CN110297984A - Information transmission dynamics system, construction method, device and medium based on microblogging - Google Patents
Information transmission dynamics system, construction method, device and medium based on microblogging Download PDFInfo
- Publication number
- CN110297984A CN110297984A CN201910411316.2A CN201910411316A CN110297984A CN 110297984 A CN110297984 A CN 110297984A CN 201910411316 A CN201910411316 A CN 201910411316A CN 110297984 A CN110297984 A CN 110297984A
- Authority
- CN
- China
- Prior art keywords
- public sentiment
- state
- index
- outburst
- users
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9536—Search customisation based on social or collaborative filtering
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The present invention provides a kind of information transmission dynamics system, construction method, device and medium based on microblogging, comprising: the state of user in microblogging is divided into sensitization, forwarding state or/and read state and immune state;Information transmission dynamics model is constructed according to the variation of User Status;The following number of users in each state and variation tendency are predicted according to above- mentioned information transmission dynamics model by the monitoring of accumulative transfer amount or/and accumulative amount of reading to microblogging.The forwarding or/and reading information that above system, construction method, device and medium externally announce microblogging are analyzed, and are predicted public sentiment.
Description
Technical field
The present invention relates to the dynamical system constructing technology fields that public sentiment is propagated, more particularly, to a kind of based on microblogging
Information transmission dynamics system, construction method, device and medium.
Background technique
By the end of the year 2018, Sina weibo monthly possesses 4.46 hundred million any active ues (the 32% of Zhan Zhongguo total population), is
The maximum social media of state.By taking " one, Chongqing bus falls river event " as an example, microblogging produces about 10,000 forwardings in 24 hours
User and 100,000,000 reads user, can illustrate that the information of microblogging is propagated rapidly and influenced significant.Since the network of each complexity has
There is specific topological characteristic, some articles investigate relevant characteristic measure.Due to the difference of culture, Chinese Sina
Microblogging has some different features in terms of content and information propagation.In China, Sina weibo is that an important public opinion passes
Platform is broadcast, therefore, the research of micro-blog information mechanism of transmission is increasingly becoming current research hotspot.
The information communication process of blog (Microblog) is the network reality that can be interacted by " concern " behavior connection
Existing.The user for implementing concern behavior is known as " bean vermicelli ", i.e. follower, and the user as person of being concerned that the bean vermicelli is of interest.
The diffusion model of basic structure based on blog, information mainly includes three categories: model based on communication process, based on influencing
The model of power and based on forwarding factor model.When focus to be conceived to the dynamic change of information communication process, based on dynamic
The Epidemic Model of mechanics has obtained extensive research, although this field has numerous scholars and expands extensive work,
Due to the difference of cultural traits, microblogging has some specific characteristics for being different from other social medias, and dynamic communication characteristic is also
Up for further studying.
Have some information transmission dynamics systems based on classical Epidemic Model at present, crowd is divided into and is not overlapped
Three classes: not in contact with the people (know-nothing) of rumour, actively the people (disseminator) that spreads rumors and the people no longer to spread rumors are (immune
Person).It is the refinement of propagation characteristic to first thinking that classical model improves, it is contemplated that the incomplete reading of user is gone
For researcher proposes the blog-easy infection-has infected-propagated for information and deletes (Microblog-Susceptible-
Infected-Removed, Mb-SIR) model.In addition, there is also two kinds of improved models, the first kind considers disseminator not
Always active situation, the second class consider the know-nothing of information to spreading rumors and uninterested possibility, and by pair
Twitter platform data progress numerical simulation obtains the more classical gossip propagation model of model constructed by it closer to actual conditions
Conclusion.Some scholars consider the shadow that the intimate degree between the communication effect of immune and members of society propagates information
It rings, proposes improved gossip propagation model (modified rumor spreading model, SIRe).Numerical simulation result
Show that the higher group of cohesion is easier to possess the disseminator of higher proportion, and disperse the immune concentration that is better than and be immunized, can make
The maximum rate of disseminator is reduced.In addition, another crowd of scholar has also set up and has retouched to distinguish the effects of different methods of refuting a rumour
The gossip propagation model of user browsing behavior feature is stated, analysis is the result shows that the propagation of information has certain timeliness.
It is the research that new module is added and improves communication process to another thinking that classical propagation model improves,
It is exactly other than know-nothing, disseminator and immune, it is also contemplated that other states propagated.It especially incorporates with infectiousness
Bean vermicelli quantity, the research verified using truthful data to revised model reflects the information characteristics of Sina weibo
And propagation characteristic.Researcher constructs a dynamic model (Susceptible users (S), super-spreaders
(A), normal spreaders (I) recovered users (R), SAIR) it is existing to characterize the super propagation pushed away in literary information propagation
As.In addition, super-spreader's state proposes an improved model (Ignorant spreaders (know-nothing), Super-
Spreaders (super-spreader), Stiflers model, ISJR), this demonstrate super-spreader how in the network of blog plus
Fast information is propagated, and informational influence power is expanded.Meanwhile there are many more other models by the recovery status merging of the user of different platform
Increase is new plate, including Chinese microblogging, Japan Mixi and Facebook.
In the research above to all kinds of blog propagation characteristics and transmission structure, scholars mostly concentrate on focus of attention
On the information communication theory of broad sense, the difference and connection of different data propagation characteristic are not accounted for, is only carried out with single-dimensional data real
Example verifying, without combining the expansion of true available data to combine the research of actual conditions multi-angle.
Summary of the invention
In view of the above problems, the present invention provides forwarding that a kind of pair of microblogging is externally announced or/and reading information is analyzed,
Information transmission dynamics system, construction method, device and the medium based on microblogging that public sentiment is predicted.
According to an aspect of the present invention, a kind of information transmission dynamics system constituting method based on microblogging is provided, comprising:
The state of user in microblogging is divided into three kinds of sensitization, forwarding state and immune state states, wherein described easy
Sense state indicates that user not yet contacts issuing microblog but has the state that microblogging is forwarded ability;The forwarding state indicates
Have been carried out forwarding and in active state can be to the state that other people have an impact;The immune state indicates that user turns
Lost after hair the ability of enlivening state or reading information after actively abandon forwarding microblogging state;
Information transmission dynamics model is constructed by following formula (1)-(4)
S ' (t)=- β S (t) F (t) (1)
F ' (t)=p1βS(t)F(t)-α1F(t) (2)
I ' (t)=(1-p1)βS(t)F(t)+α1F(t) (3)
E ' (t)=p1βS(t)F(t) (4)
Wherein, t is time index, and β indicates the user in sensitization in the rate of the mean exposure information of microblogging, p1
Indicate that the user in sensitization forwards the rate of microblogging, α1Indicate that the user in forwarding state becomes sluggish speed
Rate, S (t) are Susceptible population's quantity, and F (t) is forwarding crowd quantity, and I (t) is immune crowd's quantity, and E (t) is accumulative transfer amount,
S ' (t), F ' (t), I ' (t) and E ' (t) are respectively the derivative of S (t), F (t), I (t) and E (t), and the derivative is greater than 0 expression and refers to
It marks numerical value to increase, derivative indicates that index value reduces less than 0, and derivative, which is equal to 0, indicates that index value is constant;
By the monitoring to accumulative transfer amount according to above- mentioned information transmission dynamics model to following Susceptible population's quantity, turn
Hair crowd quantity and immune crowd's quantity and its variation are predicted.
According to the second aspect of the invention, a kind of information transmission dynamics system constituting method based on microblogging is provided, comprising:
The state of user in microblogging is divided into three kinds of sensitization, read state and immune state states, wherein described easy
Sense state indicates that user not yet contacts issuing microblog but has the state that microblogging carries out reading ability;The read state indicates
Have been carried out reading and in active state can be to the state that other people have an impact;The immune state indicates that user reads
The state for the ability of enlivening is lost after reading or has exceeded the state for reading validity period;
Information transmission dynamics model is constructed by following formula (8)-(11)
S ' (t)=- β S (t) R (t)+p2α2R(t) (8)
R ' (t)=β S (t) R (t)-α2R(t) (9)
I ' (t)=(1-p2)α2R(t) (10)
E′2(t)=β S (t) R (t) (11)
Wherein, R (t) is reading crowd quantity, E2It (t) is accumulative amount of reading, p2Indicate that reading user generates repeat reading
The rate of behavior, α2Indicate that user becomes sluggish rate, S ' (t), F ' (t), I ' (t) and E ' in read state2(t) respectively
For S (t), F (t), I (t) and E2(t) derivative, the derivative, which is greater than 0, indicates that index value increases, and derivative indicates index less than 0
Numerical value reduces, and derivative, which is equal to 0, indicates that index value is constant;
By the monitoring to accumulative amount of reading according to above- mentioned information transmission dynamics model to following Susceptible population's quantity, read
Reading crowd quantity and immune crowd's quantity and its variation are predicted.
According to the third aspect of the invention we, a kind of information transmission dynamics system based on microblogging is provided, comprising:
The state of user in microblogging is divided into sensitization, forwarding state and immune state three by User Status setting module
Kind state, wherein the sensitization indicates that user not yet contacts issuing microblog but has the shape that microblogging is forwarded ability
State;The forwarding state expression has been carried out forwarding and can be to the state that other people have an impact in active state;Institute
It states the state for losing the ability of enlivening after immune state expression user forwards or actively abandons the shape of forwarding microblogging after reading information
State;
Model construction module constructs information transmission dynamics model by following formula (1)-(4)
S ' (t)=- β S (t) F (t) (1)
F ' (t)=p1βS(t)F(t)-α1F(t) (2)
I ' (t)=(1-p1)βS(t)F(t)+α1F(t) (3)
E ' (t)=p1βS(t)F(t) (4)
Wherein, t is time index, and β indicates the user in sensitization in the rate of the mean exposure information of microblogging, p1
Indicate that the user in sensitization forwards the rate of microblogging, α1Indicate that the user in forwarding state becomes sluggish speed
Rate, S (t) are Susceptible population's quantity, and F (t) is forwarding crowd quantity, and I (t) is immune crowd's quantity, and E (t) is accumulative transfer amount,
S ' (t), F ' (t), I ' (t) and E ' (t) are respectively the derivative of S (t), F (t), I (t) and E (t), and the derivative is greater than 0 expression and refers to
It marks numerical value to increase, derivative indicates that index value reduces less than 0, and derivative, which is equal to 0, indicates that index value is constant;
Prediction module, the information transmission dynamics constructed by the monitoring to accumulative transfer amount according to above-mentioned model construction module
The quantity and variation of the user in different states of model prediction User Status setting module setting.
Preferably, the User Status setting module by the state of user in microblogging be divided into sensitization, read state and
Three kinds of states of immune state, wherein the sensitization indicates that user not yet contacts issuing microblog but having microblogging reads
The state of reading ability;The read state expression has been carried out reading and can have an impact to other people in active state
State;The immune state loses the state for the ability of enlivening after indicating user's reading or has exceeded the shape for reading validity period
State;
The model construction module constructs information transmission dynamics model by following formula (8)-(11)
S ' (t)=- β S (t) R (t)+p2α2R(t) (8)
R ' (t)=β S (t) R (t)-α2R(t) (9)
I ' (t)=(1-p2)α2R(t) (10)
E′2(t)=β S (t) R (t) (11)
Wherein, R (t) is reading crowd quantity, E2It (t) is accumulative amount of reading, p2Indicate that reading user generates repeat reading
The rate of behavior, α2Indicate that user becomes sluggish rate, S ' (t), F ' (t), I ' (t) and E ' in read state2(t) respectively
For S (t), F (t), I (t) and E2(t) derivative, the derivative, which is greater than 0, indicates that index value increases, and derivative indicates index less than 0
Numerical value reduces, and derivative, which is equal to 0, indicates that index value is constant;
The prediction module is propagated by the monitoring to accumulative amount of reading according to the information that above-mentioned model construction module constructs
Dynamic model predicts the quantity and variation of the user in different states of User Status setting module setting.
In addition, the present invention also provides a kind of electronic device, including memory and processor, base is stored in the memory
In the information transmission dynamics system construction procedures of microblogging, the information transmission dynamics system construction procedures based on microblogging are described
The step of processor realizes the above-mentioned information transmission dynamics system constituting method based on microblogging when executing.
In addition, including in the computer readable storage medium the present invention also provides a kind of computer readable storage medium
There are the information transmission dynamics system construction procedures based on microblogging, the information transmission dynamics system construction procedures quilt based on microblogging
Processor execute when, realize the above-mentioned information transmission dynamics system constituting method based on microblogging the step of.
Above-mentioned information transmission dynamics system, construction method, device and medium based on microblogging is to micro-blog information mechanism of transmission
It is studied, analyzes the general modfel of rich information dissemination mechanism, the forwarding externally announced microblogging, reading information are analyzed,
Construct information transmission dynamics model, carry out public sentiment index building, in advance grasp Internet streaming action to and detect micro-blog information
Propagate, the current development propagated in time to information is desk-top and change in future is judged and predicted, it is final realize public sentiment prediction and
Early warning is of great significance to maintaining social stability, building a harmonious society.
Detailed description of the invention
Fig. 1 is the schematic diagram of information transmission dynamics model SFI of the present invention;
Fig. 2 is the composition block diagram of one embodiment of the information transmission dynamics system of the present invention based on microblogging;
Fig. 3 is the signal of the public sentiment index of one embodiment of the information transmission dynamics system of the present invention based on microblogging
Figure;
Fig. 4 is that of the present invention the present invention is based on one embodiment of the construction method of the information transmission dynamics system of microblogging
Flow chart;
Fig. 5 is the schematic diagram of information transmission dynamics model SRI of the present invention;
Fig. 6 is the composition block diagram of another embodiment of the information transmission dynamics system of the present invention based on microblogging;
Fig. 7 is showing for the public sentiment index of another embodiment of the information transmission dynamics system of the present invention based on microblogging
It is intended to;
Fig. 8 is the schematic diagram of topological diagram of the present invention;
Fig. 9 is the quantity of each User Status in mono- specific embodiment of information transmission dynamics model SFI of the present invention
Matched curve figure;
Figure 10 is the number in mono- specific embodiment of information transmission dynamics model SFI of the present invention according to preceding setting time
It is predicted that accumulative transfer amount and forwarding crowd's quantity schematic diagram;
Figure 11 is the quantity of each User Status in mono- specific embodiment of information transmission dynamics model SRI of the present invention
Matched curve figure;
Figure 12 is the number in mono- specific embodiment of information transmission dynamics model SRI of the present invention according to preceding setting time
It is predicted that accumulative amount of reading and read crowd's quantity schematic diagram.
Specific embodiment
In the following description, for purposes of illustration, it in order to provide the comprehensive understanding to one or more embodiments, explains
Many details are stated.It may be evident, however, that these embodiments can also be realized without these specific details.
In other examples, one or more embodiments for ease of description, well known structure and equipment are shown in block form an.
Each embodiment according to the present invention is described in detail below with reference to accompanying drawings.
Single microblogging is the most basic unit for constituting micro-blog information and propagating, and has great research significance.
In one embodiment, it for the research of unit microblogging, defines for same microblogging, each user only forwards once
Hypothesis.Compared with classical Epidemic Model, the information transmission dynamics model SFI of quasi- building more meets the characteristic of microblogging,
Exactly consider when user checks a microblogging, has the choice of whether to forward the microblogging, i.e. whether user can continue information
It propagates, if user's selection is forwarded, forwarding state will be entered, if after a certain time, user will no longer be forwarded, i.e.,
Switch to immune state, the information transmission dynamics model SFI for intending building is as shown in Figure 1.
Fig. 2 is the composition block diagram of one embodiment of the information transmission dynamics system the present invention is based on microblogging, such as Fig. 2 institute
Show, the information transmission dynamics system includes:
First User Status setting module 1, by the state of user in microblogging be divided into sensitization (S), forwarding state (F) and
Three kinds of states of immune state (I), wherein the sensitization indicates that user not yet contacts issuing microblog but has microblogging progress
The state of transfer capability;The forwarding state expression has been carried out forwarding and can generate shadow to other people in active state
Loud state;The immune state indicates the state for losing the ability of enlivening after user's forwarding or actively abandons turning after reading information
Send out the state of microblogging;
First model construction module 2 constructs information transmission dynamics model by following formula (1)-(4)
S ' (t)=- β S (t) F (t) (1)
F ' (t)=p1βS(t)F(t)-α1F(t) (2)
I ' (t)=(1-p1)βS(t)F(t)+α1F(t) (3)
E ' (t)=p1βS(t)F(t) (4)
Wherein, t is time index, and β indicates the user in sensitization in the rate of the mean exposure information of microblogging, p1
Indicate that the user in sensitization forwards the rate of microblogging, α1Indicate that the user in forwarding state becomes sluggish speed
Rate, S (t) are Susceptible population's quantity, and F (t) is forwarding crowd quantity, and I (t) is immune crowd's quantity, and E (t) is accumulative transfer amount,
S ' (t), F ' (t), I ' (t) and E ' (t) are respectively the derivative of S (t), F (t), I (t) and E (t), and the derivative is greater than 0 expression and refers to
It marks numerical value to increase, derivative indicates that index value reduces less than 0, and derivative, which is equal to 0, indicates that index value is constant;
First prediction module 3, the letter constructed by the monitoring to accumulative transfer amount according to above-mentioned first model construction module
The quantity and variation of the user in different states of transmission dynamics model prediction the first User Status setting module setting are ceased,
In, accumulative transfer amount E (t) is the available data that microblogging is externally announced.
Preferably, first prediction module 3 also predicts public sentiment by public sentiment index, as shown in figure 3, the public sentiment index
Including the maximum public sentiment spread index of the first public sentiment explosive index, the first public sentiment propagation peak and first, first prediction module
Include:
First public sentiment explosive index predicting unit predicts the first carriage according to the following formula (5) by initial Susceptible population's quantity
Feelings explosive index, the first public sentiment explosive index indicate that the public sentiment outburst stage actively forwards the forwarding of customer impact to use by one
The quantity at family also illustrates that the severity of public sentiment event outburst
Wherein :=indicate to be defined as,For the first public sentiment explosive index, S0Indicate the number of users for being in sensitization
Initial value, 1/ α1It indicates active time, the time span of active period is defined as forwarding validity period, due to the definition to forwarding
There are certain period of time divisions, so there is forwarding validity period when user is forwarded, enabling u, (s is the use in forwarding state
Ss hour is passed through still in the quantity of the user of forwarding state, if with ratio, α in the unit time in family1Forwarding state is left,
ThenIt is availableI.e. the user in forwarding state is after ss unit time
Still it is for the ratio of the user in forwarding stateNamely forwarding validity period is that average value is
Exponential distribution;
Judge that the first public sentiment explosive index is greater than 1, is also equal to 1 less than 1;
IfShow to forward number of users decline, microblogging will be unable to further propagate;If
Show to forward number of users that will exponentially increase;IfShow to forward number of users without change;
First public sentiment propagation peak predicting unit predicts the first carriage according to the following formula (6) by initial Susceptible population's quantity
Feelings propagation peak Fmax, user forwards behavior within the unit time when the first public sentiment propagation peak expression public sentiment event is broken out
The maximum value of increment
Wherein,0Indicate the initial value of the number of users in forwarding state;
First time predicting unit needs to know specific microblogging for the performance for understanding the microblogging ecosystem in global range
Initial time, peak value explosion time and the end time of outburst increase forwarding number of users to the first public sentiment propagation peak first
Initial time is broken out as the first public sentiment at the time of setting ratio, the first public sentiment outburst initial time is predicted;It will forwarding
Peak value moment is broken out as the first public sentiment at the time of number of users increases to the first public sentiment propagation peak, peak value is broken out to the first public sentiment
Moment is predicted;As the first public sentiment at the time of forwarding number of users is dropped to first public sentiment the second setting ratio of propagation peak
Finish time is broken out, the first public sentiment outburst finish time is predicted;
First Speed predicting unit arrives the first public sentiment outburst initial time to characterize microblogging diffusion duration
The forwarding number of users that first public sentiment breaks out the unit time of peak value moment breaks out speed as the first public sentiment, realizes to the first carriage
The prediction of feelings outburst speed;By the forwarding of the unit time of the first public sentiment outburst peak value moment to the first public sentiment outburst finish time
Number of users realizes the prediction to the first public sentiment decline rate as the first public sentiment decline rate;First public sentiment is broken out and is originated
The forwarding number of users of the unit time of moment to the first public sentiment outburst finish time averagely breaks out rate, realization pair as first
First public sentiment averagely breaks out the prediction of rate;
First maximum public sentiment spread index predicting unit passes through (7) prediction the according to the following formula of initial Susceptible population's quantity
One maximum public sentiment spread index, the described first maximum public sentiment spread index indicate that public sentiment event breaks out the maximum side that can reach
Boundary's number, i.e. whole event Onset forward cumulative amount
Wherein, E0Indicate initially accumulative transfer amount, EsFor the first maximum public sentiment spread index.
Furthermore it is preferred that above- mentioned information transmission dynamics system further includes the first model training module, to the first model construction
The information transmission dynamics model of module building is trained, comprising:
First training set construction unit acquires the data of microblogging setting time as training set;
The training set of first training set construction unit building is inputted the first model construction module by the first parameter training unit
The information transmission dynamics model of building, using the parameter of Least Square Method information transmission dynamics model
Wherein, LS is the square error of parameter, and Θ is information transmission dynamics model parameter vectors, Θ=(β, α1, p1, S0),
ID indicates the cumulative actual transfer amount of training sample, tk=k, k=1,2 ... T indicates sampling time, fE(tk, Θ) and it is moment tk
Training sample input parameter vector Θ the obtained accumulative transfer amount of information transmission dynamics model.
The construction method of the above-mentioned information transmission dynamics system based on microblogging is as shown in figure 3, the construction method includes:
The state of user in microblogging is divided into three kinds of sensitization, forwarding state and immune state states by step S1,
In, the sensitization indicates that user not yet contacts issuing microblog but has the state that microblogging is forwarded ability;Described turn
The expression of hair-like state has been carried out forwarding and can be to the state that other people have an impact in active state;The immune state
It indicates the state for losing the ability of enlivening after user forwards or actively abandons the state of forwarding microblogging after reading information;
Step S2 constructs information transmission dynamics model by formula (1)-(4);
Step S3, by the monitoring to accumulative transfer amount according to above- mentioned information transmission dynamics model to Susceptible population's quantity,
Forwarding crowd quantity and immune crowd's quantity and its variation tendency (whether index derivative is greater than 0) are predicted.
In step s3, preferably further include the steps that predicting public sentiment by public sentiment index, the public sentiment index includes the
The maximum public sentiment spread index of one public sentiment explosive index, the first public sentiment propagation peak and first, the step include:
The first public sentiment explosive index is predicted according to formula (5) by initial Susceptible population's quantity, and first public sentiment is quick-fried
It sends out exponential representation public sentiment and breaks out the quantity that the stage actively forwards the forwarding user of customer impact by one, also illustrate that public sentiment event is quick-fried
The severity of hair;
Judge that the first public sentiment explosive index is greater than 1, is also equal to 1 less than 1;
IfShow to forward number of users decline, microblogging will be unable to further propagate;If
Show to forward number of users that will exponentially increase;IfShow to forward number of users without change;
The first public sentiment propagation peak F is predicted according to formula (6) by initial Susceptible population's quantitymax, first public sentiment
Propagation peak indicates the maximum value of user's forwarding behavior increment within the unit time when outburst of public sentiment event;
Forwarding number of users is broken out as the first public sentiment at the time of increase to first public sentiment the first setting ratio of propagation peak
Initial time tb, the prediction that initial time is broken out to the first public sentiment is realized by formula (6);
Forwarding number of users is broken out into peak value moment t as the first public sentiment at the time of increase to the first public sentiment propagation peakmax,
The prediction that peak value moment is broken out to the first public sentiment is realized by formula (6);
It is broken out at the time of forwarding number of users is dropped to first public sentiment the second setting ratio of propagation peak as the first public sentiment
Finish time te, the prediction that finish time is broken out to the first public sentiment is realized by formula (6);
By the forwarding number of users of the unit time of the first public sentiment outburst initial time to the first public sentiment outburst peak value moment
Speed V is broken out as the first public sentimento, realize the prediction to the first public sentiment outburst speed
Wherein, r1For the first setting ratio;
By the forwarding number of users of the unit time of the first public sentiment outburst peak value moment to the first public sentiment outburst finish time
As the first public sentiment decline rate Vd, realize the prediction to the first public sentiment decline rate
Wherein, r2For the second setting ratio;
By the forwarding number of users of the unit time of the first public sentiment outburst initial time to the first public sentiment outburst finish time
Rate V is averagely broken out as firsta, realize the prediction that rate is averagely broken out to the first public sentiment
By initial Susceptible population's quantity according to the maximum public sentiment spread index of formula (7) prediction first, described first most
The public sentiment event that big public sentiment spread index indicates breaks out the maximum boundary number that can reach, i.e. whole event Onset forwarding is tired
Metering.
Preferably, the construction method further includes the training step of the information transmission dynamics model, and the step includes:
The data of microblogging setting time are acquired as training set, input information transmission dynamics model;
Using the parameter of Least Square Method information transmission dynamics model.
Microblogging propagate final result just only and parameter S0It is related.It is however noted that not with epidemiological study
Together, susceptible number of users initial in microblogging is usually unknown, it is preferable that when event starts, the initial user of microblogging is issued,
Only publisher's issuing microblog sets primary condition: E0=F0=1, I0=0, S0=N-F0=N-1, N indicate closed in spaces, table
Show the set for being likely to be exposed all users of this microblogging.
Above- mentioned information transmission dynamics system and its construction method also construct public sentiment index, realize and pass through information transmission dynamics mould
Type obtains the collateral information that can more characterize public sentiment propagation condition.
In another embodiment of the present invention, the outburst of public sentiment event is frequently accompanied by the generation of many topics, and one
Topic is made of many information, this makes the discussion of event and propagates efficient and information concentration.Topic amount of reading is as microblogging pair
Another important information of outer publication contributes an amount of reading when a user browses to a microblogging, that is, enters read state,
Since user may repeat repeatedly to read, i.e., susceptible person is become with certain ratio again, then remaining user is no longer right
Public sentiment has of interest i.e. into immune state.Whether user selects forwarding simultaneously, it is built upon on the basis of having read, then
Amount of reading can describe the scale and concerned degree of event outburst to a certain extent.In view of user may be to the same topic
It is repeatedly read, i.e., user behavior has repeatability, intends the information transmission dynamics model SRI of single topic propagation of building as schemed
Shown in 5, information transmission dynamics system can be constructed using multiple microbloggings of a topic, enhancement information transmission dynamics model adapts to
Popularity, information transmission dynamics system can also be constructed using each microblogging, when multiple microbloggings building using topic
When information transmission dynamics system, add up the sum of the accumulative amount of reading of multiple microbloggings that amount of reading is the topic, other parameters (S
(t), F (t), I (t)) and so on.
Fig. 6 is the composition block diagram of another embodiment of the information transmission dynamics system the present invention is based on microblogging, such as Fig. 6 institute
Show, the information transmission dynamics system includes:
The state of user in microblogging is divided into sensitization (S), read state (R) by 1 ' of second user state setting module
With three kinds of states of immune state (I), wherein the sensitization indicate user not yet contact issuing microblog still have microblogging into
The state of row reading ability;The read state expression has been carried out reading and can generate to other people in active state
The state of influence;The immune state loses the state for the ability of enlivening after indicating user's reading or has exceeded reading validity period
State;
Second model construction module, 2 ' constructs information transmission dynamics model by following formula (8)-(11)
S ' (t)=- β S (t) R (t)+p2α2R(t) (8)
R ' (t)=β S (t) R (t)-α2R(t) (9)
I ' (t)=(1-p2)α2R(t) (10)
E′2(t)=β S (t) R (t) (11)
Wherein, R (t) is reading crowd quantity, E2It (t) is accumulative amount of reading, p2Indicate that reading user generates repeat reading
The rate of behavior, α2Indicate that user becomes sluggish rate, S ' (t), F ' (t), I ' (t) and E ' in read state2(t) respectively
For S (t), F (t), I (t) and E2(t) derivative, the derivative, which is greater than 0, indicates that index value increases, and derivative indicates index less than 0
Numerical value reduces, and derivative, which is equal to 0, indicates that index value is constant;
Second prediction module, 3 ', the letter constructed by the monitoring to accumulative amount of reading according to above-mentioned second model construction module
Cease the quantity and variation of the user in different states of transmission dynamics model prediction second user state setting module setting.
Preferably, the second prediction module also passes through public sentiment index prediction public sentiment, and the public sentiment index includes that the second public sentiment is quick-fried
Several, the second public sentiment propagation peak of bristling with anger and the second maximum public sentiment spread index, second prediction module further include:
Second public sentiment explosive index predicting unit predicts the second carriage according to the following formula (12) by initial Susceptible population's quantity
Feelings explosive index, the second public sentiment explosive index indicate the severity of public sentiment event outburst
Wherein :=indicate to be defined as,For the second public sentiment explosive index, E20Indicate initially accumulative amount of reading, S0It indicates
The initial value of number of users in sensitization;
Judge that the second public sentiment explosive index is greater than 1, is also equal to 1 less than 1;
IfShow to read number of users decline, microblogging will be unable to further propagate;If
Show that reading number of users will exponentially increase;IfShow to read number of users without change;
Second public sentiment propagation peak predicting unit predicts the second carriage according to the following formula (13) by initial Susceptible population's quantity
Feelings propagation peak Rmax, user's reading behavior is within the unit time when the second public sentiment propagation peak expression public sentiment event is broken out
The maximum value of increment
Second time prediction unit, at the time of number of users being read and increase to the second public sentiment propagation peak third setting ratio
Initial time is broken out as the second public sentiment, the second public sentiment outburst initial time is predicted;Number of users will be read to increase to the
Peak value moment is broken out as the second public sentiment at the time of two public sentiment propagation peaks, the second public sentiment outburst peak value moment is predicted;
It will read and break out finish time as the second public sentiment at the time of number of users drops to the second public sentiment four setting ratio of propagation peak,
Second public sentiment outburst finish time is predicted;
Second speed predicting unit, when the second public sentiment is broken out unit of the initial time to the second public sentiment outburst peak value moment
Between reading number of users as the second public sentiment break out speed, realize to the second public sentiment outburst speed prediction;By the second public sentiment
The reading number of users of outburst peak value moment to the unit time that the second public sentiment breaks out finish time are fast as the decline of the second public sentiment
Degree realizes the prediction to the second public sentiment decline rate;Second public sentiment is broken out into initial time and breaks out finish time to the second public sentiment
Unit time reading number of users as second averagely break out rate, realization the pre- of rate is averagely broken out to the second public sentiment
It surveys;
Second maximum public sentiment spread index predicting unit passes through (14) building the according to the following formula of initial Susceptible population's quantity
Two maximum public sentiment spread indexs, the described second maximum public sentiment spread index indicate that public sentiment event breaks out the maximum side that can reach
Boundary's number, i.e. whole event Onset read cumulative amount
Wherein, E2sFor the second maximum public sentiment spread index.
Preferably, above- mentioned information transmission dynamics system further includes the second model training module, to the second model construction module
The information transmission dynamics model of building is trained, comprising:
Second training set construction unit acquires the data of microblogging setting time as training set;
Second parameter training unit, the training set that training set construction unit is constructed input information transmission dynamics model, adopt
With the parameter of Least Square Method information transmission dynamics model
Wherein, LS is the square error of parameter, and Θ ' is information transmission dynamics model parameter vectors, Θ '=(β, α2, p2,
S0), IDk' indicates training sample moment tkCumulative actual amount of reading, tk=k, k=1,2 ... T indicates sampling time, fE(tk,
Θ ') it is moment tkTraining sample input parameter vector Θ ' the obtained accumulative amount of reading of information transmission dynamics model.
The construction method construction method of above- mentioned information transmission dynamics system includes:
The state of user in microblogging is divided into three kinds of sensitization, read state and immune state states, wherein described easy
Sense state indicates that user not yet contacts issuing microblog but has the state that microblogging is forwarded ability;The read state indicates
Have been carried out reading and in active state can be to the state that other people have an impact;The immune state indicates that user reads
The state for the ability of enlivening is lost after reading or has exceeded the state for reading validity period;
Information transmission dynamics model is constructed by formula (8)-(11);
By the monitoring to accumulative amount of reading according to above- mentioned information transmission dynamics model to Susceptible population's quantity, reading crowd
Quantity and immune crowd's quantity and its variation are predicted.
Preferably, further include the steps that predicting that public sentiment, the public sentiment index are broken out including the second public sentiment by public sentiment index
The maximum public sentiment spread index of index, the second public sentiment propagation peak and second, the schematic diagram of public sentiment index is as shown in fig. 7, the step
Suddenly include:
The second public sentiment explosive index is predicted according to formula (12) by initial Susceptible population's quantity, and second public sentiment is quick-fried
Send out the severity of exponential representation public sentiment event outburst;
IfShow to read number of users decline, microblogging will be unable to further propagate;If
Show that reading number of users will exponentially increase;IfShow to read number of users without change;
The second public sentiment propagation peak R is predicted according to formula (13) by initial Susceptible population's quantitymax, second carriage
Feelings propagation peak indicates the maximum value of user's reading behavior increment within the unit time when outburst of public sentiment event;
It will read at the time of number of users increases to the second public sentiment propagation peak third setting ratio and be broken out as the second public sentiment
Initial time tb' realizes the prediction that initial time is broken out to the second public sentiment by formula (13);
It will read and break out peak value moment t as the second public sentiment at the time of number of users increases to the second public sentiment propagation peakmax',
The prediction that peak value moment is broken out to the second public sentiment is realized by formula (13);
It will read at the time of number of users drops to the second public sentiment four setting ratio of propagation peak and broken out as the second public sentiment
Finish time te' realizes the prediction that finish time is broken out to the second public sentiment by formula (13);
By the reading number of users of the unit time of the second public sentiment outburst initial time to the second public sentiment outburst peak value moment
Speed V is broken out as the second public sentimento' realizes the prediction to the second public sentiment outburst speed
Wherein, r3For third setting ratio;
By the reading number of users of the unit time of the second public sentiment outburst peak value moment to the second public sentiment outburst finish time
As the second public sentiment decline rate Vd' realizes the prediction to the second public sentiment decline rate
Wherein, r4For the 4th setting ratio;
By the reading number of users of the unit time of the second public sentiment outburst initial time to the second public sentiment outburst finish time
Rate V is averagely broken out as seconda' realizes the prediction that rate is averagely broken out to the second public sentiment
By initial Susceptible population's quantity according to the maximum public sentiment spread index of formula (14) building second, described second most
The public sentiment event that big public sentiment spread index indicates breaks out the maximum boundary number that can reach, i.e. whole event Onset is read tired
Metering.
Preferably, further include the steps that training information transmission dynamics model, the step includes: acquisition microblogging setting time
Data as training set, input information transmission dynamics model;Using the ginseng of Least Square Method information transmission dynamics model
Number.
The public sentiment explosive index of one topicIt is by parameter beta, α2And S0It determines, and public sentiment propagation peak RmaxMost
Big public sentiment spread index EsIt is by parameter beta, α2,S0And p2It is different from epidemiological study come what is determined, it is initial easy in microblogging
Feel number of users S0Usually it is unknown preferably, it is preferable that setting primary condition are as follows: E20=R0< <, I0=0, S0=N-R0。
In the various embodiments of the invention, first setting ratio, the second setting ratio, third setting ratio and the 4th
Setting ratio can be the same or different, for example, can be 0.1.
In the various embodiments described above of the invention, in Chinese public sentiment field, a maximum problem is exactly our needs
The data unit (hour) for adding up reading or forwarding number of users in model parameter can be estimated, so as to quick-fried to event
The propagation trend of hair early stage is predicted and calculates critical public sentiment index.Therefore, if the mistake of event is only used only in we
Data are gone to estimate all model parameters, model can there is very big then our single microblogging SFI model is with list topic SRI
Uncertain and unstability, it is preferred that being estimated when predicting public sentiment using partial parameters priori (a-priori)
Method, specifically, comprising:
The topological diagram that micro-blog information is propagated is constructed, as shown in figure 8, the topological mode that message is propagated, comprising: star-like, constellation
One or more of type and nebula type, wherein central point represents leader of opinion, around central point reading or forwarded correspondence
Leader of opinion bean vermicelli, the line between point represents concern relation;
The similarity of premonitoring public sentiment with the public sentiment that finished is judged according to the registration of leader of opinion, what leader of opinion was overlapped
More, similarity is higher;
The public sentiment that finished is ranked up according to the sequence of similarity from high to low, the public sentiment that finished of setting quantity before taking
Model parameter average value as with monitoring public sentiment model parameter, the mould of the highest public sentiment that finished of similarity can also be taken
Shape parameter is as the model parameter with monitoring public sentiment.
Preferably, model parameter β is determined by the compactedness of network structure, thus can when event starts just to its into
Row estimation, model parameter β indicates mean exposure rate, because the susceptible number of users and entire subnet of whole event are (in topological diagram not
With the topological subgraph of leader of opinion) compactedness just had determined that when topic starts to occur.
Furthermore it is preferred that model parameter α is usually the specific attribute of user rather than some topic or microblogging institute are peculiar
's.We are difficult to change the habit of user, because several talentes of some users can browse a microblogging, but some users are often short
Several hours just select to open Sina weibo.Therefore, identical as model parameter β, we can be in the early stage (example of event
Such as, from other events) remove estimation model parameter α.
In one particular embodiment of the present invention, the Sina weibo social activity most as Chinese current application user is flat
The propagation of platform, various information focuses primarily upon microblogging, so Sina weibo researchs and analyses the control propagated today's society information
System has great significance.Using a microblogging public sentiment in information transmission dynamics model SFI monitoring Sina weibo, comprising:
It goes to collect the quantity for accumulating forwarding user by microblogging API, as shown in table 1 below,
Table 1
Table 1 lists the data set of the accumulation transfer amount for the real event that a time span is 16 days;
The value range that model parameter is obtained using the model parameter of knowledge professional library or similar microblogging, in example as above, β
∈[2×10-6, 8 × 10-6], p1∈ [0.1,0.5], α1∈ [0.5,1.5], S0∈[7×106, 2 × 107];
According to above-mentioned value range, the initial value Θ of setting model parameter0=(4.0 × 10-6, 0.3,1.0,1.0 ×
107), using least square method obtain information transmission dynamics model parameter, and to S (t), F (t), I (t) sum E (t) curve into
Row fitting, as a result as shown in figure 9, β=2.5651 × 10-6, α1=1.0365, p1=0.1006, S0=1.0221 × 107;
Public sentiment is monitored in conjunction with above-mentioned curve and public sentiment index, as shown in figure 9, Fmax=
2.4724×105, Es=0.9243 × 106, tb=5.92 (day), te=12.57day), tmax=8.49 (day), Vo=
0.8658×105/ day, Vd=5.4538 × 104/ day, Va=3.3461 × 104/day。
Preferably, following public sentiment index can be predicted using the accumulation forwarding number of users of setting time,
For example, the accumulation forwarding number of users of first day, the 8th day second day ... is respectively adopted, using the above method to the model of SFI
Parameter and public sentiment index predicted, as shown in Figure 10, it can be seen that model parameter can be estimated from preceding 3 days data
Out, this is much earlier than peak explosion time, can not only indicate that first 3 days earliest pre-warning times occur for outgoing event, but also for
The prediction of each public sentiment index is also more ideal.Especially, after the 8th day close to public sentiment outburst peak value, we can be quasi-
Really predict our interested all indexes.
Further, it is preferable that using preceding accumulation in 3 days forwarding number of users as training set, obtained by least square method
Model parameter is obtained, so that the public sentiment index to future time predicts that prediction result is as shown in following table 2-4
Table 2
Table 3
Table 4
In another specific embodiment of the invention, public sentiment is predicted using information transmission dynamics model SRI, comprising:
It goes to collect by microblogging API and accumulates the quantity for reading user, as shown in table 5 below,
Table 5
Table 5 lists the data set of the accumulation amount of reading of a length of 4 hours topics when four acquisitions;
The initial value of setting model parameter, using least square method obtain information transmission dynamics model parameter, and to S (t),
F (t), I (t) and E2(t) curve is fitted, as a result as shown in figure 11, β=2.5791 × 10-7, α2=0.22122, p2=
0.10607, S0=3240000;
Public sentiment is monitored in conjunction with above-mentioned curve and public sentiment index, as shown in figure 11,Rmax=
1.3023×106, E2s=3.5307 ××s 106, tb'=23.12 (h), tmax=51.68 (h), te'=113.88 (h), Vo'=
4.1039×104(/h), Va'=1.2914 × 104(/h), Vd'=1.8844 × 104(/h)。
Preferably, predicting following public sentiment index for number of users can be read using the accumulation of setting time,
For example, with 8 hours accumulative amount of reading using frequency collection microblogging, using the above method to the model parameter and public sentiment of SRI
Index predicted, as shown in figure 12,3 days data can estimate Prediction Parameters (p before topic2, S0), and this when
Between put this much earlier than peak explosion time.It was found that the parameter of estimation and SRI model are combined pre- to carry out public sentiment
It surveys, real data can be ideally fitted, until there is turning point, such as time point 16 hours, 40 hours and 72 hours.Total comes
Say, after there is variation tendency in information communication process, SRI model can in 8 hours Fast Fitting, therefore have to early prediction
Good effect.
Further, it is preferable that number of users is read as training set using preceding accumulation in 40 hours, passes through least square method
Model parameter is obtained, so that the public sentiment index to future time predicts that prediction result is as shown in following table 6-8
Table 6
Table 7
Table 8
From table 6-8 can be seen that using the past 40 hours data go prediction when, public sentiment explosive indexAnd public sentiment
Break out speed (Vo') all there is lower prediction accuracy, still, initial time (t is broken out for public sentimentb') but have quite not
Wrong prediction effect for 24 hours.In particular, if propagation peak time (tmax'=50.08h) after 40 hours that public sentiment occurs,
Our partial parameters apriority can accurately predict it.Regrettably, it within public sentiment is broken out 56 hours, can only obtain
The lower public sentiment of one accuracy breaks out peak value (Rmax), but this is still worth because need not until public sentiment it is final when
Between point (110h), can know earlier public sentiment outburst peak value quantity.In addition, breaking out end time (t for public sentimente') or
Person is public sentiment decay rates (Vd'), maximum public sentiment spread index (E2s) and averagely outburst rate (VaFor '), optimal early stage
Predicted time point is 70 hours, 78 hours and 62 hours respectively.
Generally speaking, the partial parameters prior estimate method that the present invention is proposed for early prediction can help us to be advanced by
The progress of microblog topic is solved, therefore, network public-opinion can be monitored and controlled in we in time.
Information transmission dynamics system, construction method, device and medium of the present invention based on microblogging passes through cross discipline,
By mathematical theory be applied to the analysis of public opinion, Information dynamics with public sentiment event information propagation combine, Information dynamics analysis and
News public sentiment event analysis, which combines, establishes model.For the exclusive information propagating characteristic and different user's rows of Sina weibo
To be modeled respectively to the reading behavior of user and forwarding behavior, passing through transfer amount to single microblogging and single topic
Amount of reading constructs parameter and carries out parameter Estimation, and the indexs such as outburst rate, explosion time for breaking out to public sentiment are analyzed, and building is deposited
In the new public sentiment ecological models of three kinds of spread states, to be applied to the prediction, early warning and control analysis of public sentiment.
The various embodiments described above are respectively described SRI and SFI information transmission dynamics system, but of the invention and unlimited
In this, information transmission dynamics system of the invention can use the combination of SRI and SFI, to forwarding number it is predicted that while to readding
It reads data to be predicted, can also need to assign two kinds of different weights of model according to different, for example, laying particular stress on forwarding data
When (the public sentiment index of forwarding number and SFI), so that the weight of SFI is greater than the weight of SRI.
In addition, the present invention also provides a kind of electronic device, including memory and processor, base is stored in the memory
In the information transmission dynamics system construction procedures of microblogging, the information transmission dynamics system construction procedures based on microblogging are described
Processor realizes the step of the above-mentioned information transmission dynamics system constituting method based on microblogging when executing.
In the present embodiment, memory includes the readable storage medium storing program for executing of at least one type.At least one type
Readable storage medium storing program for executing can be the non-volatile memory medium of such as flash memory, hard disk, multimedia card, card-type memory.In some realities
It applies in example, the readable storage medium storing program for executing can be the internal storage unit of the electronic device, such as the hard disk of the electronic device.
In further embodiments, the readable storage medium storing program for executing is also possible to the external memory of the electronic device, such as the electricity
The plug-in type hard disk being equipped in sub-device, intelligent memory card (Smart Media Card, SMC), secure digital (Secure
Digital, SD) card, flash card (Flash Card) etc., the memory can be also used for temporarily storing exported or
The data that will be exported.
Processor can be a central processing unit (Central Processing Unit, CPU) in some embodiments,
Microprocessor or other data processing chips, program code or processing data for being stored in run memory.
Preferably, electronic device further includes network interface, optionally may include standard wireline interface and wireless interface
(such as WI-FI interface), commonly used in establishing communication connection between the electronic device, with other electronic equipments;Communication bus is used
Connection communication between these components of realization.
In addition, the embodiment of the present invention also proposes a kind of computer readable storage medium, the computer readable storage medium
In include the information transmission dynamics system construction procedures based on microblogging, it is described based on microblogging information transmission dynamics system building
When program is executed by processor, realize the above-mentioned information transmission dynamics system constituting method based on microblogging the step of.
The electronic device of the present invention and the specific embodiment of computer readable storage medium and the above-mentioned letter based on microblogging
Breath transmission dynamics system constituting method, the specific embodiment of system are roughly the same, and details are not described herein.
Although content disclosed above shows exemplary embodiment of the present invention, it should be noted that without departing substantially from power
Under the premise of benefit requires the range limited, it may be many modifications and modify.According to the side of inventive embodiments described herein
Function, step and/or the movement of method claim are not required to the execution of any particular order.In addition, although element of the invention can
It is unless explicitly limited individual element it is also contemplated that having multiple elements to be described or be required in the form of individual.
Claims (10)
1. a kind of information transmission dynamics system constituting method based on microblogging characterized by comprising
The state of user in microblogging is divided into three kinds of sensitization, forwarding state and immune state states, wherein the susceptible shape
State indicates that user not yet contacts issuing microblog but has the state that microblogging is forwarded ability;The forwarding state indicates
Carried out forward and in active state can be to the state that other people have an impact;After the immune state indicates user's forwarding
Lose the ability of enlivening state or reading information after actively abandon forwarding microblogging state;
Information transmission dynamics model is constructed by following formula (1)-(4)
S ' (t)=- β S (t) F (t) (1)
F ' (t)=p1βS(t)F(t)-α1F(t) (2)
I ' (t)=(1-p1)βS(t)F(t)+α1F(t) (3)
E ' (t)=p1βS(t)F(t) (4)
Wherein, t is time index, and β indicates the user in sensitization in the rate of the mean exposure information of microblogging, p1It indicates
User in sensitization forwards the rate of microblogging, α1Indicate that the user in forwarding state becomes sluggish rate, S
It (t) is Susceptible population's quantity, F (t) is forwarding crowd quantity, and I (t) is immune crowd's quantity, and E (t) is accumulative transfer amount, S '
(t), F ' (t), I ' (t) and E ' (t) are respectively the derivative of S (t), F (t), I (t) and E (t), and the derivative, which is greater than 0, indicates index
Numerical value increases, and derivative indicates that index value reduces less than 0, and derivative, which is equal to 0, indicates that index value is constant;
By the monitoring to accumulative transfer amount according to above- mentioned information transmission dynamics model to following Susceptible population's quantity, forwarding people
Group's quantity and immune crowd's quantity and its variation are predicted.
2. the information transmission dynamics system constituting method according to claim 1 based on microblogging, which is characterized in that further include
The step of predicting public sentiment by public sentiment index, the public sentiment index include the first public sentiment explosive index, the first public sentiment propagation peak
With the first maximum public sentiment spread index, the step includes:
The first public sentiment explosive index is predicted according to the following formula (5) by initial Susceptible population's quantity, and the first public sentiment outburst refers to
Number indicates that the public sentiment outburst stage actively forwards the quantity of the forwarding user of customer impact by one, also illustrates that the outburst of public sentiment event
Severity
Wherein :=indicate to be defined as,For the first public sentiment explosive index, S0Indicate the first of the number of users in sensitization
Initial value;
Judge that the first public sentiment explosive index is greater than 1, is also equal to 1 less than 1;
IfShow to forward number of users decline, microblogging will be unable to further propagate;IfShow
Forwarding number of users will exponentially increase;IfShow to forward number of users without change;
The first public sentiment propagation peak F is predicted according to the following formula (6) by initial Susceptible population's quantitymax, the first public sentiment propagation
Peak value indicates the maximum value of user's forwarding behavior increment within the unit time when outburst of public sentiment event
Wherein, F0Indicate the initial value of the number of users in forwarding state;
Forwarding number of users is broken out as the first public sentiment at the time of increase to first public sentiment the first setting ratio of propagation peak and is originated
Moment predicts the first public sentiment outburst initial time;
Forwarding number of users is broken out into peak value moment as the first public sentiment at the time of increase to the first public sentiment propagation peak, to the first carriage
Feelings outburst peak value moment is predicted;
Terminate at the time of forwarding number of users is dropped to first public sentiment the second setting ratio of propagation peak as the outburst of the first public sentiment
Moment predicts the first public sentiment outburst finish time;
Using the first public sentiment break out initial time to the first public sentiment outburst peak value moment unit time forwarding number of users as
First public sentiment breaks out speed, realizes the prediction to the first public sentiment outburst speed;
Using the first public sentiment break out peak value moment to the first public sentiment outburst finish time unit time forwarding number of users as
First public sentiment decline rate realizes the prediction to the first public sentiment decline rate;
Using the first public sentiment break out initial time to the first public sentiment outburst finish time unit time forwarding number of users as
First averagely breaks out rate, realizes the prediction that rate is averagely broken out to the first public sentiment;
The first maximum public sentiment spread index, the described first maximum carriage are predicted according to the following formula (7) by initial Susceptible population's quantity
The public sentiment event that feelings spread index indicates breaks out the maximum boundary number that can reach, i.e. whole event Onset forwards cumulative amount
Wherein, E0Indicate initially accumulative transfer amount, EsFor the first maximum public sentiment spread index.
3. the information transmission dynamics system constituting method according to claim 1 based on microblogging, which is characterized in that the letter
Breath transmission dynamics model training method include:
The data of microblogging setting time are acquired as training set, input information transmission dynamics model;
Using the parameter of Least Square Method information transmission dynamics model
Wherein, LS is the square error of parameter, and Θ is information transmission dynamics model parameter vectors, Θ=(β, α1, p1, S0), ID table
Show the cumulative actual transfer amount of training sample, tk=k, k=1,2 ... T indicates sampling time, fE(tk, Θ) and it is moment tk's
The accumulative transfer amount that the information transmission dynamics model of training sample input parameter vector Θ obtains.
4. a kind of information transmission dynamics system constituting method based on microblogging characterized by comprising
The state of user in microblogging is divided into three kinds of sensitization, read state and immune state states, wherein the susceptible shape
State indicates that user not yet contacts issuing microblog but has the state that microblogging carries out reading ability;The read state indicates
Carried out read and in active state can be to the state that other people have an impact;After the immune state indicates that user reads
It loses the state for the ability of enlivening or has exceeded the state for reading validity period;
Information transmission dynamics model is constructed by following formula (8)-(11)
S ' (t)=- β S (t) R (t)+p2α2R(t) (8)
R ' (t)=β S (t) R (t)-α2R(t) (9)
I ' (t)=(1-p2)α2R(t) (10)
E′2(t)=β S (t) R (t) (11)
Wherein, R (t) is reading crowd quantity, E2It (t) is accumulative amount of reading, p2Indicate that reading user generates repeat reading behavior
Rate, α2Indicate that user becomes sluggish rate, S ' (t), F ' (t), I ' (t) and E ' in read state2(t) be respectively S (t),
F (t), I (t) and E2(t) derivative, the derivative, which is greater than 0, indicates that index value increases, and derivative subtracts less than 0 expression index value
Small, derivative, which is equal to 0, indicates that index value is constant;
By the monitoring to accumulative amount of reading according to above- mentioned information transmission dynamics model to following Susceptible population's quantity, reader
Group's quantity and immune crowd's quantity and its variation are predicted.
5. the information transmission dynamics system constituting method according to claim 4 based on microblogging, which is characterized in that further include
The step of predicting public sentiment by public sentiment index, the public sentiment index include the second public sentiment explosive index, the second public sentiment propagation peak
With the second maximum public sentiment spread index, the step includes:
The second public sentiment explosive index is predicted according to the following formula (12) by initial Susceptible population's quantity, and the second public sentiment outburst refers to
Number indicates the severity of public sentiment event outburst
Wherein :=indicate to be defined as,For the second public sentiment explosive index, E20Indicate initially accumulative amount of reading, S0Expression is in
The initial value of the number of users of sensitization;
Judge that the second public sentiment explosive index is greater than 1, is also equal to 1 less than 1;
IfShow to read number of users decline, microblogging will be unable to further propagate;IfShow
Reading number of users will exponentially increase;IfShow to read number of users without change;
The second public sentiment propagation peak R is predicted according to the following formula (13) by initial Susceptible population's quantitymax, the second public sentiment biography
Broadcasting peak value indicates the maximum value of user's reading behavior increment within the unit time when outburst of public sentiment event
It will read to break out at the time of number of users increases to the second public sentiment propagation peak third setting ratio as the second public sentiment and originate
Moment predicts the second public sentiment outburst initial time;
It will read and break out peak value moment as the second public sentiment at the time of number of users increases to the second public sentiment propagation peak, to the second carriage
Feelings outburst peak value moment is predicted;
To read at the time of number of users drops to the second public sentiment four setting ratio of propagation peak terminates as the outburst of the second public sentiment
Moment predicts the second public sentiment outburst finish time;
Using the second public sentiment break out initial time to the second public sentiment outburst peak value moment unit time reading number of users as
Second public sentiment breaks out speed, realizes the prediction to the second public sentiment outburst speed;
Using the second public sentiment break out peak value moment to the second public sentiment outburst finish time unit time reading number of users as
Second public sentiment decline rate realizes the prediction to the second public sentiment decline rate;
Using the second public sentiment break out initial time to the second public sentiment outburst finish time unit time reading number of users as
Second averagely breaks out rate, realizes the prediction that rate is averagely broken out to the second public sentiment;
The second maximum public sentiment spread index, the described second maximum carriage are constructed according to the following formula (14) by initial Susceptible population's quantity
The public sentiment event that feelings spread index indicates breaks out the maximum boundary number that can reach, i.e. whole event Onset reads cumulative amount
Wherein, E2sFor the second maximum public sentiment spread index.
6. a kind of information transmission dynamics system based on microblogging characterized by comprising
The state of user in microblogging is divided into three kinds of sensitization, forwarding state and immune state shapes by User Status setting module
State, wherein the sensitization indicates that user not yet contacts issuing microblog but has the state that microblogging is forwarded ability;Institute
State forwarding state expression have been carried out forwarding and in active state can be to the state that other people have an impact;It is described immune
State indicate user forwarding after lose the ability of enlivening state or reading information after actively abandon forwarding microblogging state;
Model construction module constructs information transmission dynamics model by following formula (1)-(4)
S ' (t)=- β S (t) F (t) (1)
F ' (t)=p1βS(t)F(t)-α1F(t) (2)
I ' (t)=(1-p1)βS(t)F(t)+α1F(t) (3)
E ' (t)=p1βS(t)F(t) (4)
Wherein, t is time index, and β indicates the user in sensitization in the rate of the mean exposure information of microblogging, p1It indicates
User in sensitization forwards the rate of microblogging, α1Indicate that the user in forwarding state becomes sluggish rate, S
It (t) is Susceptible population's quantity, F (t) is forwarding crowd quantity, and I (t) is immune crowd's quantity, and E (t) is accumulative transfer amount, S '
(t), F ' (t), I ' (t) and E ' (t) are respectively the derivative of S (t), F (t), I (t) and E (t), and the derivative, which is greater than 0, indicates index
Numerical value increases, and derivative indicates that index value reduces less than 0, and derivative, which is equal to 0, indicates that index value is constant;
Prediction module, the information transmission dynamics model constructed by the monitoring to accumulative transfer amount according to above-mentioned model construction module
Predict the quantity and variation of the user in different states of User Status setting module setting.
7. the information transmission dynamics system according to claim 6 based on microblogging, which is characterized in that the prediction module is also
Predict that public sentiment, the public sentiment index include the first public sentiment explosive index, the first public sentiment propagation peak and first by public sentiment index
Maximum public sentiment spread index, the prediction module include:
First public sentiment explosive index predicting unit predicts that the first public sentiment is quick-fried according to the following formula (5) by initial Susceptible population's quantity
Bristle with anger number, the first public sentiment explosive index indicates the forwarding user's that the public sentiment outburst stage actively forwards customer impact by one
Quantity also illustrates that the severity of public sentiment event outburst
Wherein :=indicate to be defined as,For the first public sentiment explosive index, S0Indicate the first of the number of users in sensitization
Initial value;
Judge that the first public sentiment explosive index is greater than 1, is also equal to 1 less than 1;
IfShow to forward number of users decline, microblogging will be unable to further propagate;IfShow
Forwarding number of users will exponentially increase;IfShow to forward number of users without change;
First public sentiment propagation peak predicting unit predicts that the first public sentiment passes according to the following formula (6) by initial Susceptible population's quantity
Broadcast peak Fmax, user forwards behavior increment within the unit time when the first public sentiment propagation peak expression public sentiment event is broken out
Maximum value
Wherein, F0Indicate the initial value of the number of users in forwarding state;
First time predicting unit, will forwarding number of users at the time of increase to first public sentiment the first setting ratio of propagation peak as
First public sentiment breaks out initial time, predicts the first public sentiment outburst initial time;Forwarding number of users is increased to the first carriage
Peak value moment is broken out as the first public sentiment at the time of feelings propagation peak, the first public sentiment outburst peak value moment is predicted;It will turn
Finish time is broken out as the first public sentiment at the time of hair number of users drops to first public sentiment the second setting ratio of propagation peak, to the
One public sentiment outburst finish time is predicted;
First Speed predicting unit, by the unit time of the first public sentiment outburst initial time to the first public sentiment outburst peak value moment
It forwards number of users to break out speed as the first public sentiment, realizes the prediction to the first public sentiment outburst speed;First public sentiment is broken out
Peak value moment to the first public sentiment outburst finish time unit time forwarding number of users as the first public sentiment decline rate, it is real
Now to the prediction of the first public sentiment decline rate;By the unit of the first public sentiment outburst initial time to the first public sentiment outburst finish time
The forwarding number of users of time averagely breaks out rate as first, realizes the prediction that rate is averagely broken out to the first public sentiment;
First maximum public sentiment spread index predicting unit, by initial Susceptible population's quantity, (7) prediction first is most according to the following formula
Big public sentiment spread index, the described first maximum public sentiment spread index indicate that public sentiment event breaks out the maximum boundary that can reach
Number, i.e. whole event Onset forward cumulative amount
Wherein, E0Indicate initially accumulative transfer amount, EsFor the first maximum public sentiment spread index.
8. the information transmission dynamics system according to claim 6 based on microblogging, which is characterized in that the User Status is set
The state of user in microblogging is divided into three kinds of sensitization, read state and immune state states by cover half block, wherein described susceptible
State indicates that user not yet contacts issuing microblog but has the state that microblogging carries out reading ability;The read state indicates
Carried out read and in active state can be to the state that other people have an impact;The immune state indicates that user reads
The state for the ability of enlivening is lost afterwards or has exceeded the state for reading validity period;
The model construction module constructs information transmission dynamics model by following formula (8)-(11)
S ' (t)=- β S (t) R (t)+p2α2R(t) (8)
R ' (t)=β S (t) R (t)-α2R(t) (9)
I ' (t)=(1-p2)α2R(t) (10)
E′2(t)=β S (t) R (t) (11)
Wherein, R (t) is reading crowd quantity, E2It (t) is accumulative amount of reading, p2Indicate that reading user generates repeat reading behavior
Rate, α2Indicate that user becomes sluggish rate, S ' (t), F ' (t), I ' (t) and E ' in read state2It (t) is respectively S
(t), F (t), I (t) and E2(t) derivative, the derivative, which is greater than 0, indicates that index value increases, and derivative indicates index number less than 0
Value reduces, and derivative, which is equal to 0, indicates that index value is constant;
The information transmission dynamics that the prediction module is constructed by the monitoring to accumulative amount of reading according to above-mentioned model construction module
The quantity and variation of the user in different states of model prediction User Status setting module setting.
9. the information transmission dynamics system according to claim 8 based on microblogging, which is characterized in that the prediction module inspection
The public sentiment index of survey further includes the second public sentiment explosive index, the second public sentiment propagation peak and the second maximum public sentiment spread index, institute
State prediction module further include:
Second public sentiment explosive index predicting unit predicts that the second public sentiment is quick-fried according to the following formula (12) by initial Susceptible population's quantity
Bristle with anger number, the second public sentiment explosive index indicates the severity of public sentiment event outburst
Wherein :=indicate to be defined as,For the second public sentiment explosive index, E20Indicate initially accumulative amount of reading, S0Expression is in
The initial value of the number of users of sensitization;
Judge that the second public sentiment explosive index is greater than 1, is also equal to 1 less than 1;
IfShow to read number of users decline, microblogging will be unable to further propagate;IfShow
Reading number of users will exponentially increase;IfShow to read number of users without change;
Second public sentiment propagation peak predicting unit predicts that the second public sentiment passes according to the following formula (13) by initial Susceptible population's quantity
Broadcast peak value Rmax, user's reading behavior increment within unit time when the second public sentiment propagation peak expression public sentiment event is broken out
Maximum value
Second time prediction unit, will read at the time of number of users increases to the second public sentiment propagation peak third setting ratio as
Second public sentiment breaks out initial time, predicts the second public sentiment outburst initial time;Number of users will be read to increase to the second carriage
Peak value moment is broken out as the second public sentiment at the time of feelings propagation peak, the second public sentiment outburst peak value moment is predicted;It will read
It reads to break out finish time as the second public sentiment at the time of number of users drops to the second public sentiment four setting ratio of propagation peak, to the
Two public sentiments outburst finish time is predicted;
Second speed predicting unit, by the unit time of the second public sentiment outburst initial time to the second public sentiment outburst peak value moment
Number of users is read as the second public sentiment and breaks out speed, realizes the prediction to the second public sentiment outburst speed;Second public sentiment is broken out
Peak value moment to the second public sentiment outburst finish time unit time reading number of users as the second public sentiment decline rate, it is real
Now to the prediction of the second public sentiment decline rate;By the unit of the second public sentiment outburst initial time to the second public sentiment outburst finish time
The reading number of users of time averagely breaks out rate as second, realizes the prediction that rate is averagely broken out to the second public sentiment;
Second maximum public sentiment spread index predicting unit constructs second most according to the following formula (14) by initial Susceptible population's quantity
Big public sentiment spread index, the described second maximum public sentiment spread index indicate that public sentiment event breaks out the maximum boundary that can reach
Number, i.e. whole event Onset read cumulative amount
Wherein, E2sFor the second maximum public sentiment spread index.
10. the information transmission dynamics system according to any claim in claim 6-9 based on microblogging, feature exist
In, further include model training module, to model construction module building information transmission dynamics model be trained, comprising:
Training set construction unit acquires the data of microblogging setting time as training set;
Parameter training unit, the training set that training set construction unit is constructed input information transmission dynamics model, using minimum two
The parameter of multiplication estimated information transmission dynamics model.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910411316.2A CN110297984B (en) | 2019-05-17 | 2019-05-17 | Microblog-based information propagation power system, construction method, device and medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910411316.2A CN110297984B (en) | 2019-05-17 | 2019-05-17 | Microblog-based information propagation power system, construction method, device and medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110297984A true CN110297984A (en) | 2019-10-01 |
CN110297984B CN110297984B (en) | 2021-10-01 |
Family
ID=68027011
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910411316.2A Active CN110297984B (en) | 2019-05-17 | 2019-05-17 | Microblog-based information propagation power system, construction method, device and medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110297984B (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110971534A (en) * | 2019-11-13 | 2020-04-07 | 哈尔滨哈工大机器人集团嘉利通科技股份有限公司 | Government affair public opinion-oriented uplink rate regulation and control method and device |
CN111460679A (en) * | 2020-04-17 | 2020-07-28 | 中国传媒大学 | Dynamics-based synchronous cross information propagation analysis method and system |
CN112348279A (en) * | 2020-11-18 | 2021-02-09 | 武汉大学 | Information propagation trend prediction method and device, electronic equipment and storage medium |
CN113157993A (en) * | 2021-02-08 | 2021-07-23 | 电子科技大学 | Network water army behavior early warning model based on time sequence graph polarization analysis |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105357200A (en) * | 2015-11-09 | 2016-02-24 | 河海大学 | Network virus transmission behavior modeling method |
CN106096075A (en) * | 2016-05-25 | 2016-11-09 | 中山大学 | A kind of message propagation model based on social networks |
CN106127590A (en) * | 2016-06-21 | 2016-11-16 | 重庆邮电大学 | A kind of information Situation Awareness based on node power of influence and propagation management and control model |
CN106649685A (en) * | 2016-12-16 | 2017-05-10 | 南京邮电大学 | SEIAR rumor spreading procedure description method with comment and forward behaviors taken into account |
US20180018709A1 (en) * | 2016-05-31 | 2018-01-18 | Ramot At Tel-Aviv University Ltd. | Information spread in social networks through scheduling seeding methods |
CN107918610A (en) * | 2016-10-09 | 2018-04-17 | 郑州大学 | A kind of microblogging propagation model towards Time Perception |
-
2019
- 2019-05-17 CN CN201910411316.2A patent/CN110297984B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105357200A (en) * | 2015-11-09 | 2016-02-24 | 河海大学 | Network virus transmission behavior modeling method |
CN106096075A (en) * | 2016-05-25 | 2016-11-09 | 中山大学 | A kind of message propagation model based on social networks |
US20180018709A1 (en) * | 2016-05-31 | 2018-01-18 | Ramot At Tel-Aviv University Ltd. | Information spread in social networks through scheduling seeding methods |
CN106127590A (en) * | 2016-06-21 | 2016-11-16 | 重庆邮电大学 | A kind of information Situation Awareness based on node power of influence and propagation management and control model |
CN107918610A (en) * | 2016-10-09 | 2018-04-17 | 郑州大学 | A kind of microblogging propagation model towards Time Perception |
CN106649685A (en) * | 2016-12-16 | 2017-05-10 | 南京邮电大学 | SEIAR rumor spreading procedure description method with comment and forward behaviors taken into account |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110971534A (en) * | 2019-11-13 | 2020-04-07 | 哈尔滨哈工大机器人集团嘉利通科技股份有限公司 | Government affair public opinion-oriented uplink rate regulation and control method and device |
CN110971534B (en) * | 2019-11-13 | 2023-05-05 | 哈尔滨哈工智慧嘉利通科技股份有限公司 | Uplink rate regulation and control method and device for government affair public opinion |
CN111460679A (en) * | 2020-04-17 | 2020-07-28 | 中国传媒大学 | Dynamics-based synchronous cross information propagation analysis method and system |
CN111460679B (en) * | 2020-04-17 | 2021-05-25 | 中国传媒大学 | Dynamics-based synchronous cross information propagation analysis method and system |
CN112348279A (en) * | 2020-11-18 | 2021-02-09 | 武汉大学 | Information propagation trend prediction method and device, electronic equipment and storage medium |
CN112348279B (en) * | 2020-11-18 | 2024-04-05 | 武汉大学 | Information propagation trend prediction method, device, electronic equipment and storage medium |
CN113157993A (en) * | 2021-02-08 | 2021-07-23 | 电子科技大学 | Network water army behavior early warning model based on time sequence graph polarization analysis |
Also Published As
Publication number | Publication date |
---|---|
CN110297984B (en) | 2021-10-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110297984A (en) | Information transmission dynamics system, construction method, device and medium based on microblogging | |
Song et al. | Information flow modeling based on diffusion rate for prediction and ranking | |
Welch et al. | Topical semantics of twitter links | |
Agarwal et al. | Blogosphere: research issues, tools, and applications | |
Zhou et al. | C-logit stochastic user equilibrium model: formulations and solution algorithm | |
Kivimäki et al. | Two betweenness centrality measures based on randomized shortest paths | |
Ponnam et al. | Movie recommender system using item based collaborative filtering technique | |
Yang et al. | Tracking influential individuals in dynamic networks | |
Aslay et al. | Competition-based networks for expert finding | |
Zeng et al. | Uncovering the information core in recommender systems | |
Nepal et al. | A trust model-based analysis of social networks | |
Gardes et al. | Estimating extreme quantiles of Weibull tail distributions | |
CN109885656B (en) | Microblog forwarding prediction method and device based on quantification heat degree | |
Yang et al. | Recommender system-based diffusion inferring for open social networks | |
Abbas et al. | Popularity and novelty dynamics in evolving networks | |
Agarwal et al. | A social identity approach to identify familiar strangers in a social network | |
Li et al. | Social network user influence dynamics prediction | |
Roy et al. | The attention automaton: Sensing collective user interests in social network communities | |
Liu et al. | A framework to compute page importance based on user behaviors | |
von Kleist et al. | Statistical analysis of the first passage path ensemble of jump processes | |
Ying et al. | Followee recommendation in asymmetrical location-based social networks | |
Zhou et al. | Group dynamics in discussing incidental topics over online social networks | |
Chiang et al. | Exploring latent browsing graph for question answering recommendation | |
Zhou et al. | Extracting news blog hot topics based on the W2T Methodology | |
Barbieri et al. | Survival factorization on diffusion networks |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |