation and the top-N loops in the broad direction loop path, and take the union of the two as the collaborative loop topology structure of topic propagation; wherein, the deep direction of propagation is the propagation direction with a first predetermined number of node forwarding hops from the source node in the propagation process; the broad direction of propagation is the propagation direction with a second predetermined number of node forwarding hops in the propagation process; K is the first constant given by the upper-layer application; N is the second constant given by the upper-layer application.

[0020] In a third aspect, an embodiment of the present invention provides an electronic device, including:

[0021] at least one processor; and

[0022] at least one memory communicatively coupled to the processor, wherein:

[0023] The memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to execute the online social media topic propagation collaborative loop detection method provided by any one of the various implementations of the first aspect.

[0024] In a fourth aspect, an embodiment of the present invention provides a non-transitory computer-readable storage medium, which stores computer instructions, and the computer instructions enable a computer to execute the online social media topic propagation collaborative loop detection method provided by any one of the various implementation methods of the first aspect.

[0025] The embodiments of the present invention provide a method and device for detecting collaborative loops in online social media topic propagation. By extracting key topological structures of loop feedback formed by node collaboration and with propagation path dependency during the online social media topic propagation process, the method and device discover nodes with organized collaborative interaction in topic propagation and their forwarding counts, thereby providing a basis for detecting manipulation of online social media public opinion. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0027] Figure 1 A flowchart of a method for detecting collaborative loops for online social media topic propagation provided by an embodiment of the present invention;

[0028] Figure 2 A schematic diagram of the structure of an online social media topic propagation collaboration loop detection device provided by an embodiment of the present invention;

[0029] Figure 3 A schematic diagram of the physical structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0030] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. In addition, the technical features in the various embodiments or single embodiments provided by the present invention can be arbitrarily combined with each other to form a feasible technical solution. This combination is not restricted by the sequence of steps and / or structural composition mode, but must be based on the ability of ordinary technicians in this field to implement it. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0031] The embodiment of the present invention provides a method for detecting online social media topic propagation collaborative loops, see Figure 1The method includes: step 1: constructing a topic propagation forwarding network cascade graph based on the information forwarding relationship between node accounts in the online social media network under the target topic, and calculating the forwarding hop count and forwarding amount of each node; step 2: starting from each source account, using the node hop count as the weight, searching for the deep diffusion path of the topic propagation on the cascade graph, and extracting the top-K deep diffusion paths in the depth of the propagation path; step 3: starting from each source account, using the node forwarding number as the weight, searching for the broad diffusion path of the topic propagation, and extracting the top-K broad diffusion paths in the forwarding number of the propagation path; step 4: based on the obtained top-K diffusion paths in the deep direction of the topic propagation, using the node forwarding number as the weight, detecting the loop on the path, that is, the path connecting the head and tail nodes, and extracting the deep diffusion path of the topic propagation. The top-N loops before the number of loop path forwardings; Step 5: Based on the obtained Top-K diffusion paths in the broad direction of topic propagation, with the number of node forwardings as the weight, detect the loops on the path, that is, the path connecting the head and tail nodes, and extract the top-N loops before the number of loop path forwardings in the broad direction of topic propagation; Step 6: Based on the extracted top-N loops before the deep direction loop path of topic propagation and the top-N loops before the broad direction loop path, take the union of the two as the collaborative loop topology structure of topic propagation; wherein, the deep direction of propagation is the propagation direction with a first predetermined number of node forwarding hops from the source node in the propagation process; the broad direction of propagation is the propagation direction with a second predetermined number of node forwarding hops in the propagation process; K is the first constant given by the upper-layer application; N is the second constant given by the upper-layer application.

[0032] Based on the content of the above method embodiment, as an optional embodiment, the online social media topic propagation collaborative loop detection method provided in the embodiment of the present invention specifically includes the following steps: Step 1.1: Based on the online social media network forwarding relationship dataset of the target topic and the account nodes, a node set V of the cascade graph is formed, and the forwarding relationship between the nodes constitutes a directed edge set E of the cascade graph, thereby constructing a topic propagation cascade graph G(V,E); Step 1.2: Extracting the user node that publishes the original blog post in the cascade graph, and determining that the user node is the source account of the message propagated by the target topic; starting from the source account, the number of hops from the node to the source account includes: obtaining the number of hops forwarded by the node; and calculating the number of forwarding of the node according to the out-degree of the node in the cascade graph.

