CN111815028A - Method and device for predicting propagation path of sudden hot spot event - Google Patents
Method and device for predicting propagation path of sudden hot spot event Download PDFInfo
- Publication number
- CN111815028A CN111815028A CN202010525456.5A CN202010525456A CN111815028A CN 111815028 A CN111815028 A CN 111815028A CN 202010525456 A CN202010525456 A CN 202010525456A CN 111815028 A CN111815028 A CN 111815028A
- Authority
- CN
- China
- Prior art keywords
- event
- hot spot
- user information
- propagation path
- information network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 46
- 238000007477 logistic regression Methods 0.000 claims abstract description 18
- 230000006399 behavior Effects 0.000 claims abstract description 13
- 238000004590 computer program Methods 0.000 claims description 16
- 238000006243 chemical reaction Methods 0.000 claims description 8
- 238000003860 storage Methods 0.000 claims description 6
- 238000005054 agglomeration Methods 0.000 claims description 4
- 230000002776 aggregation Effects 0.000 claims description 4
- 230000005540 biological transmission Effects 0.000 abstract description 5
- 238000004458 analytical method Methods 0.000 abstract description 2
- 238000004364 calculation method Methods 0.000 description 18
- 238000010586 diagram Methods 0.000 description 10
- 230000004913 activation Effects 0.000 description 6
- 230000006870 function Effects 0.000 description 6
- 238000012986 modification Methods 0.000 description 5
- 230000004048 modification Effects 0.000 description 5
- 238000010276 construction Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 2
- 229910052799 carbon Inorganic materials 0.000 description 2
- 206010000370 Accident at home Diseases 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012067 mathematical method Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000035515 penetration Effects 0.000 description 1
- 230000000644 propagated effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/10—Text processing
- G06F40/12—Use of codes for handling textual entities
- G06F40/151—Transformation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/01—Social networking
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Economics (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Strategic Management (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Development Economics (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computing Systems (AREA)
- Primary Health Care (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention relates to the technical field of network information analysis, in particular to a method and a device for predicting a propagation path of an emergency hot spot event. The method comprises the following steps: determining an emergency hotspot event according to the social network text; establishing a user information network according to the attention behaviors among the users in the user set; and inputting the user information network into a binary logistic regression model to obtain a predicted propagation path of the sudden hot event. According to the method and the device, the user information network is built according to the attention behaviors among the users, the actual scene of information transmission is closer, the accuracy of predicting the propagation path of the emergency hot spot event is improved, the whole user information network is traversed by using a binary logistic regression model, the propagation node which can possibly forward the emergency hot spot event is obtained, and therefore the prediction of the propagation path of the emergency hot spot event is efficiently realized.
Description
Technical Field
The invention relates to the technical field of network information analysis, in particular to a method and a device for predicting a propagation path of an emergency hot spot event.
Background
With the continuous development of online social networks, more and more netizens use social networks to communicate and exchange. In the social network, information is propagated along the social network formed by the association relationship among the users through the propagation behavior among the users. Due to the characteristics of short text, weak relationship, instantaneity and the like of the social network, netizens can read interesting information on the platform at the first time, so that the information transmission speed is far faster than that of the traditional media, and often, an emergency hot event is quickly spread on the social media before the report of the main stream media and has certain influence on the society.
The quasi-emergency hot spot event is predicted in the social network environment, early warning information can be provided for the government and enterprise departments in real time, and the government and enterprise departments are helped to adopt scientific and effective control and guidance. The propagation path of the predicted information is a hot issue in the field of event prediction, and is the focus of current research.
At present, the prediction of the propagation path of the information is mainly based on models such as an information cascade model, a classification model, a game theory model and the like. Most of the prediction methods consider users as the same or similar individuals, and in a social network, each user has unique characteristics, and the characteristics often determine the propagation path of the emergency hot event, so that the accuracy of predicting the emergency hot event is influenced. In addition, the prediction methods have low computational efficiency, so that the prediction timeliness is poor.
Therefore, how to predict the propagation path of the sudden hot event with high accuracy and good timeliness is a technical problem which needs to be solved urgently at present.
Disclosure of Invention
The invention aims to provide a method and a device for predicting a propagation path of an emergency hot spot event, which can realize the prediction of the propagation path of the emergency hot spot event with high accuracy and good timeliness.
The embodiment of the invention provides the following scheme:
in a first aspect, an embodiment of the present invention provides a method for predicting a propagation path of an emergency hot spot event, where the method includes:
determining an emergency hotspot event according to the social network text;
establishing a user information network according to the attention behaviors among the users in the user set; the issuing user of the emergency hotspot event in the user set is a central node of the user information network; the other users in the user set except the issuing user of the sudden hot spot event are neighbor nodes of the user information network;
inputting the user information network into a binary logistic regression model to obtain a predicted propagation path of the emergency hot spot event;
wherein the binary logistic regression model comprises:
an input layer to receive the user information network;
the first computing layer is used for computing first forwarding probabilities of all first-degree neighbor nodes for forwarding the sudden hot spot events; the first-degree neighbor node is a neighbor node which is in one-way connection with the central node from the first-degree neighbor node in the user information network;
the second computing layer is used for computing second forwarding probabilities of forwarding the sudden hot spot events by other neighbor nodes; the other neighbor nodes are neighbor nodes except the first-degree neighbor node in the user information network;
a propagation node determining layer for determining a set of propagation nodes in the user information network; the propagation node set comprises a first-degree neighbor node with the first forwarding probability being greater than a first numerical value and other neighbor nodes with the second forwarding probability being greater than a second numerical value;
a predicted propagation path generation layer for generating a predicted propagation path of the emergency hot spot event according to a propagation node in the user information network;
and the output layer is used for outputting the predicted propagation path of the sudden hot spot event.
In a possible embodiment, the determining an emergency hotspot event according to the social network text includes:
extracting key words from the social network text by using an Encoder-Decoder model frame to obtain a plurality of groups of key word sets;
calculating semantic similarity among the plurality of groups of keyword sets by using the Hamming distance to obtain the semantic similarity among the plurality of groups of keyword sets;
according to the semantic similarity among the plurality of groups of keyword sets, event semantic similarity clustering is carried out by adopting a bottom-up agglomeration method to obtain an event set;
calculating the text number growth rate of each event in the event set;
and determining the sudden hot spot event according to the text number increase rate of each event in the event set.
In a possible embodiment, the extracting keywords in the social network text by using an Encoder-Decoder model framework to obtain a plurality of groups of keyword sets includes:
step 1, selecting a target text from the social network text;
step 2, converting the target text into a word sequence by using an Encoder code;
step 3, carrying out nonlinear conversion on the word sequence to obtain the middle semantics of the word sequence;
step 4, generating a keyword set of the target text by using Decoder decoding according to the middle semantics of the word sequence and the previously generated keyword set;
and 5, selecting a new target text from the social network text, and returning to the step 2.
In a possible embodiment, the calculating the semantic similarity between the sets of keywords by using the hamming distance to obtain the semantic similarity between the sets of keywords includes:
obtaining the sememes of each group of keyword sets in the plurality of groups of keyword sets by using a known network;
according to the sememe of each group of keyword sets, creating a sememe set corresponding to each group of keyword sets;
and calculating the semantic similarity among the plurality of groups of keyword sets according to the semantic set corresponding to each group of keyword sets.
In a possible embodiment, the clustering event semantic similarity according to semantic similarity between the plurality of groups of keyword sets by using bottom-up clustering to obtain an event set includes:
step i, taking each group of keyword sets as an event, and constructing a clustering event set;
step ii, according to the semantic similarity between any two events in the cluster event set, constructing a semantic similarity set; the semantic similarity between any two events is the semantic similarity between the keyword sets corresponding to any two events;
step iii, judging whether the maximum value in the semantic similarity set is larger than a third numerical value;
step iv, if not, taking the clustering event set as the event set;
step v, if yes, clustering the two events corresponding to the maximum value into a new event, and updating the clustering event set;
step vi, setting the semantic similarity between the new event and any event in the updated clustered event set as the mean of the semantic similarities between the two events corresponding to the maximum value and the any event respectively, and returning to step ii.
In one possible embodiment, the calculating a text number growth rate of each event in the event set includes:
acquiring the text number of each event in the event set in a plurality of time stamps by using a time sliding window;
and calculating the text number growth rate of each event in the event set according to the text number of each event in a plurality of timestamps.
In a possible embodiment, the determining the emergency hot spot event according to the text number increase rate of each event in the event set includes:
selecting texts with the text number growth rate larger than a fourth numerical value from the event set, and constructing a suspected emergency hot spot event set;
and determining the sudden hot spot event according to the suspected sudden hot spot event set.
In a second aspect, an embodiment of the present invention provides an apparatus for predicting a propagation path of an emergency hot spot event, including:
the emergency hot event determining module is used for determining an emergency hot event according to the social network text;
the user information network establishing module is used for establishing a user information network according to the attention behaviors among the users in the user set; the issuing user of the emergency hotspot event in the user set is a central node of the user information network; the other users in the user set except the issuing user of the sudden hot spot event are neighbor nodes of the user information network;
the predicted propagation path acquisition module is used for inputting the user information network into a binary logistic regression model to acquire a predicted propagation path of the sudden hot event;
wherein the binary logistic regression model comprises:
an input layer to receive the user information network;
the first computing layer is used for computing first forwarding probabilities of all first-degree neighbor nodes for forwarding the sudden hot spot events; the first-degree neighbor node is a neighbor node which is in one-way connection with the central node from the first-degree neighbor node in the user information network;
the second computing layer is used for computing second forwarding probabilities of forwarding the sudden hot spot events by other neighbor nodes; the other neighbor nodes are neighbor nodes except the first-degree neighbor node in the user information network;
a propagation node determining layer for determining a set of propagation nodes in the user information network; the propagation node set comprises a first-degree neighbor node with the first forwarding probability being greater than a first numerical value and other neighbor nodes with the second forwarding probability being greater than a second numerical value;
a predicted propagation path generation layer for generating a predicted propagation path of the emergency hot spot event according to a propagation node in the user information network;
and the output layer is used for outputting the predicted propagation path of the sudden hot spot event.
In a possible embodiment, the emergency hot spot event determining module includes:
the keyword set acquisition module is used for extracting keywords from the social network text by using an Encoder-Decoder model frame to obtain a plurality of groups of keyword sets;
the keyword set semantic similarity acquisition module is used for calculating semantic similarity among the plurality of groups of keyword sets by utilizing the Hamming distance and acquiring the semantic similarity among the plurality of groups of keyword sets;
the event set acquisition module is used for clustering event semantic similarity by adopting a bottom-up agglomeration method according to the semantic similarity among the plurality of groups of keyword sets to acquire an event set;
the first calculation module is used for calculating the text number growth rate of each event in the event set;
and the first determining module is used for determining the sudden hot spot event according to the text number increase rate of each event in the event set.
In a possible embodiment, the keyword set obtaining module includes:
the target text selection module is used for selecting a target text from the social network text;
the word sequence conversion module is used for converting the target text into a word sequence by using Encoder coding;
the middle semantic obtaining module is used for carrying out nonlinear conversion on the word sequence to obtain the middle semantic of the word sequence;
a keyword set generating module, configured to generate a keyword set of the target text by using Decoder decoding according to the intermediate semantics of the word sequence and a previously generated keyword set;
and the target text updating module is used for selecting a new target text from the social network text and returning the new target text to the word sequence conversion module.
In a possible embodiment, the keyword set semantic similarity obtaining module includes:
the semantic source obtaining module is used for obtaining the semantic source of each group of keyword sets in the plurality of groups of keyword sets by using a knowledge network;
an original set obtaining module, configured to create an original set corresponding to each group of keyword sets according to the original of each group of keyword sets;
and the keyword set semantic similarity calculation module is used for calculating the semantic similarity among the plurality of groups of keyword sets according to the semantic set corresponding to each group of keyword sets.
In one possible embodiment, the event set obtaining module includes:
the cluster event set construction module is used for constructing a cluster event set by taking each group of keyword sets as an event;
the semantic similarity set construction module is used for constructing a semantic similarity set according to the semantic similarity between any two events in the clustering event set; the semantic similarity between any two events is the semantic similarity between the keyword sets corresponding to any two events;
the first judgment module is used for judging whether the maximum value in the semantic similarity set is larger than a third numerical value or not;
an event set obtaining module, configured to use the clustered event set as the event set when a maximum value in the semantic similarity set is not greater than a third value;
a clustering event set updating module, configured to cluster, when a maximum value in the semantic similarity set is greater than a third value, two events corresponding to the maximum value into a new event, and update the clustering event set;
and the semantic similarity updating module is used for setting the semantic similarity between the new event and any event in the updated clustered event set as the mean value of the semantic similarities between the two events corresponding to the maximum value and the any event respectively, and returning the mean value to the semantic similarity set building module.
In one possible embodiment, the first calculation module includes:
the text number counting module is used for acquiring the text number of each event in the event set in a plurality of time stamps by utilizing a time sliding window;
and the second calculation module is used for calculating the text number increase rate of each event in the event set according to the text number of each event in a plurality of timestamps.
In one possible embodiment, the first determining module includes:
the suspected emergency hot event set building module is used for selecting texts with the text number growth rate larger than a fourth numerical value from the event set and building a suspected emergency hot event set;
and the second determining module is used for determining the unexpected hot spot event according to the suspected unexpected hot spot event set.
In a third aspect, an embodiment of the present invention provides a device for predicting a propagation path of an emergency hot spot event, including:
a memory for storing a computer program;
a processor configured to execute the computer program to implement the steps of the method for predicting a propagation path of an emergency hot spot event according to any one of the first aspect.
In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, where the computer program is executed by a processor to implement the steps of the method for predicting a propagation path of an emergency hot spot event according to any one of the first aspect.
Compared with the prior art, the invention has the following advantages and beneficial effects:
according to the method and the device, the user information network is built according to the attention behaviors among the users, the actual scene of information transmission is closer, the accuracy of predicting the propagation path of the emergency hot spot event is improved, the whole user information network is traversed by using a binary logistic regression model, the propagation node which can possibly forward the emergency hot spot event is obtained, and therefore the prediction of the propagation path of the emergency hot spot event is efficiently realized.
Drawings
In order to more clearly illustrate the embodiments of the present specification or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present specification, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a flowchart of a method for predicting a propagation path of an emergency hot spot event according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of a device for predicting a propagation path of an emergency hot spot event according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, rather than all embodiments, and all other embodiments obtained by those skilled in the art based on the embodiments of the present invention belong to the scope of protection of the embodiments of the present invention.
Referring to fig. 1, fig. 1 is a flowchart illustrating a method for predicting a propagation path of an emergency hot spot event according to an embodiment of the present invention, including steps 11 to 13.
And step 11, determining the emergency hot spot event according to the social network text.
Specifically, the social network text is generated by a plurality of users publishing information and forwarding information within a period of time, and can be acquired from social network applications such as microblog and facebook through a crawler technology.
Specifically, whether an emergency hot spot event, such as an emergency of a certain star, an international situation change, an emergency major event or an accident at home and abroad, exists in the social network text can be determined through experience.
And step 12, establishing a user information network according to the attention behaviors among the users in the user set. And issuing the emergency hotspot event in the user set, wherein the issuing user of the emergency hotspot event in the user set is a central node of the user information network. And the other users in the user set except the issuing user of the sudden hot spot event are neighbor nodes of the user information network.
Specifically, in the social network, a user who pays attention to an emergency hotspot event publishing user can learn the emergency hotspot event more quickly, and most probably forwards and diffuses the emergency hotspot event at the earliest time.
Specifically, the in-degree of any node in the user information network is the attention amount of the user corresponding to the node, and the out-degree of any node is the attention amount of the user corresponding to the node.
And step 13, inputting the user information network into a binary logistic regression model to obtain the predicted propagation path of the sudden hot spot event.
Wherein the binary logistic regression model comprises:
an input layer for receiving the user information network.
And the first computing layer is used for computing a first forwarding probability of forwarding the sudden hot spot event by each first-degree neighbor node. The first-degree neighbor node is a neighbor node which is in one-way connection with the central node from the first-degree neighbor node in the user information network.
Specifically, a better calculation method of the first forwarding probability is provided.
Firstly, defining a superior neighbor node, and if a one-way line pointing to another neighbor node from a certain neighbor node exists between the certain neighbor node and the other neighbor node, making the other neighbor node a superior neighbor node of the certain neighbor node. In other words, if a user corresponding to a certain neighbor node concerns a user corresponding to another neighbor node in the social network, the other neighbor node is a higher-level neighbor node of the certain neighbor node.
Calculating a first forwarding probability P of each one-degree neighbor node for forwarding the sudden hot spot event according to the following formularetweet(n1)：
Wherein c is a central node and n1Is any one-degree neighbor node, miIs a one-degree neighbor node n1The ith upper-level neighbor node of (1),
for the central node c to one degree neighbor node n1Forwarding the activation probability, x, of the sudden hot event1The influence of a central node c is Deg, the degree of entry of the central node c is Act, the activity of the central node c is Hot, the previous degree of the social network of the user corresponding to the central node c is Hot, and p1、p2And p3Set weights, x, corresponding to Deg, Act and Hot respectively2Is a central node c and a first-degree neighbor node n1Intimacy between, x3Is a one-degree neighbor node n1Whether x is present in the @ information of the central node c4Is a one-degree neighbor node n1Degree of interest in sudden hot spot events, λ1、λ2、λ3And λ4Are respectively x1、x2、x3And x4The weight is set.
Is a one-degree neighbor node n1I upper neighbor node miTo one-degree neighbor node n1And forwarding the activation probability of the sudden hot spot event.
A general calculation upper-level neighbor node m is providedjFor its subordinate neighbor node nkForwarding activation probability of the sudden hot spot event
The formula for the calculation of (a) is,
the specific calculation of (a) can be carried out by applying the following specific formula.
Wherein x is1The influence of a central node c is Deg, the degree of entry of the central node c is Act, the activity of the central node c is Hot, the previous degree of the social network of the user corresponding to the central node c is Hot, and p1、p2And p3Respectively corresponding to the set weights of Deg, Act and Hot,
is a superior neighbor node mjThe influence of (a) on the magnetic field,
is a superior neighbor node mjThe degree of penetration of the (c) is,
is a superior neighbor node mjThe activity of the active carbon is higher than that of the active carbon,
is a superior neighbor node mjThe past popularity of the corresponding user's social network,
and
are respectively as
And
the corresponding weight value is set, and the weight value,
is a superior neighbor node mjAnd a lower neighbor node nkThe degree of intimacy between the two,
for lower neighbor node nkWhether it appears in the upper neighbor node mjIn the @ information of (a) to (b),
for lower neighbor node nkThe degree of interest in the incident of a hot spot,
and
are respectively x1、
And
the weight is set.
When the first forwarding probability is calculated, the intimacy degree between users and the interest degree of the users in the information are used as the influence factors for calculating the activation probability, the influence of the users in information transmission is further described, and the accuracy of prediction is improved. And further, a central node (such as a node with a large influence, such as a star) is taken into an influence factor during iterative propagation, and the influence of the original publisher is considered. The influence of the user when encountering the event is reduced to the maximum extent, the actual situation is fitted, and the prediction accuracy is further improved. When individual propagation behaviors are predicted, the activation behaviors of the active nodes to the inactive nodes are regarded as independent events in combination with actual conditions, the independent events do not affect each other, and the forwarding probability of the inactive nodes is calculated by utilizing the independence of the events. And a mathematical method model in probability theory is used for restoring the real situation, so that the prediction accuracy is improved, and the error is reduced.
And the second calculation layer is used for calculating second forwarding probabilities of forwarding the sudden hot spot event by other neighbor nodes.
And the other neighbor nodes are neighbor nodes except the first-degree neighbor node in the user information network.
Specifically, a better calculation method of the second forwarding probability is provided.
Here with other neighbor nodes nkFor example, the second forwarding probability P is illustratedretweet(nk) In which the neighbor node mjFor other neighbor nodes nkThe specific calculation formula of the jth superior neighbor node is as follows:
wherein the content of the first and second substances,
for other neighbor nodes nkJ upper neighbor node mjFor other neighbor nodes nkAnd forwarding the activation probability of the sudden hot spot event, wherein the specific calculation process is detailed above.
After calculating, the second forwarding probability Pretweet(nk) Then, the user information network can be usedAnd traversing other neighbor nodes to finally obtain a second forwarding probability of forwarding the sudden hot spot event by each other neighbor node.
A propagation node determining layer for determining a set of propagation nodes in the user information network. Wherein the set of propagation nodes includes one-degree neighbor nodes having the first forwarding probability greater than a first value and other neighbor nodes having the second forwarding probability greater than a second value.
Specifically, the first numerical value and the second numerical value are both 0.5, and if the first forwarding probability of a certain degree of neighbor nodes is greater than 0.5, the node will forward the sudden hot spot event issued by the central node; if the second forwarding probability of some other neighbor node is greater than 0.5, the node forwards the sudden hot spot event forwarded by the neighbor node at the upper level.
And a predicted propagation path generation layer for generating a predicted propagation path of the emergency hot spot event according to the propagation node in the user information network.
And the output layer is used for outputting the predicted propagation path of the sudden hot spot event.
Specifically, a cost function of the binary logistic regression model is further provided, so that relevant parameters in the binary logistic regression model can be quickly adjusted through a training set, and the propagation path can be accurately predicted.
The cost function is:
in a possible embodiment, in order to determine the emergency hot spot event more accurately and efficiently, the invention further provides a better scheme, and the specific scheme is as follows:
the determining of the emergency hot spot event according to the social network text comprises steps 21 to 25.
And step 21, extracting key words from the social network text by using an Encoder-Decoder model frame to obtain a plurality of groups of key word sets.
Specifically, the method of this step includes steps 211 to 215.
Step 211, selecting a target text from the social network text.
Specifically, the target text W is an information text published or forwarded by a certain user at the ith time in the social network text.
And 212, converting the target text into a word sequence by using the Encoder coding.
Specifically, the word sequence of the converted target text W is represented as: w ═<w1,w2...wm>。
Step 213, performing nonlinear transformation on the word sequence to obtain the middle semantics of the word sequence.
Specifically, the intermediate semantic M corresponding to the target text W is expressed as: m ═ f (w)1,w2...wm)。
And 214, generating a keyword set of the target text by using Decoder decoding according to the intermediate semantics of the word sequence and the previously generated keyword set.
Since the keyword set k corresponding to each text before the i-th time can be obtained through steps 211 to 215 before1,k2...ki-1Then the set k of keywords generated at the ith timeiExpressed as: k is a radical ofi＝g(M,k1,k2...ki-1)。
Step 215, selecting a new target text from the social network text, and returning to step 212.
Specifically, in the step, a new target text is selected from the social network text again, so that the social network text is traversed through the steps, and all keyword sets are obtained.
And step 22, calculating the semantic similarity among the plurality of groups of keyword sets by using the Hamming distance, and acquiring the semantic similarity among the plurality of groups of keyword sets.
Specifically, the method for implementing this step includes steps 221 to 223.
For convenience of explanation of the scheme, two sets of keyword sets K are used in steps 221 to 223x(tx1,tx2...txp) And Ky(ty1,ty2...tyq) By way of example, a process for calculating semantic similarity between two sets of keywords is illustrated, where txi，tyiIs the tf-idf value of the keyword.
Step 221, obtaining the sememes of each group of keyword sets in the plurality of groups of keyword sets by using the known network.
Specifically, since the keyword set cannot be directly converted into the codeword set, the lexeme is required to be used as an intermediate link.
Step 222, according to the sememes of each group of keyword sets, creating the sememe sets corresponding to each group of keyword sets.
Specifically, the sememes of all keywords in the keyword set are obtained based on the known network, and two keyword sets K are further createdx(tx1,tx2...txp) And Ky(ty1,ty2...tyq) Corresponding set of sememes Dx(dx1,dx2...dxs)，Dy(dy1,dy2...dys). Wherein d isxi，dyiEach of the primitive frequencies is an integer of 0 or more, and is set to 0 as an initial value. Traversing corresponding original meaning set D according to the original meanings of the keywords in the keyword setx，Dy. Each pass will be DxMiddle corresponds to KxThe corresponding frequency of the keyword sememe is increased by 1, DyThe same is true.
Step 223, calculating semantic similarity between the plurality of groups of keyword sets according to the semantic set corresponding to each group of keyword sets.
Specifically, the finally obtained semantic set Dx，DyThe non-zero element in the code word is set to be 1 to obtain the corresponding code word set Mx，My. Because the frequency of the sememes in the sememe set has high and low scores, the sememe set is divided into a plurality of sememe subsets according to the frequency, and the sememe subsets are called as Hamming sets. Then, setting weight W for each Hamming set according to the segmentation principlei。
Specifically, the semantic similarity between two sets of keyword sets must be calculated first, and the calculation method is as follows:
wherein H (M)x,My) Is a set of code words Mx，MyS is the dimension of the codeword set.
The semantic similarity between the two sets of keywords is:
where n is the number of divided hamming sets.
And step 23, clustering event semantic similarity by adopting a bottom-up clustering method according to the semantic similarity among the plurality of groups of keyword sets to obtain an event set.
Specifically, the method for implementing this step includes steps 231 to 236.
And 231, taking each group of keyword sets as an event, and constructing a clustering event set.
Specifically, each set of keywords is from a text (e.g., microblog text), which can represent a separate event.
Step 232, according to the semantic similarity between any two events in the cluster event set, a semantic similarity set is constructed. And the semantic similarity between any two events is the semantic similarity between the keyword sets corresponding to any two events.
Specifically, since each set of keywords is regarded as an event, the semantic similarity between events is equal to the semantic similarity between corresponding sets of keywords.
Specifically, the invention uses an nxn similarity matrix as a semantic similarity set, where n is the total number of events in the event cluster set, a certain element in the similarity matrix represents the semantic similarity between certain two events, and the main diagonal element of the similarity matrix is set to 0.
In step 233, it is determined whether the maximum value in the semantic similarity set is greater than a third value.
Specifically, here the third value is 0.8.
And step 234, if not, taking the clustering event set as the event set.
Specifically, if the maximum value in the semantic similarity set is not greater than 0.8, it indicates that clustering cannot be performed between events in the event clustering set, and the clustering operation is terminated to obtain the final event set.
And 235, if yes, clustering the two events corresponding to the maximum value into a new event, and updating the clustering event set.
Specifically, if the maximum value in the semantic similarity set is greater than 0.8, it indicates that two events corresponding to the maximum value in the semantic similarity set are very similar and can be regarded as the same event, so that the two events are clustered into a new event, the new event is put into the clustered event set, and the two events are deleted from the clustered event set, thereby completing the update of the clustered event set.
Step 236, setting the semantic similarity between the new event and any event in the updated clustered event set as the mean of the semantic similarities between the two events corresponding to the maximum value and the any event respectively, and returning to step 232.
Specifically, since the clustering event set is updated, the semantic similarity between the events in the clustering event set needs to be recalculated, and the semantic similarity set can be updated only by calculating the semantic similarity between a new event generated by a newly added cluster and other events. The calculation method adopted by the invention is that the average value of the original semantic similarity between the two events corresponding to the maximum value and any event is calculated, and the obtained average value is used as the semantic similarity between the new event and any event.
And 24, calculating the text number growth rate of each event in the event set.
Specifically, the method for implementing this step includes steps 241 to 242.
And 241, acquiring the number of texts of each event in the event set in a plurality of timestamps by using a time sliding window.
And 242, calculating the text number increase rate of each event in the event set according to the text number of each event in a plurality of timestamps.
And step 25, determining the sudden hot spot event according to the text number increase rate of each event in the event set.
Specifically, the method for implementing this step includes steps 251 to 252.
And 251, selecting texts with the text number growth rate larger than a fourth numerical value from the event set, and constructing a suspected emergency hot spot event set.
Step 252, determining the hot burst event according to the suspected hot burst event set.
Specifically, any event in the suspected emergency event set may be selected as the emergency event, or the event with the largest text number increase rate may be selected as the emergency event.
Based on the same inventive concept as the method, an embodiment of the present invention further provides a device for predicting a propagation path of an unexpected hot spot event, as shown in fig. 2, which is a schematic structural diagram of the embodiment of the device, and the device includes:
the emergency hot spot event determining module 31 is configured to determine an emergency hot spot event according to the social network text;
a user information network establishing module 32, configured to establish a user information network according to the attention behavior among the users in the user set; the issuing user of the emergency hotspot event in the user set is a central node of the user information network; the other users in the user set except the issuing user of the sudden hot spot event are neighbor nodes of the user information network;
a predicted propagation path obtaining module 33, configured to input the user information network into a binary logistic regression model, and obtain a predicted propagation path of the emergency hot spot event;
wherein the binary logistic regression model comprises:
an input layer to receive the user information network;
the first computing layer is used for computing first forwarding probabilities of all first-degree neighbor nodes for forwarding the sudden hot spot events; the first-degree neighbor node is a neighbor node which is in one-way connection with the central node from the first-degree neighbor node in the user information network;
the second computing layer is used for computing second forwarding probabilities of forwarding the sudden hot spot events by other neighbor nodes; the other neighbor nodes are neighbor nodes except the first-degree neighbor node in the user information network;
a propagation node determining layer for determining a set of propagation nodes in the user information network; the propagation node set comprises a first-degree neighbor node with the first forwarding probability being greater than a first numerical value and other neighbor nodes with the second forwarding probability being greater than a second numerical value;
a predicted propagation path generation layer for generating a predicted propagation path of the emergency hot spot event according to a propagation node in the user information network;
and the output layer is used for outputting the predicted propagation path of the sudden hot spot event.
In a possible embodiment, the emergency hot spot event determining module 31 includes:
the keyword set acquisition module is used for extracting keywords from the social network text by using an Encoder-Decoder model frame to obtain a plurality of groups of keyword sets;
the keyword set semantic similarity acquisition module is used for calculating semantic similarity among the plurality of groups of keyword sets by utilizing the Hamming distance and acquiring the semantic similarity among the plurality of groups of keyword sets;
the event set acquisition module is used for clustering event semantic similarity by adopting a bottom-up agglomeration method according to the semantic similarity among the plurality of groups of keyword sets to acquire an event set;
the first calculation module is used for calculating the text number growth rate of each event in the event set;
and the first determining module is used for determining the sudden hot spot event according to the text number increase rate of each event in the event set.
In a possible embodiment, the keyword set obtaining module includes:
the target text selection module is used for selecting a target text from the social network text;
the word sequence conversion module is used for converting the target text into a word sequence by using Encoder coding;
the middle semantic obtaining module is used for carrying out nonlinear conversion on the word sequence to obtain the middle semantic of the word sequence;
a keyword set generating module, configured to generate a keyword set of the target text by using Decoder decoding according to the intermediate semantics of the word sequence and a previously generated keyword set;
and the target text updating module is used for selecting a new target text from the social network text and returning the new target text to the word sequence conversion module.
In a possible embodiment, the keyword set semantic similarity obtaining module includes:
the semantic source obtaining module is used for obtaining the semantic source of each group of keyword sets in the plurality of groups of keyword sets by using a knowledge network;
an original set obtaining module, configured to create an original set corresponding to each group of keyword sets according to the original of each group of keyword sets;
and the keyword set semantic similarity calculation module is used for calculating the semantic similarity among the plurality of groups of keyword sets according to the semantic set corresponding to each group of keyword sets.
In one possible embodiment, the event set obtaining module includes:
the cluster event set construction module is used for constructing a cluster event set by taking each group of keyword sets as an event;
the semantic similarity set construction module is used for constructing a semantic similarity set according to the semantic similarity between any two events in the clustering event set; the semantic similarity between any two events is the semantic similarity between the keyword sets corresponding to any two events;
the first judgment module is used for judging whether the maximum value in the semantic similarity set is larger than a third numerical value or not;
an event set obtaining module, configured to use the clustered event set as the event set when a maximum value in the semantic similarity set is not greater than a third value;
a clustering event set updating module, configured to cluster, when a maximum value in the semantic similarity set is greater than a third value, two events corresponding to the maximum value into a new event, and update the clustering event set;
and the semantic similarity updating module is used for setting the semantic similarity between the new event and any event in the updated clustered event set as the mean value of the semantic similarities between the two events corresponding to the maximum value and the any event respectively, and returning the mean value to the semantic similarity set building module.
In one possible embodiment, the first calculation module includes:
the text number counting module is used for acquiring the text number of each event in the event set in a plurality of time stamps by utilizing a time sliding window;
and the second calculation module is used for calculating the text number increase rate of each event in the event set according to the text number of each event in a plurality of timestamps.
In one possible embodiment, the first determining module includes:
the suspected emergency hot event set building module is used for selecting texts with the text number growth rate larger than a fourth numerical value from the event set and building a suspected emergency hot event set;
and the second determining module is used for determining the unexpected hot spot event according to the suspected unexpected hot spot event set.
Based on the same inventive concept as that in the foregoing embodiments, an embodiment of the present invention further provides a device for predicting a propagation path of an unexpected hotspot event, including a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of any one of the foregoing methods when executing the program.
Based on the same inventive concept as in the previous embodiments, embodiments of the present invention further provide a computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, implements the steps of any of the methods described above.
The technical scheme provided by the embodiment of the invention at least has the following technical effects or advantages:
according to the method and the device, the user information network is built according to the attention behaviors among the users, the actual scene of information transmission is closer, the accuracy of predicting the propagation path of the emergency hot spot event is improved, the whole user information network is traversed by using a binary logistic regression model, the propagation node which can possibly forward the emergency hot spot event is obtained, and therefore the prediction of the propagation path of the emergency hot spot event is efficiently realized.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (modules, systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all such alterations and modifications as fall within the scope of the invention.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.
Claims (10)
1. A method for predicting a propagation path of an emergency hot spot event, the method comprising:
determining an emergency hotspot event according to the social network text;
establishing a user information network according to the attention behaviors among the users in the user set; the issuing user of the emergency hotspot event in the user set is a central node of the user information network; the other users in the user set except the issuing user of the sudden hot spot event are neighbor nodes of the user information network;
inputting the user information network into a binary logistic regression model to obtain a predicted propagation path of the emergency hot spot event;
wherein the binary logistic regression model comprises:
an input layer to receive the user information network;
the first computing layer is used for computing first forwarding probabilities of all first-degree neighbor nodes for forwarding the sudden hot spot events; the first-degree neighbor node is a neighbor node which is in one-way connection with the central node from the first-degree neighbor node in the user information network;
the second computing layer is used for computing second forwarding probabilities of forwarding the sudden hot spot events by other neighbor nodes; the other neighbor nodes are neighbor nodes except the first-degree neighbor node in the user information network;
a propagation node determining layer for determining a set of propagation nodes in the user information network; the propagation node set comprises a first-degree neighbor node with the first forwarding probability being greater than a first numerical value and other neighbor nodes with the second forwarding probability being greater than a second numerical value;
a predicted propagation path generation layer for generating a predicted propagation path of the emergency hot spot event according to a propagation node in the user information network;
and the output layer is used for outputting the predicted propagation path of the sudden hot spot event.
2. The method for predicting the propagation path of the emergency hot spot event according to claim 1, wherein the determining the emergency hot spot event according to the social network text comprises:
extracting key words from the social network text by using an Encoder-Decoder model frame to obtain a plurality of groups of key word sets;
calculating semantic similarity among the plurality of groups of keyword sets by using the Hamming distance to obtain the semantic similarity among the plurality of groups of keyword sets;
according to the semantic similarity among the plurality of groups of keyword sets, event semantic similarity clustering is carried out by adopting a bottom-up agglomeration method to obtain an event set;
calculating the text number growth rate of each event in the event set;
and determining the sudden hot spot event according to the text number increase rate of each event in the event set.
3. The method for predicting propagation paths of unexpected hot events according to claim 2, wherein the extracting keywords in the social network text by using an encor-Decoder model framework to obtain several groups of keyword sets comprises:
step 1, selecting a target text from the social network text;
step 2, converting the target text into a word sequence by using an Encoder code;
step 3, carrying out nonlinear conversion on the word sequence to obtain the middle semantics of the word sequence;
step 4, generating a keyword set of the target text by using Decoder decoding according to the middle semantics of the word sequence and the previously generated keyword set;
and 5, selecting a new target text from the social network text, and returning to the step 2.
4. The method for predicting the propagation path of the unexpected hotspot event according to claim 3, wherein the calculating the semantic similarity between the plurality of groups of keyword sets by using the hamming distance to obtain the semantic similarity between the plurality of groups of keyword sets comprises:
obtaining the sememes of each group of keyword sets in the plurality of groups of keyword sets by using a known network;
according to the sememe of each group of keyword sets, creating a sememe set corresponding to each group of keyword sets;
and calculating the semantic similarity among the plurality of groups of keyword sets according to the semantic set corresponding to each group of keyword sets.
5. The method for predicting the propagation path of the unexpected hotspot event according to claim 4, wherein the step of clustering the semantic similarity of the events by using a bottom-up clustering method according to the semantic similarity among the plurality of groups of keyword sets to obtain the event set comprises:
step i, taking each group of keyword sets as an event, and constructing a clustering event set;
step ii, according to the semantic similarity between any two events in the cluster event set, constructing a semantic similarity set; the semantic similarity between any two events is the semantic similarity between the keyword sets corresponding to any two events;
step iii, judging whether the maximum value in the semantic similarity set is larger than a third numerical value;
step iv, if not, taking the clustering event set as the event set;
step v, if yes, clustering the two events corresponding to the maximum value into a new event, and updating the clustering event set;
step vi, setting the semantic similarity between the new event and any event in the updated clustered event set as the mean of the semantic similarities between the two events corresponding to the maximum value and the any event respectively, and returning to step ii.
6. The method according to claim 5, wherein the calculating a text number increase rate of each event in the event set comprises:
acquiring the text number of each event in the event set in a plurality of time stamps by using a time sliding window;
and calculating the text number growth rate of each event in the event set according to the text number of each event in a plurality of timestamps.
7. The method according to claim 6, wherein the determining the sudden hot event according to the text number increase rate of each event in the event set comprises:
selecting texts with the text number growth rate larger than a fourth numerical value from the event set, and constructing a suspected emergency hot spot event set;
and determining the sudden hot spot event according to the suspected sudden hot spot event set.
8. An apparatus for predicting a propagation path of an emergency hot spot event, comprising:
the emergency hot event determining module is used for determining an emergency hot event according to the social network text;
the user information network establishing module is used for establishing a user information network according to the attention behaviors among the users in the user set; the issuing user of the emergency hotspot event in the user set is a central node of the user information network; the other users in the user set except the issuing user of the sudden hot spot event are neighbor nodes of the user information network;
the predicted propagation path acquisition module is used for inputting the user information network into a binary logistic regression model to acquire a predicted propagation path of the sudden hot event;
wherein the binary logistic regression model comprises:
an input layer to receive the user information network;
the first computing layer is used for computing first forwarding probabilities of all first-degree neighbor nodes for forwarding the sudden hot spot events; the first-degree neighbor node is a neighbor node which is in one-way connection with the central node from the first-degree neighbor node in the user information network;
the second computing layer is used for computing second forwarding probabilities of forwarding the sudden hot spot events by other neighbor nodes; the other neighbor nodes are neighbor nodes except the first-degree neighbor node in the user information network;
a propagation node determining layer for determining a set of propagation nodes in the user information network; the propagation node set comprises a first-degree neighbor node with the first forwarding probability being greater than a first numerical value and other neighbor nodes with the second forwarding probability being greater than a second numerical value;
a predicted propagation path generation layer for generating a predicted propagation path of the emergency hot spot event according to a propagation node in the user information network;
and the output layer is used for outputting the predicted propagation path of the sudden hot spot event.
9. A prediction apparatus for a propagation path of an emergency hotspot event, comprising:
a memory for storing a computer program;
a processor for executing the computer program to carry out the steps of the method of any one of claims 1 to 7.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, is adapted to carry out the steps of the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010525456.5A CN111815028A (en) | 2020-06-10 | 2020-06-10 | Method and device for predicting propagation path of sudden hot spot event |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010525456.5A CN111815028A (en) | 2020-06-10 | 2020-06-10 | Method and device for predicting propagation path of sudden hot spot event |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111815028A true CN111815028A (en) | 2020-10-23 |
Family
ID=72846483
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010525456.5A Pending CN111815028A (en) | 2020-06-10 | 2020-06-10 | Method and device for predicting propagation path of sudden hot spot event |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111815028A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112686765A (en) * | 2020-12-09 | 2021-04-20 | 天津大学 | Information propagation rule mining method based on social network |
| CN114648025A (en) * | 2022-05-18 | 2022-06-21 | 国网浙江省电力有限公司信息通信分公司 | Power grid data processing method and system based on multi-dimensional evolution diagram in power field |
| CN116719936A (en) * | 2023-06-15 | 2023-09-08 | 湖北大学 | Network unreliable information early detection method based on ensemble learning |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120158630A1 (en) * | 2010-12-17 | 2012-06-21 | Microsoft Corporation | Information propagation probability for a social network |
| CN106599243A (en) * | 2016-12-20 | 2017-04-26 | 西南石油大学 | Method and system for predicting micro-blog forwarding path based on micro-blog themes |
| CN108733816A (en) * | 2018-05-21 | 2018-11-02 | 重庆人文科技学院 | A kind of microblogging incident detection method |
| CN108897784A (en) * | 2018-06-08 | 2018-11-27 | 北京航空航天大学 | One emergency event dimensional analytic system based on social media |
| CN108932311A (en) * | 2018-06-20 | 2018-12-04 | 天津大学 | The method of incident detection and prediction |
| CN110543590A (en) * | 2019-01-23 | 2019-12-06 | 北京信息科技大学 | Method for detecting microblog emergency |
| CN111241421A (en) * | 2020-01-14 | 2020-06-05 | 西安电子科技大学 | User forwarding behavior prediction method based on social context information |
-
2020
- 2020-06-10 CN CN202010525456.5A patent/CN111815028A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120158630A1 (en) * | 2010-12-17 | 2012-06-21 | Microsoft Corporation | Information propagation probability for a social network |
| CN106599243A (en) * | 2016-12-20 | 2017-04-26 | 西南石油大学 | Method and system for predicting micro-blog forwarding path based on micro-blog themes |
| CN108733816A (en) * | 2018-05-21 | 2018-11-02 | 重庆人文科技学院 | A kind of microblogging incident detection method |
| CN108897784A (en) * | 2018-06-08 | 2018-11-27 | 北京航空航天大学 | One emergency event dimensional analytic system based on social media |
| CN108932311A (en) * | 2018-06-20 | 2018-12-04 | 天津大学 | The method of incident detection and prediction |
| CN110543590A (en) * | 2019-01-23 | 2019-12-06 | 北京信息科技大学 | Method for detecting microblog emergency |
| CN111241421A (en) * | 2020-01-14 | 2020-06-05 | 西安电子科技大学 | User forwarding behavior prediction method based on social context information |
Non-Patent Citations (1)
| Title |
|---|
| XICHAN NIE等: "Method to Predict Bursty Hot Events On Twitter b ased on User Relationship Network", 《IEEE ACCESS》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112686765A (en) * | 2020-12-09 | 2021-04-20 | 天津大学 | Information propagation rule mining method based on social network |
| CN114648025A (en) * | 2022-05-18 | 2022-06-21 | 国网浙江省电力有限公司信息通信分公司 | Power grid data processing method and system based on multi-dimensional evolution diagram in power field |
| CN116719936A (en) * | 2023-06-15 | 2023-09-08 | 湖北大学 | Network unreliable information early detection method based on ensemble learning |
| CN116719936B (en) * | 2023-06-15 | 2023-12-26 | 湖北大学 | Network unreliable information early detection method based on ensemble learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111815028A (en) | Method and device for predicting propagation path of sudden hot spot event | |
| CN105144227A (en) | Social media impact assessment | |
| CN110659678B (en) | User behavior classification method, system and storage medium | |
| CN110995485B (en) | Social message propagation range prediction method without topological structure | |
| Li et al. | Exploiting microblog conversation structures to detect rumors | |
| CN110297885B (en) | Method, device and equipment for generating real-time event abstract and storage medium | |
| CN112131888A (en) | Method, device and equipment for analyzing semantic emotion and storage medium | |
| Bai et al. | A rumor detection model incorporating propagation path contextual semantics and user information | |
| Liu et al. | Siga: social influence modeling integrating graph autoencoder for rating prediction | |
| EP3871115A1 (en) | Data retrieval | |
| Zhang et al. | Causalrd: A causal view of rumor detection via eliminating popularity and conformity biases | |
| He et al. | Purify and generate: Learning faithful item-to-item graph from noisy user-item interaction behaviors | |
| CN114116692B (en) | Mask and bidirectional model-based missing POI track completion method | |
| CN115438648A (en) | Propagation content enhancement method and system based on user behavior prediction | |
| Wu et al. | Neural gender prediction in microblogging with emotion-aware user representation | |
| Mirakyan | Abcde: Approximating betweenness-centrality ranking with progressive-dropedge | |
| Liu et al. | Prediction model for non-topological event propagation in social networks | |
| Ji et al. | Attention-based graph neural network for news recommendation | |
| Xie | Research on Weibo user behavior system for subjective perception and big data mining technology | |
| Liu et al. | Dual-view hypergraph attention network for news recommendation | |
| Capretz et al. | Improving effort estimation by voting software estimation models | |
| CN104331483A (en) | Method and equipment for detecting area events based on short text data | |
| Tai et al. | Predicting information diffusion using the inter-and intra-path of influence transitivity | |
| Liu et al. | Promoting generalization for exact solvers via adversarial instance augmentation | |
| Liu et al. | A cascade information diffusion prediction model integrating topic features and cross-attention |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20201023 |
|
| RJ01 | Rejection of invention patent application after publication |