Search

CN-117453972-B - Construction method of topic topology fingerprint diagram of online social platform

CN117453972BCN 117453972 BCN117453972 BCN 117453972BCN-117453972-B

Abstract

The invention provides a method for constructing a topic transmission topological fingerprint graph of an online social platform, which comprises the steps of constructing a topic transmission network graph, determining source nodes, extracting maximum connected subgraphs, calculating the forwarding quantity of each node, the hop count of each node and the forwarding strength of adjacent edges of the subgraphs, starting from each subgraph source node, taking the forwarding strength as a weight and the shortest weighted path from the source node as a filtering function, searching forward for a connected component sequence, recording the starting time step and the duration of each connected component, merging the same connected components, starting from the searching ending node recorded by each subgraph, searching reversely, obtaining a ring sequence by searching reversely, recording the starting time step and the duration of each ring structure, merging the same ring structure, and constructing the topic transmission fingerprint graph according to a connected component sequence set and a ring sequence set. The method can effectively extract the remarkable topological structure of the communication path and the circuit formed in the topic transmission process of the online social platform and quantify the duration time of the communication path and the circuit.

Inventors

  • ZENG YUANYUAN
  • JIANG HAO

Assignees

  • 武汉大学

Dates

Publication Date
20260512
Application Date
20231031

Claims (9)

