CN103888498B - Information-pushing method, device, terminal and server - Google Patents
Information-pushing method, device, terminal and server Download PDFInfo
- Publication number
- CN103888498B CN103888498B CN201210562450.0A CN201210562450A CN103888498B CN 103888498 B CN103888498 B CN 103888498B CN 201210562450 A CN201210562450 A CN 201210562450A CN 103888498 B CN103888498 B CN 103888498B
- Authority
- CN
- China
- Prior art keywords
- interest
- probability
- scene
- product
- matrix
- 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.)
- Active
Links
Abstract
The invention discloses a kind of information-pushing method, device, terminal and server, belong to field of computer technology.The described method includes：Obtain the sequence of scenes and the corresponding sequence of interest of the sequence of scenes for being gradually switched to n-th current of scene in the default period from the 1st scene, n >=2；By the scene composition in the sequence of scenes and preset model into scene set, the interest in the sequence of interest and the model is combined into interest set, the model parameter of the model is calculated according to the scene set and the interest set；The current corresponding current interest of n-th of scene according to the sequence of scenes, the interest set and the model parameter calculation；According to the current interest to terminal pushed information.The present invention solves the problem of terminal is not modeled to the corresponding interest of window according to scene, have impact on the accuracy pushed to terminal into row information, has reached the effect for the accuracy for improving information push.
Description
Technical field
The present invention relates to field of computer technology, more particularly to a kind of information-pushing method, device, terminal and server.
Background technology
User produces user data during application program is used, such as, user browses door information website, microblogging
Or use UGC caused by player（User Generated Content, user-generated content）Or status data etc..With
The addition of function in the abundant and application program of application program species, the use that user produces during using application program
User data is also more and more, and abundant user data allows user data of the terminal by same user in a variety of application programs
Merged, and the interest of user is predicted according to the data after fusion, and then may be interested to terminal push user
Information.
In the prior art, terminal establishes a model for user in advance, the window information and use of terminal operating in the model
The interest at family is one-to-one relation, then when terminal from previous windows exchange be current window when, according to current window information
Interest corresponding with current window can be found in a model, to terminal push and the relevant information of the interest.Such as model
In to pre-set the corresponding interest of web page windows be news, then terminal detects current window when being web page windows, in a model
Interest corresponding with web page windows is searched, obtains current events, then to terminal push and the relevant information of current events.
In the implementation of the present invention, inventor has found that the prior art has at least the following problems：
When the same window of terminal is run in different scenes, the corresponding interest of the window may be different, such as, morning fortune
The corresponding interest of capable web page windows may be current events, afternoon operation the corresponding interest of web page windows may be entertainment news
Deng, and terminal is not modeled the corresponding interest of window according to scene, have impact on the accuracy to terminal pushed information.
The content of the invention
The corresponding interest of window is not modeled according to scene in order to solve terminal, have impact on to terminal pushed information
Accuracy the problem of, an embodiment of the present invention provides a kind of information-pushing method, device, terminal and server.The technology
Scheme is as follows：
On the one hand, there is provided a kind of information-pushing method, the described method includes：
Obtain sequence of scenes and the institute for being gradually switched to n-th current of scene in the default period from the 1st scene
State the corresponding sequence of interest of sequence of scenes, n >=2；
By the scene composition in the sequence of scenes and preset model into scene set, by the sequence of interest and the mould
Interest in type is combined into interest set, and the model that the model is calculated according to the scene set and the interest set is joined
Number；
N-th current of scene pair according to the sequence of scenes, the interest set and the model parameter calculation
The current interest answered；
According to the current interest to terminal pushed information.
On the other hand, there is provided a kind of information push-delivery apparatus, described device include：
First acquisition module, current n-th is gradually switched to for obtaining in the default period from the 1st scene
The sequence of scenes of scape and the corresponding sequence of interest of the sequence of scenes, n >=2；
First computing module, for the sequence of scenes and the field in preset model for obtaining first acquisition module
Scape is combined into scene set, and the sequence of interest that first acquisition module obtains is combined into the interest in the model
Interest set, the model parameter of the model is calculated according to the scene set and the interest set；
Second computing module, for the sequence of scenes, the interest set obtained according to first acquisition module
Current interest corresponding with n-th of scene current described in the model parameter calculation that first computing module calculates；
First pushing module, believes for being pushed according to the current interest that second computing module calculates to terminal
Breath.
Another further aspect, there is provided a kind of terminal, the terminal include information push-delivery apparatus as described above.
Another aspect, there is provided a kind of server, the server include information push-delivery apparatus as described above.
The beneficial effect that technical solution provided in an embodiment of the present invention is brought is：
By by the scene composition in the sequence of scenes and preset model into scene set, by the sequence of interest and institute
State the interest in model and be combined into interest set, the model of the model is calculated according to the scene set and the interest set
Parameter；N-th current of scene is corresponding according to the sequence of scenes, the interest set and the model parameter calculation
Current interest；According to the current interest to terminal pushed information, it is corresponding to window not emerging according to scene to solve terminal
The problem of interest is modeled, the problem of have impact on the accuracy to terminal pushed information, has reached raising information push accuracy
Effect.
Brief description of the drawings
To describe the technical solutions in the embodiments of the present invention more clearly, make required in being described below to embodiment
Attached drawing is briefly described, it should be apparent that, drawings in the following description are only some embodiments of the present invention, for
For those of ordinary skill in the art, without creative efforts, other can also be obtained according to these attached drawings
Attached drawing.
Fig. 1 is the method flow diagram for the information-pushing method that the embodiment of the present invention one provides；
Fig. 2 is the method flow diagram of information-pushing method provided by Embodiment 2 of the present invention；
Fig. 3 is the method flow diagram for the information-pushing method that the embodiment of the present invention three provides；
Fig. 4 is the structure diagram for the information push-delivery apparatus that the embodiment of the present invention four provides；
Fig. 5 is the structure diagram for the information push-delivery apparatus that the embodiment of the present invention five provides.
Embodiment
To make the object, technical solutions and advantages of the present invention clearer, below in conjunction with attached drawing to embodiment party of the present invention
Formula is described in further detail.
Embodiment one
Please refer to Fig.1, the method flow diagram of the information-pushing method provided it illustrates the embodiment of the present invention one, the information
Method for pushing can be applied in terminal, which can be smart television, smart mobile phone or tablet computer etc.；Alternatively,
The information-pushing method can also be applied in server.The embodiment of the present invention is applied to be said exemplified by terminal in this way
Method bright, then that the information pushes, including：
Step 102：Obtain the scene for being gradually switched to n-th current of scene in the default period from the 1st scene
Sequence and the corresponding sequence of interest of sequence of scenes, n >=2；
Wherein, scene is the operation of the terminal-pair page or application program.For example when being operated to the page, scene can
To include but not limited to：Content of pages, subject of Web site, the page pool clicked on where behavior, current visual angle produced in the short time
Domain, the link etc. of current page；When being operated to application program, scene can include but is not limited to：Application program is run
Time, application category, user's use habit, user generated content (UGC), user group distribution, the period of using terminal, end
Hold type, terminal versions number etc..
Interest is the information that user is obtained by the page or application program.For example interest can include but is not limited to：Information
Technology IT, real estate, dress ornament, personal belongings, industry articles for use, decoration, traffic, education, finance, service, retail, game, consumption,
Medical treatment, amusement etc..
Step 104：By the scene composition in sequence of scenes and preset model into scene set, by sequence of interest and model
Interest be combined into interest set, according to scene set and the model parameter of interest set computation model；
Step 106：It is corresponding current according to n-th current of scene of sequence of scenes, interest set and model parameter calculation
Interest；
Step 108：According to current interest to terminal pushed information.
Terminal is referred to user recommend with the relevant information of interest in current scene into row information push.
In conclusion information-pushing method provided in an embodiment of the present invention, by by the sequence of scenes and preset model
In scene composition into scene set, the interest in the sequence of interest and the model is combined into interest set, according to institute
State scene set and the interest set calculates the model parameter of the model；According to the sequence of scenes, the interest set
Current interest corresponding with n-th of scene current described in the model parameter calculation；Pushed away according to the current interest to terminal
Deliver letters breath, solve terminal and the corresponding interest of window is not modeled according to scene, have impact on to terminal pushed information
The problem of accuracy, reached the effect for improving information push accuracy.
Embodiment two
Please refer to Fig.2, it illustrates the method flow diagram of information-pushing method provided by Embodiment 2 of the present invention, the information
Method for pushing can be applied in terminal, which can be smart television, smart mobile phone or tablet computer etc.；Alternatively,
The information-pushing method can also be applied in server.The embodiment of the present invention is applied to be said exemplified by terminal in this way
It is bright, then the information-pushing method, including：
Step 202：Obtain the scene for being gradually switched to n-th current of scene in the default period from the 1st scene
Sequence sequence of interest corresponding with the sequence of scenes, n >=2；
Wherein, scene is the operation of the terminal-pair page or application program.For example when being operated to the page, scene can
To include but not limited to：Content of pages, subject of Web site, the page pool clicked on where behavior, current visual angle produced in the short time
Domain, the link etc. of current page；When being operated to application program, scene can include but is not limited to：Application program is run
Time, application category, user's use habit, user generated content (UGC), user group distribution, the period of using terminal, end
Hold type, terminal versions number etc..
Specifically, when the page or application program in user's operation terminal, terminal can pass through the behaviour of log recording execution
Make, then can pre-set a period, read the operation that terminal performs in daily record in real time within the period, you can obtain
Sequence of scenes.The default period voluntarily can set and adjust, and the present embodiment is not construed as limiting.
For example user has been first turned on news pages, player is then opened, news pages is then returned to, then obtains
The sequence of scenes taken is " news pages, player, news pages ".
Interest is the information that user is obtained by the page or application program.For example interest can include but is not limited to：Information
Technology IT, real estate, dress ornament, personal belongings, industry articles for use, decoration, traffic, education, finance, service, retail, game, consumption,
Medical treatment, amusement etc..
Further, when user is by terminal to server requested webpage information, server can be to the letter of terminal request
Breath is recorded, then the information recorded in terminal-pair server is read out, with reference to the log information of terminal, you can is got each
The corresponding interest of a scene, so that it is determined that sequence of interest corresponding with sequence of scenes.
For example user has browsed financial finance and economics in news pages, have selected folk song in player, returns to news pages
When browsed social news, then can be by " financial finance and economics, folk song, social news " conduct " news pages, player, news page
The corresponding sequence of interest in face ".
Step 204：By the scene composition in sequence of scenes and preset model into scene set, by sequence of interest and model
Interest be combined into interest set, according to scene set and the model parameter of interest set computation model, which includes
Probability matrix, transition probability matrix and emission matrix；
Terminal can also establish model in local to scene interest corresponding with the scene, which can include scene collection
Conjunction, interest set and model parameter, for predicting the corresponding interest of scene, or prediction later scene, so that server is according to pre-
The interest or later scene measured push terminal into row information, information is pushed more hommization and intelligence.
Specifically, according to scene set and the model parameter of interest set computation model, can include：
Pre-set probability matrix or obtain previous probability matrix；
For i-th of interest in m interest of interest set, i-th of interest is calculated successively and is transferred to j-th interest
Transition probability, obtains transition probability matrix；
For i-th of interest in interest set, the emission probability that i-th of interest produces k-th of scene is calculated successively, is obtained
To emission matrix；
Wherein, 1≤i≤m, 1≤j≤m and 1≤k≤n.
Wherein, the probability of i-th of interest is pre-set, such as, terminal sets the probability of each interest
It is equal, or the probability of each interest is set based on experience value etc., the present embodiment is not construed as limiting.Further, terminal may be used also
To be adjusted to probability matrix.
When terminal calculates transition probability matrix, i-th of interest can be transferred to the number divided by i-th of j-th of interest
A interest is taken up the total degree of transfer, is obtained i-th of interest and is transferred to j-th of interest transition probability, it is, of course, also possible to logical
Cross other modes and calculate transition probability matrix, the present embodiment is not construed as limiting.
When terminal calculates emission matrix, i-th of interest can be produced to the number divided by i-th of interest of k-th of scene
The total degree of scene is produced, the emission probability that i-th of interest produces k-th of scene is obtained, it is, of course, also possible to by other means
Emission matrix is calculated, the present embodiment is not construed as limiting.
In addition, the present embodiment does not limit the computation sequence of probability matrix, probability transfer matrix and emission matrix.
Step 206：It is corresponding current according to n-th current of scene of sequence of scenes, interest set and model parameter calculation
Interest；
Specifically, it is corresponding current emerging according to n-th current of scene of sequence of scenes, interest set and model parameter calculation
Interest, can include：
It is when calculating the probability of each interest in the 1st scene, i-th of interest is initial in probability matrix
Probability and i-th of interest produce the product of the emission probability of the 1st scene as i-th of interest at the 1st in emission matrix
Probability in scene；
When calculating the probability of each interest in q-th of scene, j-th of interest is first calculated successively in previous scene
Probability and j-th of interest be transferred in transition probability matrix i-th of interest transition probability the first product, determine first
The first product of maximum in product, then calculate maximum first sum of products, i-th of interest and q-th scene is produced in emission matrix
Second product of emission probability, using probability of second product as i-th of interest in q-th of scene, 2≤q≤n-1；
When calculating the probability of each interest in n-th of scene, j-th of interest is first calculated successively in previous scene
Probability and j-th of interest be transferred in transition probability matrix i-th of interest transition probability the first product, determine first
The first product of maximum in product, then calculate maximum first sum of products, i-th of interest and n-th scene is produced in emission matrix
Second product of emission probability, probability of second product as i-th of interest in n-th of scene is determined in the second product
The second product of maximum, using the corresponding interest of maximum second product as current interest.
Such as, it is assumed that interest collection is combined into { a, b, c }, and scene collection is combined into { g, h }, and the sequence of scenes currently obtained is ghgh,
And probability matrixTransition probability matrixEmission matrixThen
The calculating process of the corresponding current interest of n-th of scene is as follows：
Intermediate quantity δ is incorporated hereinn(j), wherein δn(j) it is the probability of interest j in n-th of scene；
The probability calculation of each interest in 1st scene g：
δ1(a)=π1·b1(g)=1·0.7=0.7；
The probability calculation of each interest in 2nd scene h：
δ2(a)=δ1(a)·a11·b1(h)=0.7·0.4·0.3=0.084；
δ2(b)=δ1(a)·a12·b2(h)=0.7·0.6·0.6=0.252；
The probability calculation of each interest in 3rd scene g：
δ3(a)=δ2(a)·a11·b1(g)=0.084·0.4·0.7=0.02352；
δ3(b)=max{δ2(a)·a12,δ2(b)·a22}·b2(g)=max{0.084·0.6,0.0252·0.8}·0.4
=0.08064；
δ3(c)=δ2(b)·a23·b3(g)=0.252·0.2·0.8=0.04032；
The probability calculation of each interest in 4th scene h：
δ4(a)=δ3(a)·a11·b1(h)=0.02352·0.4·0.3=0.0028224；
δ4(b)=max{δ3(a)·a12,δ3(b)·a22}·b2(h)=max{0.014112,0.064512}·0.6=
0.0387072；
δ4(c)=max{δ3(b)·a23,δ3(c)·a32}·b3(h)=max{0.016128,0.04032}·0.2=
0.008064；
According to result of calculation, the maximum probability of b in the 4th scene h, then terminal determine that b corresponds to for the 4th scene h
Current interest.