[0033] Based on the content of the above method embodiment, as an optional embodiment, the online social media topic propagation collaborative loop detection method provided in the embodiment of the present invention, wherein the forwarding number of the node is calculated based on the out-degree of the node in the cascade graph, includes:

[0034] fw(v)=OD(v),v∈V

[0035] Where fw(v) is the number of forwarding of node v, and OD(v) is the out-degree of the node.

[0036] Specifically, the set of all nodes participating in topic propagation forms the vertex set V of the cascade graph. Based on the forwarding relationship between the source node and the target node of the message forwarding, a set of directed edges E is constructed between all nodes, resulting in the topic propagation forwarding cascade graph G(V,E). The source account can be determined by whether the node account node has an incoming edge. Based on this, each node calculates the number of hops from the source account and the forwarding count based on the out-degree of each node.

[0037] Based on the content of the above method embodiment, as an optional embodiment, the online social media topic propagation collaborative loop detection method provided in the embodiment of the present invention specifically includes the following steps: Step 2.1: Starting from any source account, visit this node, and start from the neighboring nodes that have not been visited by the source account in turn, using the number of hops of the neighboring nodes as weights, and perform depth-first search according to the weights from large to small until the search of K paths is completed, or all nodes in the cascade graph that have paths connected to the source account are visited; during the search process, calculate and save the depth of the path, that is, the number of hops from the current search path to the source account; Step 2.2: If there is still a source account that has not been visited at this time, select another source account that has not been visited as the starting point and repeat the above search process; Step 2.3: If the number of search paths in step 2.1 reaches K, and the depth of the new search path in step 2.2 is greater than the depth of the search path in step 2.1, then replace it and retain the path with the top-K depth in the current steps 2.1 and 2.2; the specific calculation method is:

[0038] min(p∈P d )=p new ,if|P d |==K

[0039] Among them, P d is the depth path set currently searched by step 2.1, p new is the new path found in step 2.2, |P d |Get the potential of the deep direction path set, that is, the number of paths; min(p∈P d ) Calculate the path with the minimum weight, i.e., the minimum path depth, in the current depth-direction path set; Step 2.4: If the number of search paths in step 2.1 does not reach K, then the new search path in step 2.2 is added to the search path set. The specific calculation method is:

[0040] P d =P d ∪{p new},if|P d | <K

[0041] Among them, P d is the depth path set currently searched by step 2.1, p new is the new path searched in step 2.2, and |Pd| is the potential of the deep direction path set, that is, the number of paths.

[0042] Specifically, the top-K deep paths constructed in steps 2.1 and 2.2 are retained until the total number of paths searched in steps 2.1 and 2.2 reaches K, or all nodes in the cascade graph that have paths connecting to the source account have been visited. This results in the deep diffusion paths of the top-K deep propagation paths.

[0043] Based on the content of the above method embodiment, as an optional embodiment, the online social media topic propagation collaborative loop detection method provided in the embodiment of the present invention specifically includes the following steps: Step 3.1: Starting from a source account, accessing nodes, starting from the neighboring nodes that have not been visited by the source account in turn, using the forwarding number of the neighboring nodes as weights, and performing a depth-first search according to the weights from large to small until the search of N paths is completed, or all nodes in the cascade graph that have paths connected to the source account are visited; during the search process, the forwarding number of the path is calculated and saved, that is, the sum of the forwarding numbers of each node on the current search path; Step 3.2: If there is still a source account that has not been visited at this time, another source account that has not been visited is selected as the starting point, and the above search process is repeated; Step 3.3: If the number of search paths in step 3.1 reaches K, and the depth of the new search path in step 3.2 is greater than the depth of the search path in step 3.1, then a replacement is performed, and the diffusion path with the top-K forwarding numbers in the current steps 3.1 and 3.2 is retained; the specific calculation method is:

[0044] min(p∈P b )=p new1 ,if|P b |==K

[0045] Among them, P b is the set of broad-direction paths currently searched in step 3.1, p new1 is the new path found in step 3.2, |P b |Get the potential of the broad path set, that is, the number of paths; min(p∈P b ) Calculate the path with the minimum weight in the current broad direction path set, i.e., the path with the minimum forwarding number; Step 3.4: If the number of search paths in step 3.1 does not reach K, then the new search path in step 3.2 is added to the search path set. The specific calculation method is: P b =P b ∪{p new1},if|Pb | <K。

[0046] Specifically, the top-K forwarding paths from steps 3.1 and 3.2 are retained until the total number of paths searched in steps 3.1 and 3.2 reaches K, or all nodes in the cascade graph that have paths connecting to the source account have been visited. This results in the broad diffusion paths of the top-K forwarding paths.

[0047] Based on the content of the above method embodiment, as an optional embodiment, the online social media topic propagation collaborative loop detection method provided in the embodiment of the present invention, step 4 specifically includes: step 4.1: according to step 2, extracting a deep diffusion path set; setting all path nodes to an unvisited state, starting from any unvisited node on any path, using the forwarding number of neighboring nodes as a weight, and performing a depth-first search from large to small according to the weight until the current search node has an edge with the visited node, then extracting the loop structure, that is, the searched end-to-end path structure, recording the node sequence of the loop path and the forwarding number of the loop path; or until all nodes on the current path are visited; step 4.2: if there are still unvisited paths in the deep diffusion path set composed of the top-K diffusion paths in the deep direction, then select another unvisited path for search and repeat the above search process; step 4.3: extracting the loop structure information of the top-N loop paths before the total forwarding number, including: the node sequence of the loop path and the forwarding number of the loop path.

[0048] Based on the content of the above method embodiment, as an optional embodiment, the online social media topic propagation collaborative loop detection method provided in the embodiment of the present invention, step 5 specifically includes: step 5.1: according to step 3, extracting a broad-direction diffusion path set consisting of the top-K diffusion paths in the broad direction; setting all path nodes to an unvisited state, starting from any unvisited node on any path, using the forwarding number of neighboring nodes as a weight, and performing a depth-first search from large to small according to the weight until the current search node has an edge with the visited node, then extracting the loop structure, that is, the searched end-to-end path structure, recording the node sequence of the loop path and the forwarding number of the loop path; or until all nodes on the current path are visited; step 5.2: if there are still paths in the broad-direction diffusion path set that have not been visited at this time, select another unvisited path for search, and repeat the above search process; step 5.3: extracting the loop structure information of the top-N before the total forwarding number of the loop path, including: the nodes of the loop path and the total forwarding number of the loop path.

[0049] In another embodiment, step 6 takes the union of the extracted top-N loops in the deep direction and the top-N loops in the broad direction of the topic propagation loop path as the loop topology structure result for the topic propagation. In a specific implementation, the loop topology structure is the top-N loops in the deep direction of the topic propagation loop path, the node sequence of the broad direction loop path, and the number of loop path forwardings, as well as the top-N loops in the deep direction of the topic propagation loop path, the node sequence of the broad direction loop path, and the number of loop path forwardings.

[0050] The embodiment of the present invention provides a method for detecting collaborative loops in online social media topic propagation. By extracting key topological structures of loop feedback formed by node collaboration and with propagation path dependency during the online social media topic propagation process, the method discovers nodes with organized collaborative interaction in topic propagation and their forwarding counts, thereby providing a basis for detecting manipulation of online social media public opinion.

[0051] The implementation basis of each embodiment of the present invention is to implement it through programmed processing by a device with processor function. Therefore, in engineering practice, the technical solutions and functions of each embodiment of the present invention can be encapsulated into various modules. Based on this reality, on the basis of the above embodiments, an embodiment of the present invention provides an online social media topic propagation collaborative loop detection device, which is used to execute the online social media topic propagation collaborative loop detection method in the above method embodiment. Figure 2The device includes: a first main module, which is used to build a topic propagation forwarding network cascade graph based on the information forwarding relationship between node accounts in the online social media network under the target topic, and calculate the number of hops and forwarding amount forwarded by each node; a second main module, which is used to search for the deep diffusion path of the topic propagation on the cascade graph from each source account, and extract the top-K deep diffusion paths in the depth of the propagation path; a third main module, which is used to search for the broad diffusion path of the topic propagation from each source account, and extract the top-K broad diffusion paths in the forwarding number of the propagation path; a fourth main module, which is used to detect the loop on the path, that is, the path connecting the head and tail nodes, based on the obtained deep diffusion path of the topic propagation, and extract the deep loop path of the topic propagation. The top-N loops before the forwarding number; the fifth main module is used to detect the loops on the path, that is, the path connecting the head and tail nodes, according to the obtained broad-direction diffusion path of the topic propagation, and extract the top-N loops before the forwarding number of the broad-direction loop path of the topic propagation; the sixth main module is used to realize the extracted top-N loops before the deep-direction loop path of the topic propagation and the top-N loops before the broad-direction loop path, and take the union of the two as the collaborative loop topology structure of the topic propagation; wherein, the deep direction of propagation is the propagation direction with a first predetermined number of node forwarding hops from the source node in the propagation process; the broad direction of propagation is the propagation direction with a second predetermined number of node forwarding hops in the propagation process; K is the first constant given by the upper-layer application; N is the second constant given by the upper-layer application.

[0052] The embodiment of the present invention provides an online social media topic propagation collaborative loop detection device, which adopts Figure 2 Several modules in it extract the key topological structure of loop feedback composed of node collaboration with propagation path dependence in the process of online social media topic propagation, discover the nodes with organized collaborative interaction in topic propagation and their forwarding number information, and provide a basis for discovering online social media public opinion manipulation.

[0053] It should be noted that the device in the device embodiment provided by the present invention can be used to implement the method in the above-mentioned method embodiment as well as the method in other method embodiments provided by the present invention. The only difference is that the corresponding functional modules are set. The principle is basically the same as the principle of the above-mentioned device embodiment provided by the present invention. As long as those skilled in the art refer to the specific technical solutions in other method embodiments on the basis of the above-mentioned device embodiment, obtain the corresponding technical means and the technical solutions composed of these technical means by combining technical features, and ensure the practicality of the technical solutions, they can improve the device in the above-mentioned device embodiment to obtain the corresponding device class embodiment, thereby obtaining the corresponding device class embodiment for implementing the methods in other method class embodiments. For example:

[0054] Based on the content of the above-mentioned device embodiment, as an optional embodiment, the online social media topic propagation collaborative circuit detection device provided in the embodiment of the present invention further includes: a first sub-module, which is used to implement step 1, specifically including: step 1.1: based on the online social media network forwarding relationship data set of the target topic and the account nodes, a node set V of the cascade graph is formed, and the forwarding relationship between the nodes constitutes a directed edge set E of the cascade graph, thereby constructing a topic propagation cascade graph G(V,E); step 1.2: extracting the user node that publishes the original blog post in the cascade graph, and determining that the user node is the source account of the message propagated by the target topic; starting from the source account, the number of hops from the node to the source account includes: obtaining the number of hops forwarded by the node; and calculating the number of forwarding of the node according to the out-degree of the node in the cascade graph.

[0055] Based on the content of the above device embodiment, as an optional embodiment, the online social media topic propagation collaborative loop detection device provided in the embodiment of the present invention further includes: a second submodule for implementing the calculation of the forwarding number of the node based on the out-degree of the node in the cascade graph, including:

[0056] fw(v)=OD(v),v∈V

[0057] Where fw(v) is the number of forwarding of node v, and OD(v) is the out-degree of the node.

[0058] Based on the content of the above-mentioned device embodiment, as an optional embodiment, the online social media topic propagation collaborative loop detection device provided in the embodiment of the present invention further includes: a third submodule, which is used to implement step 2, specifically including: step 2.1: starting from any source account, access this node, and starting from the neighboring nodes that have not been visited by the source account in turn, using the number of hops of the neighboring nodes as weights, and performing depth-first search according to the weights from large to small, until the search of K paths is completed, or all nodes in the cascade graph that have paths connected to the source account are visited; during the search process, the depth of the path is calculated and saved, that is, the number of hops from the current search path to the source account; step 2.2: if there is still a source account that has not been visited at this time, select another source account that has not been visited as the starting point and repeat the above search process; step 2.3: if the number of search paths in step 2.1 reaches K, and the depth of the new search path in step 2.2 is greater than the depth of the search path in step 2.1, then replace it and retain the path with the top-K depth in the current steps 2.1 and 2.2; the specific calculation method is:

[0059] min(p∈P d )=p new ,if|P d |==K

[0060] Among them, Pd is the depth path set currently searched by step 2.1, p new is the new path found in step 2.2, |P d |Get the potential of the deep direction path set, that is, the number of paths; min(p∈P d ) Calculate the path with the minimum weight, i.e., the minimum path depth, in the current depth-direction path set; Step 2.4: If the number of search paths in step 2.1 does not reach K, then the new search path in step 2.2 is added to the search path set. The specific calculation method is:

[0061] P d =P d ∪{p new},if|P d | <K

[0062] Among them, P d is the depth path set currently searched by step 2.1, p new is the new path searched in step 2.2, and |Pd| is the potential of the deep direction path set, that is, the number of paths.

[0063] Based on the content of the above-mentioned device embodiment, as an optional embodiment, the online social media topic propagation collaborative loop detection device provided in the embodiment of the present invention further includes: a fourth submodule, which is used to implement step 3, specifically including: step 3.1: starting from a source account to visit the node, starting from the neighboring nodes that have not been visited by the source account in turn, using the forwarding number of the neighboring nodes as the weight, and performing a depth-first search according to the weight from large to small until the search of N paths is completed, or all nodes in the cascade graph that have a path connected to the source account are visited; during the search process, the forwarding number of the path is calculated and saved, that is, the sum of the forwarding numbers of each node on the current search path; step 3.2: if there is still a source account that has not been visited at this time, another source account that has not been visited is selected as the starting point, and the above search process is repeated; step 3.3: if the number of search paths reaches K in step 3.1, and the depth of the new search path in step 3.2 is greater than the depth of the search path in step 3.1, then a replacement is performed, and the diffusion path with the top-K forwarding numbers in the current steps 3.1 and 3.2 is retained; the specific calculation method is:

[0064] min(p∈P b )=p new1 ,if|P b |==K

[0065] Among them, P b is the set of broad-direction paths currently searched in step 3.1, p new1 is the new path found in step 3.2, |P b|Get the potential of the broad path set, that is, the number of paths; min(p∈P b ) Calculate the path with the minimum weight in the current broad direction path set, i.e., the path with the minimum forwarding number; Step 3.4: If the number of search paths in step 3.1 does not reach K, then the new search path in step 3.2 is added to the search path set. The specific calculation method is: P b =P b ∪{p new1},if|P b | <K。

[0066] Based on the content of the above-mentioned device embodiment, as an optional embodiment, the online social media topic propagation collaborative loop detection device provided in the embodiment of the present invention further includes: a fifth submodule, which is used to implement step 4, specifically including: step 4.1: according to step 2, extracting a deep diffusion path set; setting all path nodes to an unvisited state, starting from any unvisited node on any path, using the forwarding number of neighboring nodes as a weight, and performing a depth-first search from large to small according to the weight until the current search node has an edge with the visited node, then extracting the loop structure, that is, the searched end-to-end path structure, recording the node sequence of the loop path and the forwarding number of the loop path; or until all nodes on the current path are visited; step 4.2: if there are still paths in the deep diffusion path set composed of the top-K diffusion paths in the deep direction that have not been visited, then select another unvisited path for search, and repeat the above search process; step 4.3: extracting the loop structure information of the top-N before the total forwarding number of the loop path, including: the node sequence of the loop path and the forwarding number of the loop path.

[0067] Based on the content of the above-mentioned device embodiment, as an optional embodiment, the online social media topic propagation collaborative loop detection device provided in the embodiment of the present invention further includes: a sixth submodule, which is used to implement step 5, specifically including: step 5.1: according to step 3, extracting a set of broad-direction diffusion paths; setting all path nodes to an unvisited state, starting from any unvisited node on any path, using the forwarding number of neighboring nodes as a weight, and performing a depth-first search from large to small according to the weight until the current search node has an edge with the visited node, then extracting the loop structure, that is, the searched end-to-end path structure, recording the node sequence of the loop path and the forwarding number of the loop path; or until all nodes on the current path are visited; step 5.2: if there are still paths in the broad-direction diffusion path set composed of the top-K diffusion paths in the broad direction that have not been visited, then select another unvisited path for search, and repeat the above search process; step 5.3: extracting the loop structure information of the top-N before the total forwarding number of the loop path, including: the nodes of the loop path and the total forwarding number of the loop path.

[0068] The method of the embodiment of the present invention is implemented by electronic devices, so it is necessary to introduce the relevant electronic devices. Based on this purpose, the embodiment of the present invention provides an electronic device, such as Figure 3 As shown, the electronic device includes: at least one processor, a communications interface, at least one memory, and a communications bus, wherein the at least one processor, the communications interface, and the at least one memory communicate with each other via the communications bus. The at least one processor can call logic instructions in the at least one memory to execute all or part of the steps of the methods provided in the aforementioned method embodiments.

[0069] In addition, the logic instructions in the at least one memory mentioned above can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present invention, 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, server, or network device, etc.) to execute all or part of the steps of the method described in each method embodiment of the present invention. 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.

[0070] The device embodiments described above are merely illustrative. 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, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0071] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course can also be implemented by hardware. Based on this understanding, the essence of the above technical solution or the part that contributes to the existing technology can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiment.

[0072] The flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions and operations of the systems, methods and computer program products according to multiple embodiments of the present invention. Based on this understanding, each box in the flowchart or block diagram can represent a module, program segment or part of the code, and the module, program segment or part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, or sometimes in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0073] It should be noted that the terms "comprise," "include," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus comprising a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, the elements defined by the phrase "comprise..." do not preclude the presence of additional identical elements in the process, method, article, or apparatus comprising the elements.

[0074] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for detecting collaborative loops in online social media topic propagation, characterized in that: include: Step 1: Based on the information forwarding relationship between node accounts in the online social media network under the target topic, a topic propagation forwarding network cascade graph is constructed, and the number of hops and forwarding volume of each node are calculated; Step 2: Starting from each source account, with the number of node hops as the weight, the deep diffusion path of the topic propagation is searched on the cascade graph, and the deep diffusion path of the top-K in the depth of the propagation path is extracted; Step 3: Starting from each source account, with the number of node forwarding as the weight, the broad diffusion path of the topic propagation is searched, and the broad diffusion path of the top-K in the forwarding number of the propagation path is extracted; Step 4: Based on the obtained top-K diffusion path in the deep direction of the topic propagation, with the number of node forwarding as the weight, the loop on the path, that is, the path connecting the head and tail nodes, is detected, and the deep loop path of the topic propagation is extracted. The top-N circuits before the forwarding number; Step 5: According to the obtained top-K diffusion paths in the broad direction of topic propagation, with the node forwarding number as the weight, detect the circuits on the path, that is, the path connecting the head and tail nodes, and extract the top-N circuits before the forwarding number of the broad direction circuit path of topic propagation; Step 6: According to the extracted top-N circuits before the deep direction circuit path of topic propagation and the top-N circuits before the broad direction circuit path, take the union of the two as the collaborative circuit topology structure of topic propagation; wherein, the deep direction of propagation is the propagation direction with a first predetermined number of node forwarding hops from the source node in the propagation process; the broad direction of propagation is the propagation direction with a second predetermined number of node forwarding hops in the propagation process; K is the first constant given by the upper-layer application; N is the second constant given by the upper-layer application.

2. The online social media topic propagation collaborative loop detection method according to claim 1, characterized in that: Step 1 specifically includes: Step 1.1: Based on the online social media network forwarding relationship dataset of the target topic and the account nodes, a node set V of the cascade graph is formed, and the forwarding relationship between the nodes constitutes a directed edge set E of the cascade graph, thereby constructing a topic propagation cascade graph G(V,E); Step 1.2: Extracting the user node that publishes the original blog post in the cascade graph, and determining that the user node is the source account of the message propagation of the target topic; starting from the source account, the number of hops of the node forwarding is obtained by the number of hops from the node to the source account; and calculating the number of forwarding of the node according to the out-degree of the node in the cascade graph.

3. The online social media topic propagation collaborative loop detection method according to claim 2, characterized in that: The method of calculating the forwarding number of a node according to the out-degree of the node in the cascade graph includes: fw(v)=OD(v),v∈V Wherein, node v is any node in the node set V, fw(v) is the forwarding number of node v, and OD(v) is the node out-degree.

4. The online social media topic propagation collaborative loop detection method according to claim 3, characterized in that: Step 2 specifically includes: Step 2.1: Starting from any source account, visit this node, starting from the neighboring nodes that have not been visited by the source account in turn, using the number of hops of the neighboring nodes as weight, and performing depth-first search from large to small according to the weight until the search of the specified K paths is completed, or all nodes in the cascade graph that have paths connected to the source account are visited; during the search process, calculate and save the depth of the path, that is, the number of hops from the current search path to the source account; Step 2.2: If there is still a source account that has not been visited at this time, select another source account that has not been visited as the starting point and repeat the above search process; Step 2.3: If the number of search paths in step 2.1 reaches K, and the depth of the new search path in step 2.2 is greater than the depth of the search path in step 2.1, then replace it and retain the path with the top-K depth in the current steps 2.1 and 2.2; the specific calculation method is: min(p∈P d )=p new ,if|P d |==K Among them, P d is the depth path set currently searched by step 2.1, p new is the new path found in step 2.2, |P d |Get the potential of the deep direction path set, that is, the number of paths; min(p∈P d ) Calculate the minimum weight of the current depth-direction path set, that is, the path with the minimum path depth; Step 2.4: If the number of search paths in step 2.1 does not reach K, then the new search path in step 2.2 is added to the search path set. The specific calculation method is: P d =P d ∪{p new },if|P d |<K Among them, P d is the depth path set currently searched by step 2.1, p new is the new path searched in step 2.2, and |Pd| is the potential of the deep direction path set, that is, the number of paths.

5. The online social media topic propagation collaborative loop detection method according to claim 4, characterized in that: Step 3 specifically includes: Step 3.1: Starting from a source account, visit the node, starting from the neighboring nodes that have not been visited by the source account in turn, using the forwarding number of the neighboring nodes as the weight, and performing a depth-first search from large to small according to the weight until the search of N paths is completed, or all nodes in the cascade graph that have paths connected to the source account are visited; during the search process, calculate and save the forwarding number of the path, that is, the sum of the forwarding numbers of each node on the current search path; Step 3.2: If there is still a source account that has not been visited at this time, select another source account that has not been visited as the starting point and repeat the above search process; Step 3.3: If the number of search paths in step 3.1 reaches K, and the depth of the new search path in step 3.2 is greater than the depth of the search path in step 3.1, then replace it and retain the diffusion path with the top-K forwarding numbers in the current steps 3.1 and 3.2; the specific calculation method is: min(p∈P b )=p new1 ,if|P b |==K Among them, P b is the set of broad-direction paths currently searched in step 3.1, p new1 is the new path found in step 3.2, |P b |Get the potential of the broad path set, that is, the number of paths; min(p∈P b ) Calculate the minimum weight of the current broad-direction path set, that is, the path with the minimum number of forwarding paths; Step 3.4: If the number of search paths in step 3.1 does not reach K, then the new search path in step 3.2 is added to the search path set. The specific calculation method is: P b =P b ∪{p new1 },if|P b |<K。 6. The online social media topic propagation collaborative loop detection method according to claim 5, characterized in that: Step 4 specifically includes: Step 4.1: According to Step 2, extract the deep-direction diffusion path set consisting of the diffusion paths of the top-K in the deep direction; set all path nodes to an unvisited state, starting from any unvisited node on any path, and using the forwarding number of neighboring nodes as the weight, perform a depth-first search from large to small according to the weight until the current search node has an edge with the visited node, then extract the loop structure, that is, the searched end-to-end path structure, record the node sequence of the loop path and the forwarding number of the loop path; or until all nodes on the current path are visited; Step 4.2: If there are still paths in the deep-direction diffusion path set that have not been visited at this time, select another path that has not been visited to search, and repeat the above search process; Step 4.3: Extract the loop structure information of the top-N before the total forwarding number of the loop path, including: the node sequence of the loop path and the forwarding number of the loop path.

7. The online social media topic propagation collaborative loop detection method according to claim 6, characterized in that: Step 5 specifically includes: Step 5.1: According to Step 3, extract the broad-direction diffusion path set consisting of the top-K diffusion paths in the broad direction; set all path nodes to an unvisited state, starting from any unvisited node on any path, and using the forwarding number of neighboring nodes as the weight, perform a depth-first search from large to small according to the weight until the current search node has an edge with the visited node, then extract the loop structure, that is, the searched end-to-end path structure, record the node sequence of the loop path and the forwarding number of the loop path; or until all nodes on the current path are visited; Step 5.2: If there are still paths in the broad-direction diffusion path set that have not been visited at this time, select another path that has not been visited to search, and repeat the above search process; Step 5.3: Extract the loop structure information of the top-N before the total forwarding number of the loop path, including: the nodes of the loop path and the total forwarding number of the loop path.

8. A device for detecting collaborative loops of online social media topic propagation, characterized in that: include: The first main module is used to build a topic propagation and forwarding network cascade graph based on the information forwarding relationship between node accounts in the online social media network under the target topic, and calculate the number of hops and forwarding volume of each node; The second main module is used to realize the search for the deep diffusion path of topic propagation on the cascade graph starting from each source account, with the number of node hops as the weight, and extract the deep diffusion path of the top-K deep diffusion path before the depth of the propagation path; the third main module is used to realize the search for the broad diffusion path of topic propagation starting from each source account, with the number of node forwarding as the weight, and extract the broad diffusion path of the top-K broad diffusion path before the forwarding number of the propagation path; the fourth main module is used to realize the detection of the loop on the path based on the obtained deep Top-K diffusion path of topic propagation, with the number of node forwarding as the weight, and extract the loop of the top-N deep loop path forwarding number of topic propagation; the fifth main module is used to realize the detection of the deep Top-K diffusion path of topic propagation based on the obtained deep Top-K diffusion path of topic propagation, with the number of node forwarding as the weight, and extract the loop of the top-N deep loop path forwarding number of topic propagation; The Top-K diffusion path in the broad direction of topic propagation uses the number of node forwarding as weight to detect the loop on the path, that is, the path connecting the head and tail nodes, and extract the Top-N loops with the largest forwarding number in the broad direction loop path of topic propagation; the sixth main module is used to implement the extracted Top-N loops in the deep direction loop path of topic propagation and the Top-N loops in the broad direction loop path, and take the union of the two as the collaborative loop topology structure of topic propagation; wherein, the deep direction of propagation is the propagation direction with a first predetermined number of node forwarding hops from the source node in the propagation process; the broad direction of propagation is the propagation direction with a second predetermined number of node forwarding hops in the propagation process; K is the first constant given by the upper-layer application; N is the second constant given by the upper-layer application.

9. An electronic device, characterized in that: include: At least one processor, at least one memory and a communication interface; wherein, The processor, memory and communication interface communicate with each other; The memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to execute the method according to any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium, characterized in that The non-transitory computer-readable storage medium stores computer instructions, which cause the computer to execute the method of any one of claims 1 to 7.

Images