  1. 1. The method for constructing the topic topology fingerprint map of the online social platform is characterized by comprising the following steps of: Aiming at target topic data of an online social platform, constructing a topic propagation network graph according to information propagation forwarding relations among node accounts of the social platform, determining a topic source node account, extracting a maximum connected subgraph starting from a source node, calculating information forwarding quantity of each node in the subgraph and hop count of the node from the source node, and calculating forwarding strength of adjacent nodes on the basis and taking the forwarding strength as edge weight; Step 2, based on the subgraph extracted by the topic propagation network graph, starting from any subgraph source node, forward searching is carried out by taking the forwarding strength as an edge weight and the shortest weighted path from the source node as a filtering function until all nodes in the subgraph are accessed, and searching ending nodes are recorded; step 3, starting from other sub-image source nodes which are not searched forward in sequence, executing step 2, forming a sequence obtained by searching a plurality of sub-image source nodes, forming a connected component sequence set comprising a starting time step and a duration, combining the same connected components obtained by searching all sub-image source nodes, wherein the starting time is the minimum starting time, and the duration is the sum of the duration of all connected components in each sequence; Step 4, for each subgraph, starting reverse search from the recorded end node, searching by taking the forwarding strength as an edge weight and taking the shortest weighted path from the source node as a filtering function until returning to the source node, filtering out a ring sequence on the topic propagation network graph, and recording the relative time steps of the connection component and the ring start and end in each group of sequences; step 5, starting from the end nodes of other sub-graph records which are not searched in the reverse direction in turn, executing the step 4, forming a plurality of sequences obtained by searching the source nodes, forming a ring sequence set comprising a starting time step and a duration, merging the same rings obtained by searching the source nodes of each sub-graph, wherein the starting time is the minimum starting time, and the duration is the sum of the durations of each ring structure in each sequence; And 6, extracting a connected component sequence set and the starting time step and duration time of each sequence in the ring sequence set to construct a coordinate pair, combining the coordinate pairs to form a coordinate set, drawing each point in a two-dimensional plane coordinate system, and constructing a fingerprint map of the key topological characteristic of the target topic transmission.
  2. 2. The method for constructing the topic topology fingerprint map of the online social platform according to claim 1, wherein the step 1 specifically includes: aiming at target topic data of an online social platform, a node set V of a topic propagation network diagram is formed according to node account information related to a target topic propagation process, a directed edge set E of a cascade diagram is formed according to topic propagation forwarding relations among nodes, and a topic propagation network diagram G (V, E) is constructed according to the directed edge set E; Step 1.2, extracting user nodes which release original blogs in a topic transmission network diagram, and determining that the user nodes are message source accounts of target topic transmission; extracting a maximum connected subgraph G ' (V ', E ') communicated with the source node by adopting a depth-first algorithm according to each source account, namely the source node; step 1.3, according to the extracted subgraph G ' (V ', E '), starting from a source account, calculating the forwarding quantity of each node according to the node forwarding relation, and calculating the forwarding strength of adjacent nodes, wherein the forwarding quantity calculating method of each node comprises the following steps: Wherein, node V is any node in node set V 'of sub-graph G', fw (V) is the forwarding number of node V, and OD (V) is node output; step 1.4, calculating the hop count of the node from the source node by taking the source node as the 0 th hop, and calculating the forwarding strength of the adjacent node: Wherein, node V is any node in node set V 'of sub graph G', s (V) is forwarding strength of node V, OD (V) is node egress, and ID (V) is node ingress.
  3. 3. The method for constructing the topic topology fingerprint map of the online social platform according to claim 1, wherein in the step 2, a filtering function with forwarding strength as an edge weight and a shortest weighted path from a source node is as follows: wherein v i is any node in the sub graph G', v s is the current source node, d (v i ,v s ) is the distance from the node v i to the source node v s , and min is the minimum function.
  4. 4. The method for building an online social platform topic topology fingerprint graph as recited in claim 3, wherein in the step 2 forward search, the search strategy of the current node v i is to select an adjacent node v j with the shortest weighted path distance from the source node v s to access, namely: Wherein v i 、v j is any node in the sub graph G', f (v i ) is the shortest weighted path length of the node v i from the source node v s , f (v j ) is the shortest weighted path length of the node v j from the source node v s , and min is a minimum function.
  5. 5. The method for constructing the topic topology fingerprint map of the online social platform according to claim 1, wherein the step 3 comprises the following sub-steps: Step 3.1, starting from other non-searched sub-graph source nodes in turn, and executing the step 2, thereby obtaining a plurality of connected component sequences searched by starting from the sub-graph source nodes, and forming a connected component sequence set containing a starting time step and duration; Step 3.2, in the connected component sequence set obtained in step 3.1, merging the same connected components obtained by starting and searching each sub-graph through the source node, wherein the starting time is the minimum starting time, and the duration is the sum of the duration of each connected component in each sequence, namely: Wherein c mi is the ith connected component obtained by filtering with v sm as a source node, c nj is the jth connected component obtained by filtering with v sn as a source node, t begin (c mi ) is the starting time step of the connected component c mi , t begin (c nj ) is the starting time step of the connected component c nj , t p (c mi ) is the duration of the connected component c mi , and t p (c nj ) is the duration of the connected component c nj .
  6. 6. The method for constructing the topic topology fingerprint map of the online social platform according to claim 1, wherein the implementation manner of the step 4 is as follows: Step 4.1, on the basis of searching in the step 3, reversely searching from the recorded end node to the source node direction aiming at each subgraph, and taking the forwarding strength as the edge weight and the shortest weighted path from the source node as the filtering function; step 4.2, executing step 4.1 until tracing back to the source node v s , according to the above-mentioned process, filtering out the ring sequence on the sub-graph G', taking the sub-graph source node as 0 time steps, taking the hop count of each node from the source node as relative time steps, recording the time steps t begin and t end of starting and ending the rings in each group of sequences, and calculating the duration time t p of the ring structure according to the above-mentioned steps, wherein the ring sequence is represented by the following modes: [ (ring 1, t begin ,t p ), (ring 2, t begin ,t p ), (ring n, t begin ,t p ) ] Wherein t p = t begin -t end .
  7. 7. The method for constructing an online social platform topic topology fingerprint map of claim 6, wherein the reverse search strategy is: In the reverse search process, starting from the current node v i , selecting the adjacent node v j with the shortest weighted path distance from the source node and the largest drop to access, namely: Wherein v i ,v j is any node in the sub graph G', f (v i ) is the shortest weighted path length of the node v i from the source node v s , f (v j ) is the shortest weighted path length of the node v j from the source node v s , and max is the maximum function.
  8. 8. The method for constructing the topic topology fingerprint map of the online social platform according to claim 1, wherein the step 5 comprises the following sub-steps: Step 5.1, starting from the end nodes of other unsearched sub-graph records in sequence, executing step 4, thus obtaining a ring sequence reversely searched by a plurality of sub-graph source nodes, and forming a ring sequence set containing a starting time step and duration; Step 5.2, in the ring sequence set obtained in step 5.1, merging the same ring structures obtained by searching each sub-graph source node, wherein the starting time is the minimum starting time, and the duration is the sum of the duration of each connected component in each sequence, namely: Wherein, l mi is the ith ring filtered by v sm as the source node, l nj is the jth ring filtered by v sn as the source node, t begin (l mi ) is the starting time step of the ring l mi , t begin (l nj ) is the starting time step of the ring l nj , t p (l mi ) is the duration of the ring l mi , and t p (l nj ) is the duration of the ring l nj .
  9. 9. The method for constructing a topic topology fingerprint map of an online social platform according to claim 1, wherein in step 6, a union of a set of connected component sequences and a set of loop sequences is extracted, starting time steps and durations of the sequences are reserved and concentrated, and coordinate pairs are constructed according to the starting time steps and durations, and are combined to form a coordinate set, and the form of the coordinate set is: {(t begin1 ,t p1 ）,（t begin2 ,t p2 ）…（t beginN ,t pN )} Wherein t begini is the time step at which the union starts, t pi is the union duration, where i=1, 2,...

Description

Construction method of topic topology fingerprint diagram of online social platform Technical Field The invention belongs to the technical field of online social platform topic transmission big data analysis, and particularly relates to a construction method of an online social platform topic topology fingerprint map. Background The internet has entered the 2.0 era, and new media such as online social platforms are becoming the main tools for topic origin and topic propagation. The current online social platform topic transmission shows the tendency and characteristics of weaponization, and the online social platform is used as a topic transmission tool and can be used as a powerful psychological intervention weapon in the network information war field. Therefore, the fractal topological structure with the potential ordered propagation characteristics in the irregular complex phenomenon and the topology fingerprint diagram are timely and effectively found in the online social platform topic propagation process, and the method is a key for understanding online social platform topic propagation. Disclosure of Invention Aiming at the defects of the prior art, the invention provides a construction method of the topic topology fingerprint diagram of the online social platform, which can effectively extract the remarkable topology structure of a communication path and a circuit formed in the topic transmission process of the online social platform and quantify the duration time of the communication path and the circuit, and the communication path and the circuit are expressed as coordinate points in the diagram, thereby facilitating further subsequent task processing and assisting topic analysis and prediction. In order to solve the technical problems, the invention adopts the following technical scheme: a construction method of an online social platform topic topology fingerprint diagram comprises the following steps: Aiming at target topic data of an online social platform, constructing a topic propagation network graph according to information propagation forwarding relations among node accounts of the social platform, determining a topic source node account, extracting a maximum connected subgraph starting from the source node, calculating information forwarding quantity of each node in the subgraph and hop count of each node from the source node, and calculating forwarding strength of adjacent nodes on the basis and taking the forwarding strength as edge weight; step 2, based on the subgraph extracted by the topic propagation network graph, starting from any subgraph source node, forward searching is carried out by taking the forwarding strength as an edge weight and the shortest weighted path from the source node as a filtering function until all nodes in the subgraph are accessed, and searching ending nodes are recorded; Step 3, starting from other sub-image source nodes which are not searched forward in sequence, executing step 2, forming a sequence obtained by searching a plurality of sub-image source nodes, forming a connected component sequence set comprising a starting time step and a duration, combining the same connected components obtained by searching all sub-image source nodes, wherein the starting time is the minimum starting time, and the duration is the sum of the duration of all the connected components in all the sequences; Step 4, for each subgraph, starting reverse search from the recorded end node, searching by taking the forwarding strength as an edge weight and taking the shortest weighted path from the source node as a filtering function until returning to the source node, filtering out a ring sequence on the topic propagation network graph, and recording the relative time steps of the connection component and the ring start and end in each group of sequences; Step 5, starting from the end nodes of other sub-graph records which are not searched in the reverse direction in turn, executing the step 4, forming a plurality of sequences obtained by searching the source nodes, forming a ring sequence set comprising a starting time step and a duration time, merging the same rings obtained by searching the source nodes of each sub-graph, wherein the starting time is the minimum starting time, and the duration time is the sum of the duration time of each ring structure in each sequence; And 6, extracting a connected component sequence set and the starting time step and duration time of each sequence in the ring sequence set to construct a coordinate pair, combining the coordinate pairs to form a coordinate set, drawing each point in a two-dimensional plane coordinate system, and constructing a fingerprint map of the key topological characteristic of the target topic transmission. Further, the step 1 specifically includes: Aiming at target topic data of an online social platform, a node set V of a topic propagation network diagram is formed according to node account information related to a targ