Further, due to terminal obtain scene and interest can not possibly be infinitely more, for physical presence without
The interest transfer got, when calculating transition probability matrix according to above-mentioned computational methods, the transfer between the interest being calculated
Probability is zero, such as, the transition probability that c is transferred to a in A matrixes is that the transition probability that 0, c is transferred to b is 0 etc..Work as transition probability
When being zero, the probability of the interest that is calculated in the scene is necessarily zero, and one calculates path and just interrupts, and have impact on model
Accuracy., can be when calculating transition probability matrix using a smoothing algorithm is added, i.e., by interest set in order to avoid such case
Transfer number between middle any two interest adds one.
For example interest collection is combined into { a, b, c }, then when calculating the transition probability of a, it is assumed that the number that a is transferred to a is 6 times, a
The number for going to b is 3 times, and the number that a is transferred to c is 0 time, then the probability that a being calculated is transferred to c is 0/（6+3）=0.Adopt
After a smoothing algorithm is added, can obtain a be transferred to a number be 7 times, a go to b number be 4 times, a is transferred to the number of c
For 1 time, the probability that a being calculated at this time is transferred to c is 1/（7+4+1）=0.08333, avoid and calculate asking for path disruption
Topic.
Similarly, when calculating emission matrix, an above-mentioned plus smoothing algorithm can also be used, is not repeated herein.
Step 208：According to current interest to terminal pushed information.
It is corresponding with the interest to server acquisition according to the current interest when terminal gets the current interest calculated
Information, and the information is shown in the terminal.
In conclusion information-pushing method provided in an embodiment of the present invention, by by the sequence of scenes and preset model
In scene composition into scene set, the interest in the sequence of interest and the model is combined into interest set, according to institute
State scene set and the interest set calculates the model parameter of the model；According to the sequence of scenes, the interest set
Current interest corresponding with n-th of scene current described in the model parameter calculation；Pushed away according to the current interest to terminal
Deliver letters breath, solve terminal and the corresponding interest of window is not modeled according to scene, have impact on to terminal pushed information
The problem of accuracy, reached the effect for improving information push accuracy.It is in addition, emerging by m for the interest set
I-th of interest in interest, calculates the transition probability that i-th of the interest is transferred to j-th of interest successively, and it is general to obtain the transfer
Rate matrix；For i-th of interest in the interest set, the transmitting that i-th of the interest produces k-th of scene is calculated successively
Probability, obtains the emission matrix, and the relation of interest in model and scene cannot be calculated by solving, and have impact on to terminal
The problem of accuracy of pushed information, reached the effect for improving information push accuracy.
Embodiment three
Please refer to Fig.3, the method flow diagram of the information-pushing method provided it illustrates the embodiment of the present invention three, the information
Method for pushing can be applied in terminal, which can be smart television, smart mobile phone or tablet computer etc.；Alternatively,
The information-pushing method can also be applied in server.The embodiment of the present invention is applied to carry out exemplified by server in this way
Illustrate, then the information-pushing method, including：
Step 302：Obtain the scene for being gradually switched to n-th current of scene in the default period from the 1st scene
Sequence and the corresponding sequence of interest of sequence of scenes, n >=2；
Wherein, scene is the operation of the terminal-pair page or application program.For example when being operated to the page, scene can
To include but not limited to：Content of pages, subject of Web site, the page pool clicked on where behavior, current visual angle produced in the short time
Domain, the link etc. of current page；When being operated to application program, scene can include but is not limited to：Application program is run
Time, application category, user's use habit, user generated content (UGC), user group distribution, the period of using terminal, end
Hold type, terminal versions number etc..
Wherein, the process of terminal acquisition scene refers to the description in step 202, does not repeat herein.Further, terminal can
Server is sent to so that sequence of scenes will be got, so that server records scene information.
Interest is the information that user is obtained by the page or application program.For example interest can include but is not limited to：Information
Technology IT, real estate, dress ornament, personal belongings, industry articles for use, decoration, traffic, education, finance, service, retail, game, consumption,
Medical treatment, amusement etc..
Further, when user is by terminal to server requested webpage information, server can be to the letter of terminal request
Breath is recorded, then server is read out the information of record, the scene information sent with reference to terminal, you can get each
The corresponding interest of scene, so that it is determined that sequence of interest corresponding with sequence of scenes.
Step 304：By the scene composition in sequence of scenes and preset model into scene set, by sequence of interest and model
Interest be combined into interest set, according to scene set and the model parameter of interest set computation model, which includes
Probability matrix, transition probability matrix and emission matrix；
Server can also establish model to scene interest corresponding with the scene, the model can include scene set,
Interest set and model parameter, for predicting the corresponding interest of scene, or prediction later scene, so that server is according to predicting
Interest or later scene to terminal into row information push, make information push more hommization and intelligence.
Specifically, according to scene set and the model parameter of interest set computation model, can include：
Pre-set probability matrix or obtain previous probability matrix；
For i-th of interest in m interest of interest set, i-th of interest is calculated successively and is transferred to j-th interest
Transition probability, obtains transition probability matrix；
For i-th of interest in interest set, the emission probability that i-th of interest produces k-th of scene is calculated successively, is obtained
To emission matrix；
Wherein, 1≤i≤m, 1≤j≤m and 1≤k≤n.
Wherein, server refers to the calculating process of model parameter the description in step 204, does not repeat herein.
Step 306：At least one forecasting sequence is obtained, each forecasting sequence adds a prediction scene including sequence of scenes,
It is the scene in scene set to predict scene, and the prediction scene in each forecasting sequence is different；
Such as sequence of scenes ghg, and scene collection is combined into { g, h }, then forecasting sequence can be ghgg, or ghgh.
Step 308：For each forecasting sequence, it is sequenced in advance according to forecasting sequence, interest set and model parameter calculation
The probability of row；
The method that server provides the probability of two kinds of calculating forecasting sequences, specifically, according to forecasting sequence, interest set
With the probability of model parameter calculation forecasting sequence, can include：
It is emerging according to the corresponding prediction of prediction scene in forecasting sequence, interest set and model parameter forward calculation forecasting sequence
The probability of interest, will predict probability of the probability as forecasting sequence of interest；Alternatively,
According to the 1st corresponding interest of scene in forecasting sequence, interest set and model parameter backwards calculation forecasting sequence
Probability, the probability using the probability of interest as forecasting sequence.
Further, according to prediction scene pair in forecasting sequence, interest set and model parameter forward calculation forecasting sequence
The probability for the prediction interest answered, including：
It is when calculating the probability of each interest in the 1st scene, i-th of interest is initial in probability matrix
Probability and i-th of interest produce the product of the emission probability of the 1st scene as i-th of interest at the 1st in emission matrix
Probability in scene；
When calculating the probability of each interest in q-th of scene, j-th of interest is first calculated successively in previous scene
Probability and j-th of interest be transferred in transition probability matrix i-th of interest transition probability the 3rd product, calculate the 3rd
The 3rd sum of products after product addition, then calculate the 3rd sum of products and i-th of interest produces q-th scene in emission matrix
4th product of emission probability, using probability of the 4th product as i-th of interest in q-th of scene, 2≤q≤n-1；
When calculating the probability of each interest in n-th of scene, j-th of interest is first calculated successively in previous scene
Probability and j-th of interest be transferred in transition probability matrix i-th of interest transition probability the 3rd product, calculate the 3rd
The 3rd sum of products after product addition, then calculate the 3rd sum of products and i-th of interest produces n-th scene in emission matrix
4th product of emission probability, probability of the 4th product as i-th of interest in n scene is determined in the 4th product
Maximum 4th product, using maximum 4th product as the probability that the corresponding prediction interest of scene is predicted in forecasting sequence.
It is alternatively, corresponding according to the 1st scene in forecasting sequence, interest set and model parameter backwards calculation forecasting sequence
The probability of interest, can include：
It is when calculating the probability of each interest in n-th of scene, i-th of interest is initial in probability matrix
Probability and i-th of interest produce the product of the emission probability of n-th of scene as i-th of interest at n-th in emission matrix
Probability in scene；
When calculating the probability of each interest in q-th of scene, j-th of interest is first calculated successively in the latter scene
Probability and j-th of interest be transferred in transition probability matrix i-th of interest transition probability the 5th product, calculate the 5th
The 5th sum of products after product addition, then calculate the 5th sum of products and i-th of interest produces q-th scene in emission matrix
6th product of emission probability, using probability of the 6th product as i-th of interest in q-th of scene, 2≤q≤n-l；
When calculating the probability of each interest in the 1st scene, j-th of interest is first calculated successively in the latter scene
Probability and j-th of interest be transferred in transition probability matrix i-th of interest transition probability the 5th product, calculate the 5th
The 5th sum of products after product addition, then calculate the 5th sum of products and i-th of interest and the 1st scene is produced in emission matrix
6th product of emission probability, probability of the 6th product as i-th of interest in the 1st scene is determined in the 6th product
The 6th product of maximum, the probability using maximum 6th product as the 1st corresponding interest of scene in forecasting sequence.
Such as to be corresponded to according to prediction scene in forecasting sequence, interest set and model parameter forward calculation forecasting sequence
Prediction interest probability exemplified by illustrate, it is assumed that interest collection is combined into { a, b, c }, and scene collection is combined into { g, h }, current prediction
Sequence of scenes is ghgh, and probability matrixTransition probability matrixEmission matrixThen the calculating process of the corresponding current interest of n-th of scene is as follows：
Intermediate quantity α is incorporated hereinn(j), wherein αn(j) it is the probability of interest j in n-th of scene；
The probability calculation of each interest in 1st scene g：
α1(a)=π1·b1(g)=1O.7=O.7,
The probability calculation of each interest in 2nd scene h：
α2(a)=α1(a)·a11·b1(h)=O.7O.4O.3=0.084；
α2(b)=α1(a)·a12.b2(h)=O.7O.6O.6=0.252；
The probability calculation of each interest in 3rd scene g：
α3(a)=α2(a)·a11·b1(g)=0.084O.4O.7=0.02352；
α3(b)=[α2(a)·a12+α2(b)·a22]·b2(g)=[0.0840.6+0.02520.8] is o.4=
0.1008；
α3(c)=α2(b)·a23·b3(g)=0.252O.2O.8=0.04032；
Predict the probability calculation of each interest in scene (the 4th scene) h：
α4(a)=α3(a)·a11·b1(h)=0.02352O.4O.3=0.0028224；
α4(b)=[α3(a)·a12+α3(b)·a22]·b2(h)=[0.023520.6+0.10080.8] 0.6=
0.0568512；
α4(c)=[α3(b)·a23+α3(c)·a32J·b3(h)=[0.10080.2+0.040321] 0.2=
0.012096；
According to result of calculation, the probability of forecasting sequence ghgh is 0.012096, and server can be calculated further
The probability of ghgg, then performs step 310.
Further, due to terminal obtain scene and interest can not possibly be infinitely more, for physical presence without
The interest transfer got, when calculating transition probability matrix according to above-mentioned computational methods, the transfer between the interest being calculated
Probability is zero, such as, the transition probability that c is transferred to a in A matrixes is that the transition probability that 0, c is transferred to b is 0 etc..Work as transition probability
When being zero, the probability of the interest that is calculated in the scene is necessarily zero, and one calculates path and just interrupts, and have impact on model
Accuracy., can be when calculating transition probability matrix using a smoothing algorithm is added, i.e., by interest set in order to avoid such case
Transfer number between middle any two interest adds one.Similarly, when calculating emission matrix, can also use above-mentioned plus one smooth
Algorithm, specific calculation process refer to the description in step 206, do not repeat herein.
Step 310：Determine the corresponding forecasting sequence of maximum probability in probability, and the pre- of forecasting sequence is corresponded in maximum probability
Survey in scene to terminal pushed information.
Server according to the corresponding forecasting sequence of determine the probability maximum probability of the probability and ghgg of the ghgh calculated, if
The corresponding forecasting sequence of maximum probability is ghgh, then server determines that prediction scene is h；If the corresponding forecasting sequence of maximum probability
Ghgg, then server determines that prediction scene is g, and to terminal pushed information in the prediction scene.
In conclusion information-pushing method provided in an embodiment of the present invention, by by the sequence of scenes and preset model
In scene composition into scene set, the interest in the sequence of interest and the model is combined into interest set, according to institute
State scene set and the interest set calculates the model parameter of the model；According to the sequence of scenes, the interest set
Current interest corresponding with n-th of scene current described in the model parameter calculation；Pushed away according to the current interest to terminal
Deliver letters breath, solve terminal and the corresponding interest of window is not modeled according to scene, have impact on to terminal pushed information
The problem of accuracy, reached the effect for improving information push accuracy.It is in addition, emerging by m for the interest set
I-th of interest in interest, calculates the transition probability that i-th of the interest is transferred to j-th of interest successively, and it is general to obtain the transfer
Rate matrix；For i-th of interest in the interest set, the transmitting that i-th of the interest produces k-th of scene is calculated successively
Probability, obtains the emission matrix, and the relation of interest in model and scene cannot be calculated by solving, and have impact on to terminal
The problem of accuracy of pushed information, reached the effect for improving information push accuracy.
Example IV
Please refer to Fig.4, the structural framing figure of the information push-delivery apparatus provided it illustrates the embodiment of the present invention four, the information
Pusher can be applied in terminal, which can be smart television, smart mobile phone or tablet computer etc.；Alternatively,
The information-pushing method can also be applied in server.The information push-delivery apparatus, including：
First acquisition module 410, current n-th is gradually switched to for obtaining in the default period from the 1st scene
The corresponding sequence of interest of sequence of scenes and sequence of scenes of a scene, n >=2；
First computing module 420, for the scene in the sequence of scenes and preset model that obtain the first acquisition module 410
Scene set is combined into, the interest in sequence of interest and model that the first acquisition module 410 is obtained is combined into interest set, root
According to scene set and the model parameter of interest set computation model；
Second computing module 430, for sequence of scenes, the interest set and first obtained according to the first acquisition module 410
The corresponding current interest of n-th of scene of the model parameter calculation that computing module 420 calculates currently；
First pushing module 440, for according to the current interest that the second computing module 430 calculates to terminal pushed information.
In conclusion information push-delivery apparatus provided in an embodiment of the present invention, by by the sequence of scenes and preset model
In scene composition into scene set, the interest in the sequence of interest and the model is combined into interest set, according to institute
State scene set and the interest set calculates the model parameter of the model；According to the sequence of scenes, the interest set
Current interest corresponding with n-th of scene current described in the model parameter calculation；Pushed away according to the current interest to terminal
Deliver letters breath, solve terminal and the corresponding interest of window is not modeled according to scene, have impact on to terminal pushed information
The problem of accuracy, reached the effect for improving information push accuracy.
Embodiment five
Fig. 5 is refer to, the structural framing figure of the information push-delivery apparatus provided it illustrates the embodiment of the present invention five, the information
Method for pushing can be applied in terminal, which can be smart television, smart mobile phone or tablet computer etc.；Alternatively,
The information-pushing method can also be applied in server.The information push-delivery apparatus, including：First acquisition module 410, first is counted
Calculate module 420, the second computing module 430 and the first pushing module 440.
First acquisition module 410, current n-th is gradually switched to for obtaining in the default period from the 1st scene
The corresponding sequence of interest of sequence of scenes and sequence of scenes of a scene, n >=2；
First computing module 420, for the scene in the sequence of scenes and preset model that obtain the first acquisition module 410
Scene set is combined into, the interest in sequence of interest and model that the first acquisition module 410 is obtained is combined into interest set, root
According to scene set and the model parameter of interest set computation model；
Second computing module 430, for sequence of scenes, the interest set and first obtained according to the first acquisition module 410
The corresponding current interest of n-th of scene of the model parameter calculation that computing module 420 calculates currently；
First pushing module 440, for according to the current interest that the second computing module 430 calculates to terminal pushed information.
Further, model parameter includes probability matrix, transition probability matrix and emission matrix, the first computing module
420 can include：
Acquiring unit 510, for pre-setting probability matrix or obtaining previous probability matrix；
First computing unit 520, for i-th of interest in the m interest for interest set, calculates i-th successively
Interest is transferred to the transition probability of j-th of interest, obtains transition probability matrix；
Second computing unit 530, the is produced for for i-th of interest in interest set, calculating i-th of interest successively
The emission probability of k scene, obtains emission matrix；
Wherein, 1≤i≤m, 1≤j≤m and 1≤k≤n.
Further, the second computing module 430 can include：
3rd computing unit 610, for when calculating the probability of each interest in the 1st scene, by acquiring unit 510
I-th of the interest that probability and second computing unit 530 of i-th of the interest obtained in probability matrix calculate is being sent out
Probability of the product for the emission probability for producing the 1st scene as i-th of interest in the 1st scene is penetrated in matrix；
4th computing unit 620, for when calculating the probability of each interest in q-th of scene, first calculates the successively
J-th of probability and first computing unit 520 calculating of j-th of the interest that three computing units 610 calculate in previous scene
Interest is transferred to the first product of the transition probability of i-th of interest in transition probability matrix, determines the maximum in the first product
First product, then calculate i-th of the interest that maximum first the second computing unit of sum of products 530 calculates and the is produced in emission matrix
Second product of the emission probability of q scene, using probability of second product as i-th of interest in q-th of scene, 2≤q≤
n-1；
5th computing unit 630, for when calculating the probability of each interest in n-th of scene, first calculates the successively
J-th of probability and first computing unit 520 calculating of j-th of the interest that four computing units 620 calculate in previous scene
Interest is transferred to the first product of the transition probability of i-th of interest in transition probability matrix, determines the maximum in the first product
First product, then calculate i-th of the interest that maximum first the second computing unit of sum of products 530 calculates and the is produced in emission matrix
Second product of the emission probability of n scene, using probability of second product as i-th of interest in n-th of scene, determines
The second product of maximum in two products, using the corresponding interest of maximum second product as current interest.
Further, which can also include：
Second acquisition module 450, for obtaining at least one forecasting sequence, each forecasting sequence adds one including sequence of scenes
A prediction scene, prediction scene is the scene in scene set, and the prediction scene in each forecasting sequence is different；
3rd computing module 460, for each forecasting sequence obtained for the second acquisition module 450, according to prediction
The probability of sequence, interest set and model parameter calculation forecasting sequence；
Second pushing module 470, for determine the 3rd computing module 460 calculate probability in the corresponding prediction of maximum probability
Sequence, and to terminal pushed information in the prediction scene that maximum probability corresponds to forecasting sequence.
Further, the 3rd computing module 460 can include：
6th computing unit 710, for forecasting sequence, interest set and the model obtained according to the second acquisition module 450
The probability of the corresponding prediction interest of scene is predicted in parameter forward calculation forecasting sequence, the probability of interest will be predicted as pre- sequencing
The probability of row；Alternatively,
7th computing unit 720, for forecasting sequence, interest set and the model obtained according to the second acquisition module 450
The probability of the 1st corresponding interest of scene in parameter backwards calculation forecasting sequence, using the probability of interest as the general of forecasting sequence
Rate.
Further, the 6th computing unit 710, for when calculating the probability of each interest in the 1st scene, will obtain
Take i-th that probability and second computing unit 530 of i-th of the interest of the acquisition of unit 510 in probability matrix calculate
The product that a interest produces the emission probability of the 1st scene in emission matrix is general in the 1st scene as i-th of interest
Rate；
When calculating the probability of each interest in q-th of scene, first calculate that the 6th computing unit 710 calculates successively the
J-th of the interest that probability and first computing unit 520 of the j interest in previous scene calculate is in transition probability matrix transfer
The 3rd product of the transition probability of i-th of interest is moved on to, calculates the 3rd sum of products after the 3rd product addition, then calculate the 3rd and multiply
Product and i-th of the interest calculated with the second computing unit 530 produce the 4th of the emission probability of q-th of scene in emission matrix
Product, using probability of the 4th product as i-th of interest in q-th of scene, 2≤q≤n-1；
When calculating the probability of each interest in n-th of scene, first calculate that the 6th computing unit 710 calculates successively the
J-th of the interest that probability and first computing unit 520 of the j interest in previous scene calculate is in transition probability matrix transfer
The 3rd product of the transition probability of i-th of interest is moved on to, calculates the 3rd sum of products after the 3rd product addition, then calculate the 3rd and multiply
Product and i-th of the interest calculated with the second computing unit 530 produce the 4th of the emission probability of n-th of scene in emission matrix
Product, using probability of the 4th product as i-th of interest in n scene, determines the 4th product of maximum in the 4th product, will
Maximum 4th product is as the probability that the corresponding prediction interest of scene is predicted in forecasting sequence.
Further, the 7th computing unit 720, for when calculating the probability of each interest in n-th of scene, will obtain
Take i-th that probability and second computing unit 530 of i-th of the interest of the acquisition of unit 510 in probability matrix calculate
The product that a interest produces the emission probability of n-th of scene in emission matrix is general in n-th of scene as i-th of interest
Rate；
When calculating the probability of each interest in q-th of scene, first calculate that the 7th computing unit 720 calculates successively the
J-th of the interest that probability and first computing unit 520 of the j interest in the latter scene calculate is in transition probability matrix transfer
The 5th product of the transition probability of i-th of interest is moved on to, calculates the 5th sum of products after the 5th product addition, then calculate the 5th and multiply
Product and i-th of the interest calculated with the second computing unit 530 produce the 6th of the emission probability of q-th of scene in emission matrix
Product, using probability of the 6th product as i-th of interest in q-th of scene, 2≤q≤n-1；
When calculating the probability of each interest in the 1st scene, first calculate that the 7th computing unit 720 calculates successively the
J-th of the interest that probability and first computing unit 520 of the j interest in the latter scene calculate is in transition probability matrix transfer
The 5th product of the transition probability of i-th of interest is moved on to, calculates the 5th sum of products after the 5th product addition, then calculate the 5th and multiply
Product and i-th of the interest calculated with the second computing unit 530 produce the 6th of the emission probability of the 1st scene in emission matrix
Product, using probability of the 6th product as i-th of interest in the 1st scene, determines the 6th product of maximum in the 6th product,
Probability using maximum 6th product as the 1st corresponding interest of scene in forecasting sequence.
In conclusion information push-delivery apparatus provided in an embodiment of the present invention, by by the sequence of scenes and preset model
In scene composition into scene set, the interest in the sequence of interest and the model is combined into interest set, according to institute
State scene set and the interest set calculates the model parameter of the model；According to the sequence of scenes, the interest set
Current interest corresponding with n-th of scene current described in the model parameter calculation；Pushed away according to the current interest to terminal
Deliver letters breath, solve terminal and the corresponding interest of window is not modeled according to scene, have impact on to terminal pushed information
The problem of accuracy, reached the effect for improving information push accuracy.It is in addition, emerging by m for the interest set
I-th of interest in interest, calculates the transition probability that i-th of the interest is transferred to j-th of interest successively, and it is general to obtain the transfer
Rate matrix；For i-th of interest in the interest set, the transmitting that i-th of the interest produces k-th of scene is calculated successively
Probability, obtains the emission matrix, and the relation of interest in model and scene cannot be calculated by solving, and have impact on to terminal
The problem of accuracy of pushed information, reached the effect for improving information push accuracy.
It should be noted that：The information push-delivery apparatus that above-described embodiment provides into row information when pushing, only with above-mentioned each
The division progress of function module, can be as needed and by above-mentioned function distribution by different work(for example, in practical application
Energy module is completed, i.e., the internal structure of information push-delivery apparatus is divided into different function modules, described above complete to complete
Portion or partial function.In addition, the information push-delivery apparatus that above-described embodiment provides belongs to same with information-pushing method embodiment
Design, its specific implementation process refer to embodiment of the method, and which is not described herein again.
The embodiments of the present invention are for illustration only, do not represent the quality of embodiment.
One of ordinary skill in the art will appreciate that hardware can be passed through by realizing all or part of step of above-described embodiment
To complete, relevant hardware can also be instructed to complete by program, the program can be stored in a kind of computer-readable
In storage medium, storage medium mentioned above can be read-only storage, disk or CD etc..
The foregoing is merely presently preferred embodiments of the present invention, is not intended to limit the invention, it is all the present invention spirit and
Within principle, any modification, equivalent replacement, improvement and so on, should all be included in the protection scope of the present invention.
Claims (14)
- A kind of 1. information-pushing method, it is characterised in that the described method includes：Obtain in the default period and be gradually switched to the sequence of scenes of n-th current of scene and the field from the 1st scene The corresponding sequence of interest of scape sequence, n >=2；By the scene composition in the sequence of scenes and preset model into scene set, by the sequence of interest and the model Interest be combined into interest set, the model parameter of the model, institute are calculated according to the scene set and the interest set Stating model parameter includes probability matrix, transition probability matrix and emission matrix；N-th current of scene is corresponding according to the sequence of scenes, the interest set and the model parameter calculation Current interest；According to the current interest to terminal pushed information；Wherein, the model parameter that the model is calculated according to the scene set and the interest set, including：Pre-set probability matrix or obtain previous probability matrix；For i-th of interest in m interest of the interest set, calculate successively i-th of the interest be transferred to j-th it is emerging The transition probability of interest, obtains the transition probability matrix, 1≤i≤m, 1≤j≤m；For i-th of interest in the interest set, the transmitting for calculating i-th of the interest generation, k-th of scene successively is general Rate, obtains the emission matrix, 1≤k≤n.
- 2. information-pushing method according to claim 1, it is characterised in that it is described according to the sequence of scenes, it is described emerging Interest set current interest corresponding with n-th of scene current described in the model parameter calculation, including：It is when calculating the probability of each interest in the 1st scene, i-th of interest is initial in the probability matrix Probability and i-th of the interest are produced in the emission matrix described in the product conduct of the emission probability of the 1st scene Probability of i-th of the interest in the 1st scene；When calculating the probability of each interest in q-th of scene, it is general in previous scene that j-th of interest is first calculated successively Rate and j-th of the interest are transferred to the first product of the transition probability of i-th of interest in the transition probability matrix, determine The first product of maximum in first product, then i-th of interest described in maximum first sum of products is calculated in the transmitting The second product of the emission probability of q-th of scene is produced in matrix, is existed second product as i-th of the interest Probability in q-th of scene, 2≤q≤n-1；When calculating the probability of each interest in n-th of scene, it is general in previous scene that j-th of interest is first calculated successively Rate and j-th of the interest are transferred to the first product of the transition probability of i-th of interest in the transition probability matrix, determine The first product of maximum in first product, then i-th of interest described in maximum first sum of products is calculated in the transmitting The second product of the emission probability of n-th of scene is produced in matrix, is existed second product as i-th of the interest Probability in n-th of scene, determines the second product of maximum in second product, and maximum second product is corresponded to Interest as the current interest.
- 3. information-pushing method according to claim 1, it is characterised in that the method, further includes：At least one forecasting sequence is obtained, each forecasting sequence adds a prediction scene, the prediction including the sequence of scenes Scene is the scene in the scene set, and the prediction scene in each forecasting sequence is different；It is pre- according to the forecasting sequence, the interest set and the model parameter calculation for each forecasting sequence The probability of row is sequenced；Determine the corresponding forecasting sequence of maximum probability in the probability, and the prediction field of forecasting sequence is corresponded in the maximum probability Terminal pushed information described in Jing Zhongxiang.
- 4. information-pushing method according to claim 3, it is characterised in that it is described according to the forecasting sequence, it is described emerging The probability of forecasting sequence described in interest set and the model parameter calculation, including：Scene is predicted according in forecasting sequence described in the forecasting sequence, the interest set and the model parameter forward calculation The probability of corresponding prediction interest, the probability using the probability of the prediction interest as the forecasting sequence；Alternatively,According to the 1st field in forecasting sequence described in the forecasting sequence, the interest set and the model parameter backwards calculation The probability of the corresponding interest of scape, the probability using the probability of the interest as the forecasting sequence.
- 5. information-pushing method according to claim 4, it is characterised in that it is described according to the forecasting sequence, it is described emerging The probability of interest set prediction interest corresponding with scene is predicted in forecasting sequence described in the model parameter forward calculation, including：It is when calculating the probability of each interest in the 1st scene, i-th of interest is initial in the probability matrix Probability and i-th of the interest are produced in the emission matrix described in the product conduct of the emission probability of the 1st scene Probability of i-th of the interest in the 1st scene；When calculating the probability of each interest in q-th of scene, it is general in previous scene that j-th of interest is first calculated successively Rate and j-th of the interest are transferred to the 3rd product of the transition probability of i-th of interest in the transition probability matrix, calculate The 3rd sum of products after 3rd product addition, then the 3rd sum of products and i-th of the interest are calculated in the transmitting The 4th product of the emission probability of q-th of scene is produced in matrix, the 4th product is existed as i-th of the interest Probability in q-th of scene, 2≤q≤n-1；When calculating the probability of each interest in n-th of scene, it is general in previous scene that j-th of interest is first calculated successively Rate and j-th of the interest are transferred to the 3rd product of the transition probability of i-th of interest in the transition probability matrix, calculate The 3rd sum of products after 3rd product addition, then the 3rd sum of products and i-th of the interest are calculated in the transmitting The 4th product of the emission probability of n-th of scene is produced in matrix, the 4th product is existed as i-th of the interest Probability in the n scene, determines the 4th product of maximum in the 4th product, using maximum 4th product as institute State the probability that the corresponding prediction interest of scene is predicted in forecasting sequence.
- 6. information-pushing method according to claim 4, it is characterised in that it is described according to the forecasting sequence, it is described emerging Interest gathers the probability of interest corresponding with the 1st scene in forecasting sequence described in the model parameter backwards calculation, including：It is when calculating the probability of each interest in n-th of scene, i-th of interest is initial in the probability matrix Probability and i-th of the interest are produced in the emission matrix described in the product conduct of the emission probability of n-th of scene Probability of i-th of the interest in n-th of scene；When calculating the probability of each interest in q-th of scene, it is general in the latter scene that j-th of interest is first calculated successively Rate and j-th of the interest are transferred to the 5th product of the transition probability of i-th of interest in the transition probability matrix, calculate The 5th sum of products after 5th product addition, then the 5th sum of products and i-th of the interest are calculated in the transmitting The 6th product of the emission probability of q-th of scene is produced in matrix, the 6th product is existed as i-th of the interest Probability in q-th of scene, 2≤q≤n-1；When calculating the probability of each interest in the 1st scene, it is general in the latter scene that j-th of interest is first calculated successively Rate and j-th of the interest are transferred to the 5th product of the transition probability of i-th of interest in the transition probability matrix, calculate The 5th sum of products after 5th product addition, then the 5th sum of products and i-th of the interest are calculated in the transmitting The 6th product of the emission probability of the 1st scene is produced in matrix, the 6th product is existed as i-th of the interest Probability in 1st scene, determines the 6th product of maximum in the 6th product, using maximum 6th product as The probability of the 1st corresponding interest of scene in the forecasting sequence.
- 7. a kind of information push-delivery apparatus, it is characterised in that described device includes：First acquisition module, n-th current of scene is gradually switched to for obtaining in the default period from the 1st scene Sequence of scenes and the corresponding sequence of interest of the sequence of scenes, n >=2；First computing module, for the sequence of scenes and the scene group in preset model for obtaining first acquisition module Scene set is synthesized, the sequence of interest that first acquisition module obtains and the interest in the model are combined into interest Set, the model parameter of the model is calculated according to the scene set and the interest set, and the model parameter is included just Beginning probability matrix, transition probability matrix and emission matrix；Second computing module, for the sequence of scenes, the interest set and institute obtained according to first acquisition module State the corresponding current interest of n-th of scene current described in the model parameter calculation of the first computing module calculating；First pushing module, for according to the current interest that second computing module calculates to terminal pushed information；Wherein, first computing module includes：Acquiring unit, for pre-setting probability matrix or obtaining previous probability matrix；First computing unit, for i-th of interest in the m interest for the interest set, calculates described i-th successively Interest is transferred to the transition probability of j-th of interest, obtains the transition probability matrix, 1≤i≤m, 1≤j≤m；Second computing unit, for for i-th of interest in the interest set, calculating i-th of the interest successively and producing The emission probability of k-th of scene, obtains the emission matrix, 1≤k≤n.
- 8. information push-delivery apparatus according to claim 7, it is characterised in that second computing module includes：3rd computing unit, for when calculating the probability of each interest in the 1st scene, the acquiring unit to be obtained I-th of the interest that probability and second computing unit of i-th of the interest in the probability matrix calculate is in institute The product of the emission probability that the 1st scene is produced in emission matrix is stated as i-th of the interest in the 1st scene In probability；4th computing unit, for when calculating the probability of each interest in q-th of scene, first calculating the 3rd meter successively J-th of the interest that probability and first computing unit of j-th of the interest of unit calculating in previous scene calculate is calculated to exist The first product of the transition probability of i-th of interest is transferred in the transition probability matrix, is determined in first product most Big first product, then i-th of interest of the second computing unit calculating described in maximum first sum of products is calculated in the transmitting The second product of the emission probability of q-th of scene is produced in matrix, is existed second product as i-th of the interest Probability in q-th of scene, 2≤q≤n-1；5th computing unit, for when calculating the probability of each interest in n-th of scene, first calculating the 4th meter successively J-th of the interest that probability and first computing unit of j-th of the interest of unit calculating in previous scene calculate is calculated to exist The first product of the transition probability of i-th of interest is transferred in the transition probability matrix, is determined in first product most Big first product, then i-th of interest of the second computing unit calculating described in maximum first sum of products is calculated in the transmitting The second product of the emission probability of n-th of scene is produced in matrix, is existed second product as i-th of the interest Probability in n-th of scene, determines the second product of maximum in second product, and maximum second product is corresponded to Interest as the current interest.
- 9. information push-delivery apparatus according to claim 7, it is characterised in that described device further includes：Second acquisition module, for obtaining at least one forecasting sequence, each forecasting sequence adds one including the sequence of scenes Predict scene, the prediction scene is the scene in the scene set, and the prediction scene in each forecasting sequence is not Together；3rd computing module, for each forecasting sequence obtained for second acquisition module, according to the pre- sequencing The probability of row, the interest set and forecasting sequence described in the model parameter calculation；Second pushing module, the corresponding forecasting sequence of maximum probability in the probability calculated for determining the 3rd computing module, And to the terminal pushed information in the prediction scene that the maximum probability corresponds to forecasting sequence.
- 10. information push-delivery apparatus according to claim 9, it is characterised in that the 3rd computing module includes：6th computing unit, for forecasting sequence, the interest set and the mould obtained according to second acquisition module The probability of the corresponding prediction interest of scene is predicted in forecasting sequence described in shape parameter forward calculation, by the probability of the prediction interest Probability as the forecasting sequence；Alternatively,7th computing unit, for forecasting sequence, the interest set and the mould obtained according to second acquisition module The probability of the 1st corresponding interest of scene in forecasting sequence described in shape parameter backwards calculation, using the probability of the interest as institute State the probability of forecasting sequence.
- 11. information push-delivery apparatus according to claim 10, it is characterised in that the 6th computing unit, by based on When calculating the probability of each interest in the 1st scene, by i-th of the interest that the acquiring unit obtains in the probability square I-th of the interest that probability and second computing unit in battle array calculate produces described 1st in the emission matrix Probability of the product of the emission probability of scene as i-th of the interest in the 1st scene；When calculating the probability of each interest in q-th of scene, first calculate successively j-th that the 6th computing unit calculates J-th of the interest that probability and first computing unit of the interest in previous scene calculate is in the transition probability matrix The 3rd product of the transition probability of i-th of interest is transferred to, calculates the 3rd sum of products after the 3rd product addition, then calculate I-th of the interest that 3rd sum of products and second computing unit calculate produces described q-th in the emission matrix 4th product of the emission probability of scene, the 4th product is general in q-th of scene as i-th of the interest Rate, 2≤q≤n-1；When calculating the probability of each interest in n-th of scene, first calculate successively j-th that the 6th computing unit calculates J-th of the interest that probability and first computing unit of the interest in previous scene calculate is in the transition probability matrix The 3rd product of the transition probability of i-th of interest is transferred to, calculates the 3rd sum of products after the 3rd product addition, then calculate I-th of the interest that 3rd sum of products and second computing unit calculate produces described n-th in the emission matrix 4th product of the emission probability of scene, the 4th product is general in the n scene as i-th of the interest Rate, determines the 4th product of maximum in the 4th product, using maximum 4th product as being predicted in the forecasting sequence The probability of the corresponding prediction interest of scene.
- 12. information push-delivery apparatus according to claim 10, it is characterised in that the 7th computing unit, by based on When calculating the probability of each interest in n-th of scene, by i-th of the interest that the acquiring unit obtains in the probability square I-th of the interest that probability and second computing unit in battle array calculate produces described n-th in the emission matrix Probability of the product of the emission probability of scene as i-th of the interest in n-th of scene；When calculating the probability of each interest in q-th of scene, first calculate successively j-th that the 7th computing unit calculates J-th of the interest that probability and first computing unit of the interest in the latter scene calculate is in the transition probability matrix The 5th product of the transition probability of i-th of interest is transferred to, calculates the 5th sum of products after the 5th product addition, then calculate I-th of the interest that 5th sum of products and second computing unit calculate produces described q-th in the emission matrix 6th product of the emission probability of scene, the 6th product is general in q-th of scene as i-th of the interest Rate, 2≤q≤n-1；When calculating the probability of each interest in the 1st scene, first calculate successively j-th that the 7th computing unit calculates J-th of the interest that probability and first computing unit of the interest in the latter scene calculate is in the transition probability matrix The 5th product of the transition probability of i-th of interest is transferred to, calculates the 5th sum of products after the 5th product addition, then calculate I-th of the interest that 5th sum of products and second computing unit calculate produces described 1st in the emission matrix 6th product of the emission probability of scene, it is general in the 1st scene using the 6th product as i-th of the interest Rate, determines the 6th product of maximum in the 6th product, using maximum 6th product as the 1st in the forecasting sequence The probability of the corresponding interest of scene.
- 13. a kind of terminal, it is characterised in that the terminal includes the information push as any one of claim 7 to 12 Device.
- 14. a kind of server, it is characterised in that the server includes the information as any one of claim 7 to 12 Pusher.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210562450.0A CN103888498B (en) | 2012-12-21 | 2012-12-21 | Information-pushing method, device, terminal and server |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210562450.0A CN103888498B (en) | 2012-12-21 | 2012-12-21 | Information-pushing method, device, terminal and server |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103888498A CN103888498A (en) | 2014-06-25 |
| CN103888498B true CN103888498B (en) | 2018-04-13 |
Family
ID=50957216
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210562450.0A Active CN103888498B (en) | 2012-12-21 | 2012-12-21 | Information-pushing method, device, terminal and server |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103888498B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106339505A (en) * | 2016-09-27 | 2017-01-18 | 电子科技大学 | Music recommendation method based on Markov chain |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104572857B (en) * | 2014-12-17 | 2018-03-13 | 深圳先进技术研究院 | A kind of Web page push method and device |
| CN105373608A (en) * | 2015-11-13 | 2016-03-02 | 广州市久邦数码科技有限公司 | Input method based scene-mode content pushing method and system |
| CN113836438B (en) * | 2021-09-14 | 2022-09-02 | 上海任意门科技有限公司 | Method, electronic device, and storage medium for post recommendation |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6839680B1 (en) * | 1999-09-30 | 2005-01-04 | Fujitsu Limited | Internet profiling |
| CN101079063A (en) * | 2007-06-25 | 2007-11-28 | 腾讯科技（深圳）有限公司 | Method, system and apparatus for transmitting advertisement based on scene information |
| CN101866341A (en) * | 2009-04-17 | 2010-10-20 | 华为技术有限公司 | Information push method, device and system |
| CN102130933A (en) * | 2010-01-13 | 2011-07-20 | 中国移动通信集团公司 | Recommending method, system and equipment based on mobile Internet |
| CN102681999A (en) * | 2011-03-08 | 2012-09-19 | 阿里巴巴集团控股有限公司 | Method and device for collecting and sending user action information |
-
2012
- 2012-12-21 CN CN201210562450.0A patent/CN103888498B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6839680B1 (en) * | 1999-09-30 | 2005-01-04 | Fujitsu Limited | Internet profiling |
| CN101079063A (en) * | 2007-06-25 | 2007-11-28 | 腾讯科技（深圳）有限公司 | Method, system and apparatus for transmitting advertisement based on scene information |
| CN101866341A (en) * | 2009-04-17 | 2010-10-20 | 华为技术有限公司 | Information push method, device and system |
| CN102130933A (en) * | 2010-01-13 | 2011-07-20 | 中国移动通信集团公司 | Recommending method, system and equipment based on mobile Internet |
| CN102681999A (en) * | 2011-03-08 | 2012-09-19 | 阿里巴巴集团控股有限公司 | Method and device for collecting and sending user action information |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106339505A (en) * | 2016-09-27 | 2017-01-18 | 电子科技大学 | Music recommendation method based on Markov chain |
| CN106339505B (en) * | 2016-09-27 | 2019-09-27 | 电子科技大学 | A kind of music recommended method based on markov chain |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103888498A (en) | 2014-06-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107515909B (en) | Video recommendation method and system | |
| CN107193858B (en) | Intelligent Service application platform and method towards multi-source heterogeneous data fusion | |
| CN105046515B (en) | Method and device for sorting advertisements | |
| CN102075352B (en) | Method and device for predicting network user behavior | |
| CN110008408A (en) | A kind of session recommended method, system, equipment and medium | |
| CN109615128A (en) | Real estate client's conclusion of the business probability forecasting method, device and server | |
| CN109087130A (en) | A kind of recommender system and recommended method based on attention mechanism | |
| CN103795697B (en) | A kind of network media information dispensing effect simulation method and system | |
| CN108305094A (en) | A kind of user's behavior prediction method and device, electronic equipment | |
| CN103888498B (en) | Information-pushing method, device, terminal and server | |
| CN110489655A (en) | Hot content determination, recommended method, device, equipment and readable storage medium storing program for executing | |
| CN104883585B (en) | Show method, equipment and the system of media data | |
| CN103500213B (en) | Page hot-spot resource updating method and device based on pre-reading | |
| CN104992348B (en) | A kind of method and apparatus of information displaying | |
| CN109993583A (en) | Information-pushing method and device, storage medium and electronic device | |
| CN104951518A (en) | Context recommending method based on dynamic incremental updating | |
| CN105930507A (en) | Method and apparatus for obtaining Web browsing interest of user | |
| CN113254719B (en) | Online social network information propagation method based on status theory | |
| CN111523940B (en) | Deep reinforcement learning-based recommendation method and system with negative feedback | |
| CN107454346A (en) | Movie data analytic method, video production template recommend method, apparatus and equipment | |
| CN112699310A (en) | Cold start cross-domain hybrid recommendation method and system based on deep neural network | |
| CN115221396A (en) | Information recommendation method and device based on artificial intelligence and electronic equipment | |
| Wang | A survey of online advertising click-through rate prediction models | |
| CN104008204A (en) | Dynamic multi-dimensional context awareness film recommending system and achieving method thereof | |
| CN112053184B (en) | Popularization information delivery method and device, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |