CN109783629A - A kind of micro-blog event rumour detection method of amalgamation of global event relation information - Google Patents
A kind of micro-blog event rumour detection method of amalgamation of global event relation information Download PDFInfo
- Publication number
- CN109783629A CN109783629A CN201910042008.7A CN201910042008A CN109783629A CN 109783629 A CN109783629 A CN 109783629A CN 201910042008 A CN201910042008 A CN 201910042008A CN 109783629 A CN109783629 A CN 109783629A
- Authority
- CN
- China
- Prior art keywords
- event
- micro
- blog
- vector
- microblogging
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 33
- 238000005267 amalgamation Methods 0.000 title claims abstract description 14
- 239000013598 vector Substances 0.000 claims abstract description 62
- 238000007781 pre-processing Methods 0.000 claims abstract description 13
- 238000012360 testing method Methods 0.000 claims abstract description 8
- 238000000034 method Methods 0.000 claims description 14
- 239000011159 matrix material Substances 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 9
- 238000000605 extraction Methods 0.000 claims description 5
- 239000012634 fragment Substances 0.000 claims description 5
- 238000005457 optimization Methods 0.000 claims description 4
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 3
- 238000012937 correction Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 description 11
- 239000000284 extract Substances 0.000 description 4
- 238000003062 neural network model Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000005086 pumping Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000644 propagated effect Effects 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Abstract
The present invention relates to a kind of micro-blog event rumour detection methods of amalgamation of global event relation information, provide a text data preprocessing module, the paragraph vector of the microblogging text to obtain micro-blog event；There is provided an affair character automatic abstraction module, to extract the feature vector of micro-blog event；A clobal relationship characteristic abstraction module is provided, the event vector to obtain event holotopy information indicates；A classification results output module is provided, to be spliced into final event vector and obtain the final rumour testing result of micro-blog event by classification function；The analysis of micro-blog event category and versatile can be preferably carried out, rumour detection can be carried out from the angle in data consumer.
Description
Technical field
The present invention relates to social media analysis and rumour detection field, especially a kind of amalgamation of global event relation information
Micro-blog event rumour detection method.
Background technique
In recent years, micro-blog rumour event detection has obtained the great attention of many scholars at home and abroad and research institution.It is micro-
The relevant micro-blog information of micro-blog event and user information is utilized in blog rumour event detection, carries out the inspection of micro-blog rumour event
It surveys, rumour detection can be carried out from the angle of data consumer, more fine-grained information is provided, effectively improve rumour detection
The order of accuarcy for the analysis result that system provides, facilitates to study and judge personnel and providing more efficient and accurate information.This is just right
Micro-blog rumour event detection technology proposes a challenge: how to construct an effective micro-blog rumour event detection prototype
System meets its needs.Therefore, the highly desirable micro-blog rumour event detecting method for having a kind of precise and high efficiency of people.
Currently, many technical methods can be used for rumour detection.Traditional rumour detection method is from Manual definition's feature
Angle is set out, and versatility is poor.It is different from traditional rumour detection method, how neural network model is utilized, automatically extracts micro-
The problem of blog affair character, and potential feature of the integration from conventional model are micro-blog rumour event detections is crucial.Tradition
Research work be mainly based upon supervision machine learning method, this method passes through the relevant spy of building micro-blog event
Sign carries out event category using decision tree or support vector machine classifier, improves rumour testing result.This kind is based on feature
Although the method for engineering achieves certain effect, but need to expend vast resources, and is limited to the rule of engineer,
So there are also to be hoisted for its performance.
Currently, it with the fast development of deep learning, is used based on neural network model in semantic expressiveness and rumour detection
Aspect has more advantage.These models are also used for micro-blog event category by many researchers.Neural network model and engineering
Learning method is compared, and can learn affair character from data automatically, avoids a large amount of Feature Engineering, at capture visual angle and up and down
Also there is better expansion in terms of complicated semantic relation between text.But traditional neural network rumour detection model is only automatic
Time correlation feature in extraction event is detected for rumour, is not fully considered and is previously obtained preferable rumour in machine learning
The potential feature of detection effect, cannot make full use of the information of micro-blog event.Rumour detection side based on neural network model
Method does not fully consider the holotopy information between event.
Summary of the invention
In view of this, the purpose of the present invention is to propose to a kind of inspections of the micro-blog event rumour of amalgamation of global event relation information
Survey method can extract validity feature automatically, and feature is abstracted and is combined, and whether finally identify micro-blog event
It is rumour.
The present invention is realized using following scheme: a kind of micro-blog event rumour detection side of amalgamation of global event relation information
Method provides a text data preprocessing module, the paragraph vector of the microblogging text to obtain micro-blog event；One event is provided
The automatic abstraction module of feature, to extract the feature vector of micro-blog event；One clobal relationship characteristic is provided and extracts mould
Block, the event vector to obtain event holotopy information indicate；A classification results output module is provided, to be spliced into most
Whole event vector simultaneously obtains the final rumour testing result of micro-blog event by classification function；
Specifically comprised the following steps: using the method that above-mentioned each module carries out micro-blog rumour event detection
Step S1: obtaining micro-blog event, and the text data preprocessing module utilizes the doc2vec tool pair of gensim
Microblogging text in the micro-blog event is trained to obtain microblogging text fragment vector；
Step S2: the automatic abstraction module of affair character carries out the microblogging sequence in micro-blog event according to the period
Feature extraction is divided and carried out, the feature vector of micro-blog event is extracted, so that obtaining the vector of micro-blog event indicates；
Step S3: the event comprising event holotopy information is obtained using the clobal relationship characteristic abstraction module
Vector indicates；
Step S4: the classification results output module is by the vector of the obtained micro-blog event of the step S2 and the step
The event vector for the event holotopy information that rapid S3 is obtained is spliced to obtain final event vector, and utilizes sigmoid
Classification function calculates the final event vector one by one, obtains the final rumour testing result of micro-blog event.
Further, the step S1 specifically further includes the following contents: according to the micro-blog event according to the period
It is divided, then regard the microblogging text of each period and every microblogging as one section of word；After pretreatment, by tabling look-up
Every section of text can be converted into text fragment vector form from textual form.
Further, the automatic abstraction module of affair character described in step S2 includes the simple cycle with attention mechanism
Cell network layers and attention layer；The simple cycle cell network layers with attention mechanism utilize simple cycle element mesh
Network models microblogging sequence and period sequence, and using the attention layer to the microblogging sequence and the period
Sequence applies microblogging attention and period attention, and integrates the user of the ratio, personal description of querying correction signal microblogging
Ratio, average user popularity, microblogging quantity and user's ratio characteristic of certification, the vector for finally obtaining micro-blog event indicate.
Further, the attention layer includes softmax layers, and described softmax layers to by simple cycle element mesh
The vector that network layers export at various moments is handled, and the attention distribution and period sequence of every microblogging in microblogging sequence are obtained
The attention distribution of each period in column.
Further, clobal relationship characteristic abstraction module described in step S3 includes tensor building and tensor resolution two
A part；The tensor building constructs event correlation matrix using the holotopy information between the micro-blog event, it
Afterwards, using the event correlation matrix as tensor piece, tensor is constructed；Finally, using Rescal tensor resolution algorithm to described
Tensor is decomposed, and the holotopy information of event is excavated, and obtaining the event vector comprising event holotopy information indicates.
Further, the event correlation matrix particular content are as follows: the difference of user group's confidence level including two events
Off course degree, the difference degree of the text confidence level of two events, two events user group registration and two events
Other holotopies.
Further, the step S4 specifically further includes the following contents: using sigmoid classification function to described final
Event vector calculates one by one, obtains the micro-blog event prediction value according to the threshold value of setting；In the training stage, need to utilize loss
Function calculates the error of predicted value and target value, and using Adam optimization algorithm to the text data preprocessing module, described
The parameter of the automatic abstraction module of affair character, the clobal relationship characteristic abstraction module and the classification results output module
It is iterated update；If error no longer reduces or reach 100 the number of iterations, terminate the training stage, under otherwise continuing
Primary iteration；
The loss function are as follows:
Wherein, LjWhat is indicated is the true tag of j-th of event, andThat indicate is the prediction label of j-th of event, θ
It is model parameter set, includes the weight of simple cycle cell network layers and attention layer and full articulamentum in the set；N
It is the quantity of event in training set.
Compared with prior art, the invention has the following beneficial effects:
The present invention can extract validity feature automatically, and feature is abstracted and is combined, and finally identify micro-blog
Whether event is rumour.The analysis of micro-blog event category and versatile can be preferably carried out, can stand and be used in data
The angle of person carries out rumour detection.
Detailed description of the invention
Fig. 1 is the overall construction drawing of the embodiment of the present invention.
Specific embodiment
The present invention will be further described with reference to the accompanying drawings and embodiments.
As shown in Figure 1, present embodiments providing a kind of micro-blog event rumour detection of amalgamation of global event relation information
Method provides a text data preprocessing module, the paragraph vector of the microblogging text to obtain micro-blog event；One thing is provided
The automatic abstraction module of part feature, to extract the feature vector of micro-blog event；The extraction of one clobal relationship characteristic is provided
Module, the event vector to obtain event holotopy information indicate；A classification results output module is provided, to be spliced into
Final event vector simultaneously obtains the final rumour testing result of micro-blog event by classification function；
Specifically comprised the following steps: using the method that above-mentioned each module carries out micro-blog rumour event detection
Step S1: obtaining micro-blog event, and the text data preprocessing module utilizes the doc2vec tool pair of gensim
Microblogging text in the micro-blog event is trained to obtain microblogging text fragment vector；
Step S2: the automatic abstraction module of affair character carries out the microblogging sequence in micro-blog event according to the period
Feature extraction is divided and carried out, the feature vector of micro-blog event is extracted, so that obtaining the vector of micro-blog event indicates；
Step S3: the event comprising event holotopy information is obtained using the clobal relationship characteristic abstraction module
Vector indicates；
Step S4: the classification results output module is by the vector of the obtained micro-blog event of the step S2 and the step
The event vector for the event holotopy information that rapid S3 is obtained is spliced to obtain final event vector, and utilizes sigmoid
Classification function calculates the final event vector one by one, obtains the final rumour testing result of micro-blog event.
In the present embodiment:
1) data preprocessing module 1
Firstly, how description data preprocessing module 1 obtains paragraph vector.
Because the input data of neural network is usually vector, so as to the end-to-end training of model, it is therefore desirable to right first
Text data carries out vectorization expression.For the ease of the processing and analysis of data, in the data preprocessing module of the present embodiment, root
It is divided according to micro-blog event according to the period, then by the microblogging text of each period and each conduct of every microblogging
One section of word.After pretreatment, every section of text can be converted into vector form from textual form by tabling look-up.
2) the automatic abstraction module 2 of affair character
It is that the data how to obtain a upper module carry out affair character pumping that the automatic abstraction module 2 of affair character, which is described below,
It takes.This module is made of two layers simple cycle cell network layers and attention layer with attention mechanism.With attention
The core of the simple cycle cell network layers of mechanism be using simple cycle unit networks to microblogging sequence and period sequence into
Row modeling, and microblogging attention and period attention are applied to microblogging sequence and period sequence using attention layer, and whole
It closes and queries the ratio of correction signal microblogging, has the personal user's ratio described, average user popularity, microblogging quantity, the use of certification
The manual features such as family ratio, the vector for finally obtaining micro-blog event indicate.
3) clobal relationship characteristic abstraction module 3
It is described below clobal relationship characteristic abstraction module 3 is how to carry out feature pumping to the holotopy information of event
It takes.This module is made of tensor building and two parts of tensor resolution.Tensor building core be using micro-blog event it
Between holotopy information architecture event correlation matrix, the difference degree of user group's confidence level including two events, two
The difference degree of the text confidence level of event, the registration of user group of two events and other overall situations of two events are closed
System.Later, using obtained event correlation matrix as tensor piece, tensor is constructed.Finally, being calculated using Rescal tensor resolution
Method decomposes the tensor, excavates the holotopy information of event, obtain include event holotopy information event to
Amount indicates.
4) classification results output module 4
Finally, interpretive classification result output module 4.The output vector of module 2 and module 3 has obtained final micro-blog thing
Two kinds of vectors are attached by the vector of part, classification results output module, and using sigmoid classification function to gained vector by
One calculates, and obtains the micro-blog event prediction value according to the threshold value of setting.In the training stage, need to ask predicted value and target value
Error calculates when error shown in used loss function such as formula (1), and using Adam optimization algorithm to the text data
Preprocessing module, the automatic abstraction module of the affair character, the clobal relationship characteristic abstraction module and classification knot
The parameter of fruit output module is iterated update；If error no longer reduces or reach 100 the number of iterations, terminate to train rank
Section, otherwise continues iteration next time.
Wherein, LjWhat is indicated is the true tag of j-th of event, andThat indicate is the prediction label of j-th of event, θ
It is model parameter set, includes the weight of simple cycle cell network layers and attention layer and full articulamentum in the set.N
It is the quantity of event in training set.
Preferably, the present embodiment is by the period feature vector of Manual definition and with the simple cycle list of attention mechanism
The obtained period feature vector of metanetwork layer is attached, comprehensively considered in micro-blog event the feature of time correlation and
The influence that the manual features of each period detect rumour；Microblogging attention and period attention are constructed, is given priority to
Utilize the micro-blog information in micro-blog event；The holotopy information between micro-blog event is excavated.
Preferably, the attention layer contains softmax layers, and described softmax layers to by simple cycle element mesh
The vector that network layers export at various moments is handled, and the attention distribution and period sequence of every microblogging in microblogging sequence are obtained
The attention distribution of each period in column.
Preferably, classification results output module described in the present embodiment using sigmoid function to obtained text to
Amount processing, predicts the viewpoint classification of each text.Predict whether micro-blog event is rumour.
Preferably, training stage of the present embodiment in model, weight matrix are all parameters, according to the propagated forward of information and
The back-propagating of error will constantly be adjusted them, successive optimization objective function.
Preferably, clobal relationship characteristic abstraction module described in the present embodiment uses Rescal tensor resolution algorithm
The holotopy information between micro-blog event has been excavated, the vector of each micro-blog event is obtained.
The foregoing is merely presently preferred embodiments of the present invention, all equivalent changes done according to scope of the present invention patent with
Modification, is all covered by the present invention.
Claims (7)
1. a kind of micro-blog event rumour detection method of amalgamation of global event relation information, it is characterised in that: provide a text
Data preprocessing module, the paragraph vector of the microblogging text to obtain micro-blog event；An affair character is provided to extract automatically
Module, to extract the feature vector of micro-blog event；A clobal relationship characteristic abstraction module is provided, to obtain thing
The event vector of part holotopy information indicates；A classification results output module is provided, to be spliced into final event vector
And the final rumour testing result of micro-blog event is obtained by classification function；
Specifically comprised the following steps: using the method that above-mentioned each module carries out micro-blog rumour event detection
Step S1: micro-blog event is obtained, the text data preprocessing module is using the doc2vec tool of gensim to described
Microblogging text in micro-blog event is trained to obtain microblogging text fragment vector；
Step S2: the automatic abstraction module of affair character carries out the microblogging sequence in the micro-blog event according to the period
Feature extraction is divided and carried out, the feature vector of the micro-blog event is extracted, to obtain the vector table of micro-blog event
Show；
Step S3: the event vector comprising event holotopy information is obtained using the clobal relationship characteristic abstraction module
It indicates；
Step S4: the classification results output module is by the vector of the obtained micro-blog event of the step S2 and the step S3
The event vector of obtained event holotopy information is spliced to obtain final event vector, and is classified using sigmoid
Function calculates the final event vector one by one, obtains the final rumour testing result of micro-blog event.
2. a kind of micro-blog event rumour detection method of amalgamation of global event relation information according to claim 1,
Be characterized in that: the step S1 specifically further includes the following contents: it is divided according to the micro-blog event according to the period,
Then the microblogging text of each period and every microblogging are regard as one section of word；It, can will be every by tabling look-up after pretreatment
Duan Wenben is converted into text fragment vector form from textual form.
3. a kind of micro-blog event rumour detection method of amalgamation of global event relation information according to claim 1,
Be characterized in that: the automatic abstraction module of affair character described in step S2 includes the simple cycle unit networks with attention mechanism
Layer and attention layer；The simple cycle cell network layers with attention mechanism are using simple cycle unit networks to microblogging
Sequence and period sequence are modeled, and are applied using the attention layer to the microblogging sequence and the period sequence
Microblogging attention and period attention, and integrate user's ratio of the ratio, personal description of querying correction signal microblogging, be averaged
User's ratio characteristic of user's popularity, microblogging quantity and certification, the vector for finally obtaining micro-blog event indicate.
4. a kind of micro-blog event rumour detection method of amalgamation of global event relation information according to claim 3,
Be characterized in that: the attention layer include softmax layers, described softmax layers to by simple cycle cell network layers each
The vector of a moment output is handled, and is obtained each in the attention distribution and period sequence of every microblogging in microblogging sequence
The attention of period is distributed.
5. a kind of micro-blog event rumour detection method of amalgamation of global event relation information according to claim 1,
Be characterized in that: clobal relationship characteristic abstraction module described in step S3 includes tensor building and two parts of tensor resolution；
The tensor building constructs event correlation matrix using the holotopy information between the micro-blog event, later, will be described
Event correlation matrix constructs tensor as tensor piece；Finally, being divided using Rescal tensor resolution algorithm the tensor
Solution, excavates the holotopy information of event, and obtaining the event vector comprising event holotopy information indicates.
6. a kind of micro-blog event rumour detection method of amalgamation of global event relation information according to claim 4,
It is characterized in that: the event correlation matrix particular content are as follows: the difference degree of user group's confidence level including two events, two
The difference degree of the text confidence level of a event, two events user group registration and two events other are global
Relationship.
7. a kind of micro-blog event rumour detection method of amalgamation of global event relation information according to claim 1,
Be characterized in that: the step S4 specifically further includes the following contents: using sigmoid classification function to the final event vector
It calculates one by one, which is obtained according to the threshold value of setting；In the training stage, need to calculate using loss function
The error of predicted value and target value, and using Adam optimization algorithm to the text data preprocessing module, the affair character
The parameter of automatic abstraction module, the clobal relationship characteristic abstraction module and the classification results output module is iterated
It updates；If error no longer reduces or reach 100 the number of iterations, terminate the training stage, otherwise continue next time repeatedly
Generation；
The loss function are as follows:
Wherein, LjWhat is indicated is the true tag of j-th of event, andWhat is indicated is the prediction label of j-th of event, and θ is model
Parameter sets include the weight of simple cycle cell network layers and attention layer and full articulamentum in the set；N is trained
The quantity of concentration event.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910042008.7A CN109783629A (en) | 2019-01-16 | 2019-01-16 | A kind of micro-blog event rumour detection method of amalgamation of global event relation information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910042008.7A CN109783629A (en) | 2019-01-16 | 2019-01-16 | A kind of micro-blog event rumour detection method of amalgamation of global event relation information |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109783629A true CN109783629A (en) | 2019-05-21 |
Family
ID=66500831
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910042008.7A Pending CN109783629A (en) | 2019-01-16 | 2019-01-16 | A kind of micro-blog event rumour detection method of amalgamation of global event relation information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109783629A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111160040A (en) * | 2019-12-26 | 2020-05-15 | 西安交通大学 | Information reliability evaluation system and method based on multi-scale gating equilibrium interaction fusion network |
| CN111506710A (en) * | 2020-07-01 | 2020-08-07 | 平安国际智慧城市科技股份有限公司 | Information sending method and device based on rumor prediction model and computer equipment |
| CN112035669A (en) * | 2020-09-09 | 2020-12-04 | 中国科学技术大学 | Social media multi-modal rumor detection method based on propagation heterogeneous graph modeling |
| CN112464082A (en) * | 2020-11-05 | 2021-03-09 | 重庆邮电大学 | Rumor-splitting game propagation control method based on sparse representation and tensor completion |
| CN115809327A (en) * | 2023-02-08 | 2023-03-17 | 四川大学 | Real-time social network rumor detection method for multi-mode fusion and topics |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120311030A1 (en) * | 2011-05-31 | 2012-12-06 | International Business Machines Corporation | Inferring User Interests Using Social Network Correlation and Attribute Correlation |
| CN105224507A (en) * | 2015-09-29 | 2016-01-06 | 杭州天宽科技有限公司 | A kind of disappearance association rule mining method based on tensor resolution |
| CN106600052A (en) * | 2016-12-12 | 2017-04-26 | 西安交通大学 | User attribute and social network detection system based on space-time locus |
-
2019
- 2019-01-16 CN CN201910042008.7A patent/CN109783629A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120311030A1 (en) * | 2011-05-31 | 2012-12-06 | International Business Machines Corporation | Inferring User Interests Using Social Network Correlation and Attribute Correlation |
| CN105224507A (en) * | 2015-09-29 | 2016-01-06 | 杭州天宽科技有限公司 | A kind of disappearance association rule mining method based on tensor resolution |
| CN106600052A (en) * | 2016-12-12 | 2017-04-26 | 西安交通大学 | User attribute and social network detection system based on space-time locus |
Non-Patent Citations (2)
| Title |
|---|
| 廖祥文，杨定达，程学旗，陈国龙: "基于分层注意力网络的社交媒体谣言检测", 《中国科学 : 信息科学》 * |
| 廖祥文，杨定达，程学旗，陈国龙: "基于分层注意力网络的社交媒体谣言检测", 《中国科学：信息科学》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111160040A (en) * | 2019-12-26 | 2020-05-15 | 西安交通大学 | Information reliability evaluation system and method based on multi-scale gating equilibrium interaction fusion network |
| CN111506710A (en) * | 2020-07-01 | 2020-08-07 | 平安国际智慧城市科技股份有限公司 | Information sending method and device based on rumor prediction model and computer equipment |
| CN111506710B (en) * | 2020-07-01 | 2020-11-06 | 平安国际智慧城市科技股份有限公司 | Information sending method and device based on rumor prediction model and computer equipment |
| CN112035669A (en) * | 2020-09-09 | 2020-12-04 | 中国科学技术大学 | Social media multi-modal rumor detection method based on propagation heterogeneous graph modeling |
| CN112464082A (en) * | 2020-11-05 | 2021-03-09 | 重庆邮电大学 | Rumor-splitting game propagation control method based on sparse representation and tensor completion |
| CN112464082B (en) * | 2020-11-05 | 2022-06-28 | 重庆邮电大学 | Rumor-dagger game propagation control method based on sparse representation and tensor completion |
| CN115809327A (en) * | 2023-02-08 | 2023-03-17 | 四川大学 | Real-time social network rumor detection method for multi-mode fusion and topics |
| CN115809327B (en) * | 2023-02-08 | 2023-05-05 | 四川大学 | Real-time social network rumor detection method based on multimode fusion and topics |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109783629A (en) | A kind of micro-blog event rumour detection method of amalgamation of global event relation information | |
| CN108984745B (en) | Neural network text classification method fusing multiple knowledge maps | |
| CN106372402B (en) | The parallel method of fuzzy region convolutional neural networks under a kind of big data environment | |
| CN104318340B (en) | Information visualization methods and intelligent visible analysis system based on text resume information | |
| CN109062939A (en) | A kind of intelligence towards Chinese international education leads method | |
| CN108664632A (en) | A kind of text emotion sorting algorithm based on convolutional neural networks and attention mechanism | |
| CN114092832B (en) | High-resolution remote sensing image classification method based on parallel hybrid convolutional network | |
| CN107122416A (en) | A kind of Chinese event abstracting method | |
| CN107092596A (en) | Text emotion analysis method based on attention CNNs and CCR | |
| CN103207855A (en) | Fine-grained sentiment analysis system and method specific to product comment information | |
| CN112199608B (en) | Social media rumor detection method based on network information propagation graph modeling | |
| CN109710769A (en) | A kind of waterborne troops's comment detection system and method based on capsule network | |
| CN108172301A (en) | A kind of mosquito matchmaker's epidemic Forecasting Methodology and system based on gradient boosted tree | |
| CN109934261A (en) | A kind of Knowledge driving parameter transformation model and its few sample learning method | |
| CN107463966A (en) | Radar range profile's target identification method based on dual-depth neutral net | |
| CN107220525A (en) | Gene regulatory network based on RNN is built and dynamic differential analysis method | |
| CN106294313A (en) | Study embeds for entity and the word of entity disambiguation | |
| CN107846392A (en) | A kind of intrusion detection algorithm based on improvement coorinated training ADBN | |
| CN103942749B (en) | A kind of based on revising cluster hypothesis and the EO-1 hyperion terrain classification method of semi-supervised very fast learning machine | |
| CN110232280A (en) | A kind of software security flaw detection method based on tree construction convolutional neural networks | |
| CN103268607B (en) | A kind of common object detection method under weak supervision condition | |
| CN109614495A (en) | A kind of associated companies method for digging of combination knowledge mapping and text information | |
| CN110263822A (en) | A kind of Image emotional semantic analysis method based on multi-task learning mode | |
| CN102999615A (en) | Diversified image marking and retrieving method based on radial basis function neural network | |
| CN108470061A (en) | A kind of emotional semantic classification system for visual angle grade text |
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: 20190521 |
|
| RJ01 | Rejection of invention patent application after publication |