CN105095258A - Media information sorting method and apparatus and media information recommendation system - Google Patents
Media information sorting method and apparatus and media information recommendation system Download PDFInfo
- Publication number
- CN105095258A CN105095258A CN201410192707.7A CN201410192707A CN105095258A CN 105095258 A CN105095258 A CN 105095258A CN 201410192707 A CN201410192707 A CN 201410192707A CN 105095258 A CN105095258 A CN 105095258A
- Authority
- CN
- China
- Prior art keywords
- media information
- recommendation
- value
- data
- user
- 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
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Embodiments of the present invention provide a media information sorting method and apparatus and a media information recommendation system. The method comprises: receiving positive weight trigger data generated based on triggering a positive weight component, and determining top data of media information from the positive weight trigger data; receiving reply trigger data generated based on triggering a reply component, and determining reply data of the media information from the reply trigger data; determining attribute data of the media information; determining a sorting value of the media information based on the top data, reply data and attribute data of the media information; and sorting the media information based on the sorting value. According to the method provided by the embodiment of the present invention, an accuracy rate of media information sorting is improved, cheating is effectively prevented and a Matthew effect can be reduced.
Description
Technical field
Embodiment of the present invention relates to technical field of internet application, more specifically, relates to a kind of media information sort method, device and media information recommending system.
Background technology
In the current information age, various information equipment arises at the historic moment: have landline telephone, the mobile phone for Tone Via; There are the server for information resources share, process and PC; There are various televisors shown for video data etc.These equipment are all produce for solving actual demand in specific area.Along with the arrival that E-consumer, computing machine, communication (3C) are merged, notice has been put in the research that fully utilizes the information equipment of each different field by people more and more, better serves to make full use of existing resource equipment for people.
Social network refers to person-to-person relational network, this website based on social networking relationships system thinking is exactly social network website, special finger is intended to help people to set up the internet, applications service of social network, also refers to the existing ripe universal information carrier of society.Commentary Systems are simple social intercourse systems.After having browsed theme, user often can provide various comment.Preferentially can showing for concerned higher comment, therefore just relating to commenting on the problem sorted.
In the prior art, Commentary Systems sort to comment according to the quantity on top mostly, and such sortord is too dull, and are easily subject to cheating impact, and accuracy of therefore sorting is not high.And this sortord does not consider adding of the law of cooling, therefore also serious Matthew effect can be caused.
Summary of the invention
Embodiment of the present invention proposes a kind of media information sort method, to improve sequence accuracy rate.
Embodiment of the present invention also proposed a kind of media information collator, to improve sequence accuracy rate.
Embodiment of the present invention also proposed a kind of media information recommending system, to improve information recommendation quality.
The concrete scheme of embodiment of the present invention is as follows:
A kind of media information sort method, the method comprises:
Receive the positive weights trigger data produced based on triggering positive weights assembly, from described positive weights trigger data, determine the useful certificate of media information; Receiving based on triggering the reply trigger data of replying assembly and producing, from described reply trigger data, determining the reply data of described media information; Determine the attribute data of described media information;
Attribute data based on the useful certificate of described media information, the reply data of described media information and described media information determines the ranking value of described media information;
Based on described ranking value, media information is sorted.
A kind of media information collator, comprising:
Data determination unit, for receiving the positive weights trigger data produced based on triggering positive weights assembly, determines the useful certificate of media information from described positive weights trigger data; Receiving based on triggering the reply trigger data of replying assembly and producing, from described reply trigger data, determining the reply data of described media information; Determine the attribute data of described media information;
Sorting values determination unit, for determining the ranking value of described media information based on the attribute data of the useful certificate of described media information, the reply data of described media information and described media information;
Sequencing unit, for sorting to media information based on described ranking value.
A kind of media information recommending system, comprising:
Issue client terminal, for sending media information to server;
Server, for issuing described media information, receiving the positive weights trigger data produced based on triggering positive weights assembly, determining the useful certificate of media information from described positive weights trigger data; Receiving based on triggering the reply trigger data of replying assembly and producing, from described reply trigger data, determining the reply data of described media information; Determine the attribute data of described media information; Attribute data based on the useful certificate of described media information, the reply data of described media information and described media information determines the ranking value of described media information; Based on described ranking value, media information is sorted.
As can be seen from technique scheme, in embodiments of the present invention, receive the positive weights trigger data produced based on triggering positive weights assembly, from described positive weights trigger data, determine the useful certificate of media information; Receiving based on triggering the reply trigger data of replying assembly and producing, from described reply trigger data, determining the reply data of described media information; Determine the attribute data of described media information; Attribute data based on the useful certificate of described media information, the reply data of described media information and described media information determines the ranking value of described media information; Based on described ranking value, media information is sorted, as can be seen here, embodiment of the present invention no longer sorts to the media information such as such as commenting on according to the quantity on top, but based on media information useful certificate, reply the ranking value that data and media information attribute data comprehensively determine media information, thus can effectively prevent cheating, improve sequence accuracy.And, after application embodiment of the present invention, regularly can determine to eliminate and recommend media information, thus effectively reduce the impact of Matthew effect.
In addition, after application embodiment of the present invention, no longer determine the credit rating of user according to the individual subscriber behavior such as login duration, dispatch frequency of user.On the contrary, to serve as a fill-in and the credit rating at very useful family determines to send the credit rating of the user of media information according to the history at very useful family, very useful family, taken into full account the behavior of other user, the user credit degree accuracy rate therefore determined is high.In addition, embodiment of the present invention avoids the defect according to easily there is cheating in individual subscriber behavior decision user credit degree, thus further increases user credit degree accuracy rate.
Accompanying drawing explanation
Fig. 1 is according to embodiment of the present invention media information sort method process flow diagram;
Fig. 2 is according to embodiment of the present invention top relation chain schematic diagram;
Fig. 3 is according to embodiment of the present invention media information collator schematic diagram;
Fig. 4 is according to embodiment of the present invention media information recommending system first structural drawing;
Fig. 5 is according to embodiment of the present invention media information recommending system second structural drawing.
Embodiment
For making the object, technical solutions and advantages of the present invention clearly, below in conjunction with accompanying drawing, the present invention is described in further detail.
Succinct and directly perceived in order to what describe, hereafter by description some representational embodiments, the solution of the present invention is set forth.Details a large amount of in embodiment only understands the solution of the present invention for helping.But these details can be not limited to when clearly, technical scheme of the present invention realizes.In order to avoid unnecessarily fuzzy the solution of the present invention, some embodiments do not describe meticulously, but only give framework.Hereinafter, " comprising " refers to " including but not limited to ", " according to ... " refer to " at least basis ..., but be not limited to only basis ... ".Due to the speech habits of Chinese, when hereinafter not particularly pointing out the quantity of a composition, mean that this composition can be one also can be multiple, or can be regarded as at least one.
Fig. 1 is the media information sort method process flow diagram according to embodiment of the present invention.
As shown in Figure 1, the method comprises:
Step 101: receive the positive weights trigger data produced based on triggering positive weights assembly, determine the useful certificate of media information from described positive weights trigger data; Receiving based on triggering the reply trigger data of replying assembly and producing, from described reply trigger data, determining the reply data of described media information; Determine the attribute data of described media information;
Step 102: based on media information useful certificate, reply the ranking value of data and media information attribute data determination media information;
Step 103: media information is sorted based on ranking value.
Subscription client has positive weights assembly.Positive weights assembly can comprise various types of positive weights icon (button such as pushed up).User browses when releasing news and triggers positive weights icon, can produce trigger data.Trigger data is for expressing the front praise to releasing news, and these release news and comprise media information.
Server receives user triggers the positive weights trigger data that positive weights assembly produces, and from positive weights trigger data, determine the useful certificate for media information that user triggers positive weights icon and produces.
Subscription client also has reply assembly.Reply assembly and can comprise various types of reply icon (ejecting the button of replying frame after such as triggering).User browses when releasing news and triggers reply assembly, can produce reply frame, and user can reply in frame the reply trigger data issued for releasing news.Reply trigger data for expressing the comment to releasing news, these release news and comprise media information.
Server receives user triggers the reply trigger data of replying assembly and producing, and determines that user triggers the reply data for media information of replying icon and producing from replying trigger data.
In one embodiment, the method comprises further:
To the ranking value execution time attenuation operations of media information, the ranking value after acquisition time decay;
Carrying out sequence to media information is:
Based on the ranking value after time decay, media information is sorted.
In one embodiment, the useful certificate of media information comprises the credit rating Pi of the user of top media information, and reply data and comprise the number of users P replying this media information, media information attribute data is the length value C of this media information; The ranking value execution time attenuation operations of media information is comprised: with time attenuation function (t
comm+ 2)
gthe ranking value of decay media information;
t (i) is the ranking value after time decay, wherein:
Work as t
iwith
between mistiming when being greater than predetermined value, t
commfor this predetermined value; Work as t
iwith
between mistiming when being not more than this predetermined value, t
commfor t
iwith
difference, wherein t
ifor the issuing time of this media information;
the issuing time mean value of all media informations under theme corresponding to this media information; N is the number of users of top media information; G is the parameter pre-set.Such as, this predetermined value can be 24.
In the application, (t
comm+ 2)
gbe a time attenuation function, be similar to Newton's law of cooling; G is a variable parameter, and this parameter can change according to different application.
In one embodiment, the method comprises further:
The media information of predetermined number is selected, using as recommendation media information according to ranking results;
Show and recommend media information.
In fact, ranking value computing formula can bring Matthew effect.Matthew effect refers to the phenomenon that powerhouse is stronger, weak person is more weak.For the purpose of justice, in the ranking value of various Commentary Systems calculates, Matthew effect should be avoided.Joining day attenuation factor can reduce Matthew effect, but also cannot avoid Matthew effect completely.
In order to better comment being put into the position of recommendation, embodiment of the present invention is introduced the comment to recommendation by a kind of real time record and superseded mechanism and is analyzed, the comment that superseded effect is worse.Eliminative mechanism obtains based on to draw a conclusion: in predetermined number (such as 10) the comment position of recommending, the comment votes amount that obtains and position have nothing to do.Based on this prerequisite, can record in each same time interval, the polled data that all recommendation positions produce, then eliminate the bad comment of feedback according to data analysis
In one embodiment, the method comprises further:
Determine in predetermined time interval, recommend the useful certificate of media information;
Media information is recommended according to determining to eliminate based on recommending serving as a fill-in of media information;
Recommend media information according to ranking results selection candidate, and utilize candidate to recommend media information to replace eliminating recommendation media information.
In one embodiment, based on recommending serving as a fill-in of media information to recommend media information to comprise according to determining to eliminate:
When t is-1, the superseded value t of calculated recommendation media information i, determines that this recommendation media information recommends media information for eliminating;
n is the number of recommending media information.
In one embodiment,
Up (i) is for have to the media information that user i sends all very useful family pushing work; P (j) performs to the media information that user i sends the credit rating pushing the very useful family j of work; N (j) is that the history of very useful family j is served as a fill-in, and E is the escape factor pre-set.
In one embodiment, the method comprises further:
When the iteration convergence condition pre-set does not meet, iteratively upgrade the credit rating Pi of user i; Wherein iteration convergence condition comprises: the difference of the front and back iteration result of the credit rating Pi of user i is less than the first threshold value pre-set, and/or iterations reaches the second threshold value pre-set.
Media information can be the news that various news channel sends, and also can be the micro-blog information that individual sends, and can also be the various comment follow-up in Usenet, living forum, circle of friends, and embodiment of the present invention is to this and indefinite.
In conjunction with above-mentioned algorithm, provide instantiation below.
Such as, under certain theme news (being assumed to topic1), there is comment comm1, and for this comment comm1, there is A, B, C, D tetra-user tops; The issuing time of this comment comm1 is: before 28 hours; And the averaging time of all comments under this theme news topic1 is 3 hours, wherein user A credit rating is 1, the credit rating of user B is 2, the credit rating of user C is 24, the credit rating 128 that user D is, arranging G value is 0.5, and comment comm1 length is 120, and the reply comment number for this comment comm1 is 2; Assuming that when the mistiming between the averaging time of all comments under the issuing time commented on and theme news topic1 is greater than predetermined value (such as 24 hours), t
commfor this predetermined value, be 24 hours; When comment on issuing time and theme news topic1 under all comments averaging time between mistiming be not more than this predetermined value time, t
commfor comment issuing time and theme news topic1 under all comments averaging time between mistiming.
Under the comment issuing time of comm1 and theme news topic1, the difference of the averaging time of all comments is (28-3)=25 hour >24 hour, therefore t
commbe 24.
Then the mark ranking value T of this comment comm1 is:
For another example, such as, under certain theme news (being assumed to topic1), there is comment comm2, and for this comment comm2, there is A, B, C tri-user tops; The issuing time of this comment comm1 is: before 22 hours; And the averaging time of all comments under this theme news topic1 is 3 hours, wherein user A credit rating is 1, and the credit rating of user B is 2, the credit rating of user C is 24, arranging G value is 0.5, and comment comm2 length is 100, and the reply comment number for this comment comm2 is 2; Assuming that when the mistiming between the averaging time of all comments under the issuing time commented on and theme news topic1 is greater than predetermined value (such as 24 hours), t
commfor this predetermined value, be 24 hours; When comment on issuing time and theme news topic1 under all comments averaging time between mistiming be not more than this predetermined value time, t
commfor comment issuing time and theme news topic1 under all comments averaging time between mistiming.
Under the comment issuing time of comm2 and theme news topic1, the difference of the averaging time of all comments is (22-3)=19 hour <24 hour, therefore t
commbe 19.
Then the mark ranking value T of this comment comm1 is:
In the above-described example, the credit rating of the user of top media information is related to.For the ease of calculating, also can the credit rating of all users be set to identical.
More preferably, embodiment of the present invention can also calculate the credit rating of user based on certain algorithm.
Such as, can to serve as a fill-in based on the history at each very useful family and the credit rating at each very useful family determines that each very useful family is to the contribution factor of user credit degree sending media information, wherein contribution factor is directly proportional to the credit rating at very useful family, and the history at contribution factor and very useful family serve as a fill-in be inversely proportional to; The contribution factor of each very useful family to the user credit degree sending media information is sued for peace, to obtain the credit rating of the user sending media information.
Such as, suppose that user A has issued media information, and user B, C, D are after browsing media information, all make the top of pushing up to user A and operate.And user B pushed up altogether 1 time (comprising this time top) in history; User C pushed up altogether 3 times (comprising this time top) in history; User D pushed up altogether 4 times (comprising this time top) in history; Suppose that the credit rating of user A is PeopleRank (A); The credit rating of user B is PeopleRank (B); The credit rating of user C is PeopleRank (C); Then
In one embodiment, if be exactly a very large digraph being gathered by useful certificate of all media informations of whole system.The contribution factor of each very useful family to the user credit degree sending media information is sued for peace, comprises to obtain the user credit degree sending media information:
Determine the credit rating Pi of the user i sending media information;
Up (i) is for have to the media information that user i sends all very useful family pushing work; P (j) performs the credit rating pushing the very useful family j of work; N (j) is that the history of very useful family j is served as a fill-in, and E is the escape factor pre-set.
Because media information continues in renewal, therefore media information by useful certificate also at continuous updating, therefore can being served as a fill-in to calculate according to successive ignition and sent the credit rating of the user of media information according to the media information upgraded.When after the credit rating convergence of user sending media information, be the PeopleRank value of user, this value is absolute value, can carry out all users be when in use divided into a N layer according to certain interval to user, each level gives certain value, facilitates backstage to manage user.
In one embodiment,
When the iteration convergence condition pre-set does not meet, iteratively upgrade the credit rating P (i) of user i; Wherein iteration convergence condition comprises: the difference of the front and back iteration result of the credit rating P (i) of user i is less than the first threshold value pre-set, and/or iterations reaches the second threshold value pre-set.
In one embodiment, to serve as a fill-in based on the history at very useful family, very useful family and credit rating that the credit rating at very useful family determines to send the user of media information comprises:
History based on very useful family, very useful family is served as a fill-in and is determined transition matrix;
All user credit degree vectors of assignment are in advance multiplied with transition matrix, to obtain all user credit degree matrixes;
When the iteration convergence condition pre-set does not meet, iteratively upgrade all user credit degree matrixes; Wherein iteration convergence condition comprises: the 3rd threshold value having the difference of the front and back iteration result of at least one user to be less than in all user credit degree matrixes to pre-set, and/or iterations reaches the 4th threshold value pre-set.
Numerous embodiments can be adopted specifically to calculate user credit degree of the present invention.Such as, the problem of the stationary distribution asking Markov chain can be converted in the calculation, carry out problem analysis from the angle of markov chain.
Such as, Fig. 2 is according to embodiment of the present invention top relation chain schematic diagram.
As seen from Figure 2, user A has pushed up user B, user C and user D; User B has pushed up user C and user A; User C has pushed up user D; User D has pushed up user A and user B.
The credit rating initial value that can pre-set each user is 1, then transition matrix M is:
Wherein transition matrix the first row first is classified as 0, represents that the number of times of the very useful family A of user A is 0; Transition matrix the first row second is classified as 1/2, represent that the number of times of the very useful family A of user B is 1, and user A is 2 by the total degree pushed up; Transition matrix the first row second is classified as 0, represents that the number of times of the very useful family A of user C is 0; Transition matrix the first row the 4th is classified as 1/2, represent that the number of times of the very useful family A of user D is 1, and user A is 2 by the total degree pushed up.
Similarly, transition matrix second row first is classified as 1/3, represents that the number of times of user A very useful family B is 1, and user B to be pushed up total degree be 3; Transition matrix second row second is classified as 0, represents that the number of times of the very useful family B of user B is 0; Transition matrix second row the 3rd is classified as 0, represents that the number of times of the very useful family B of user C is 0; Transition matrix second row the 4th is classified as 1/2, represent that the number of times of the very useful family B of user D is 1, and user B is 2 by the total degree pushed up.
Similarly, transition matrix the third line first is classified as 1/3, represents that the number of times of user A very useful family C is 1, and user C to be pushed up total degree be 3; Transition matrix the third line second is classified as 1/2, represent that the number of times of the very useful family C of user B is 1, and user C is 2 by the total degree pushed up; Transition matrix the third line the 3rd is classified as 0, represents that the number of times of the very useful family C of user C is 0; Transition matrix the third line the 4th is classified as 0, represents that the number of times of the very useful family C of user D is 0.
Similarly, transition matrix fourth line first is classified as 1/3, represents that the number of times of user A very useful family D is 1, and user D to be pushed up total degree be 3; Transition matrix fourth line second is classified as 0, represents that the number of times of the very useful family D of user B is 0; Transition matrix fourth line the 3rd is classified as 1, represents that the number of times of the very useful family D of user C is 1; Transition matrix fourth line the 4th is classified as 0, represents that the number of times of the very useful family D of user D is 0.
For each user, the credit rating initial value that can pre-set each user is 1, then credit rating vector matrix is
The result that matrix M is multiplied by vector matrix V is exactly the credit rating vector matrix after this calculates, namely
Therefore, the credit rating of user A is 1; The credit rating of user B is 5/6; The credit rating of user C is 5/6; The credit rating of user D is 4/3.
In iterative computation, vector matrix V is multiplied by transition matrix M again, is iteration result, for
Then, continue again to be multiplied by transition matrix M with vector matrix V, until convergence appears in the credit rating of each user.
In actual use, the data accumulating certain hour can make PeopleRank more effective.After obtaining the PeopleRank data of each user, a PeopleRank value is had to each user, can change a mode to understand, the value of PeopleRank value representative of consumer in Commentary Systems of user, all behaviors of that user can have relation with this value.Such as, PeopleRank value can be utilized to replace the assessment of the useful certificate of user: before, one withstands on top time introducing formula carries out calculating and only represents a radix, after quoting PeopleRank, carry out of the value on the top of user according to PeopleRank is replaced, can more effective recommendation comment warmly like this.In the use of reality, embody to some extent, successful.
In addition in user management, the liveness of the reaction user of the credit rating of user, has good value for the assessment of high-quality user and definition.
In one embodiment, the method comprises further:
Based on the credit rating order of each user, each user is sorted;
According to ranking results, determine that the user of predetermined number is using as recommendation user, will recommend the review information of user as recommendation review information; And/or according to ranking results, determine that the user of predetermined number is using as shielding user, using the review information of shielding user as shielding review information.
Based on above-mentioned labor, embodiment of the present invention also proposed a kind of user credit degree determining device.
Fig. 3 is according to embodiment of the present invention media information collator structural drawing.
As shown in Figure 3, media information collator comprises:
Data determination unit 301, for receiving the positive weights trigger data produced based on triggering positive weights assembly, determines the useful certificate of media information from described positive weights trigger data; Receiving based on triggering the reply trigger data of replying assembly and producing, from described reply trigger data, determining the reply data of described media information; Determine the attribute data of described media information;
Sorting values determination unit 302, for based on media information useful certificate, reply the ranking value of data and media information attribute data determination media information;
Sequencing unit 303, for sorting to media information based on ranking value.
In one embodiment, comprise attenuation processing unit 304 further, for the ranking value execution time attenuation operations to media information, the ranking value after acquisition time decay;
Sequencing unit 303, for sorting to media information based on the ranking value after time decay.
In one embodiment, the useful certificate of media information comprises the credit rating Pi of the user of top media information, and reply data and comprise the number of users P replying this media information, media information attribute data is the length value C of this media information; The ranking value execution time attenuation operations of media information is comprised: with time attenuation function (t
comm+ 2)
gthe ranking value of decay media information;
t (i) is the ranking value after time decay, wherein:
Work as t
iwith
between mistiming when being greater than predetermined value, t
commfor this predetermined value; Work as t
iwith
between mistiming when being not more than this predetermined value, t
commfor t
iwith
difference, wherein t
ifor the issuing time of this media information;
the issuing time mean value of all media informations under theme corresponding to this media information; N is the number of users of top media information; G is the parameter pre-set.
In one embodiment, also comprise and recommend media information determining unit 305, for selecting the media information of predetermined number according to ranking results, using as recommendation media information, and show and recommend media information.
In one embodiment, media information determining unit 305 is recommended, also for determining the useful certificate of recommending media information in predetermined time interval; Media information is recommended according to determining to eliminate based on recommending serving as a fill-in of media information; Recommend media information according to ranking results selection candidate, and utilize candidate to recommend media information to replace eliminating recommendation media information.
In one embodiment, recommending media information determining unit 305, for the superseded value t of calculated recommendation media information i, determining that this recommendation media information recommends media information for eliminating when t is-1;
n is the number of recommending media information.
In the present invention's application can being applied with various media releasing.Such as, can be applied to the news that news channel sends, also can be the micro-blog information that individual sends, and can also be that various in Usenet, living forum, circle of friends are posted, and embodiment of the present invention is to this and indefinite.
Fig. 4 is according to embodiment of the present invention media information recommending system first structural drawing;
As shown in Figure 4, this system comprises:
Multiple issue client terminal 401, for sending media information to server 402;
Server 402, for publication medium information, receives the positive weights trigger data produced based on triggering positive weights assembly, determines the useful certificate of media information from described positive weights trigger data; Receiving based on triggering the reply trigger data of replying assembly and producing, from described reply trigger data, determining the reply data of described media information; Determine the attribute data of described media information; Based on media information useful certificate, reply the ranking value of data and media information attribute data determination media information; Based on ranking value, media information is sorted.Figure 4 shows that the schematic diagram of ranking results list.
Wherein, user can perform the process sending media information on various issue client terminal 401, these issue client terminals can include, but are not limited to: functional mobile phone, smart mobile phone, palm PC, PC (PC), panel computer or personal digital assistant (PDA), etc.
Although enumerate the instantiation of issue client terminal above in detail, those skilled in the art can recognize, these are enumerated is only purposes of illustration, is not intended to limit the present invention the protection domain of embodiment.And the browser that these publishing sides can adopt specifically can comprise the Safari of Firefox, Apple of InternetExplorer, Mozilla of Microsoft, the browsers such as Opera, GoogleChrome, GreenBrowser.
Although list some conventional browsers above in detail, those skilled in the art can recognize, embodiment of the present invention is not limited to these browsers, but go for the file that can be used for arbitrarily in display web page server or archives economy and allow the application (App) of user and file interaction, these application can be various browsers common at present, also can be other the application programs arbitrarily with web page browsing function.
In one embodiment, server 402, also for the ranking value execution time attenuation operations to media information, the ranking value after acquisition time decay, and based on the ranking value after time decay, media information is sorted.
In one embodiment 402, server, also for selecting the media information of predetermined number according to ranking results, using as recommendation media information; Show and recommend media information.
Fig. 5 is according to embodiment of the present invention media information recommending system second structural drawing;
As shown in Figure 5, this system comprises:
Multiple issue client terminal 401, for sending media information to server 402;
Server 402, for publication medium information, receives the positive weights trigger data produced based on triggering positive weights assembly, determines the useful certificate of media information from described positive weights trigger data; Receiving based on triggering the reply trigger data of replying assembly and producing, from described reply trigger data, determining the reply data of described media information; Determine the attribute data of described media information; Based on media information useful certificate, reply the ranking value of data and media information attribute data determination media information; Based on ranking value, media information is sorted.
Wherein, user can perform the process sending media information on various issue client terminal 401, these issue client terminals can include, but are not limited to: functional mobile phone, smart mobile phone, palm PC, PC (PC), panel computer or personal digital assistant (PDA), etc.
Although enumerate the instantiation of issue client terminal above in detail, those skilled in the art can recognize, these are enumerated is only purposes of illustration, is not intended to limit the present invention the protection domain of embodiment.And the browser that these publishing sides can adopt specifically can comprise the Safari of Firefox, Apple of InternetExplorer, Mozilla of Microsoft, the browsers such as Opera, GoogleChrome, GreenBrowser.
Although list some conventional browsers above in detail, those skilled in the art can recognize, embodiment of the present invention is not limited to these browsers, but go for the file that can be used for arbitrarily in display web page server or archives economy and allow the application (App) of user and file interaction, these application can be various browsers common at present, also can be other the application programs arbitrarily with web page browsing function.
In one embodiment, server 502, also for determining the useful certificate of recommending media information in predetermined time interval; Media information is recommended according to determining to eliminate based on recommending serving as a fill-in of media information; Recommend media information according to ranking results selection candidate, and utilize candidate to recommend media information to replace eliminating recommendation media information.
As shown in Figure 5, server is after determining multiple superseded recommendation media information, can be formed to eliminate and recommend media information list, and recommend media information according to precalculated media information ranking results selection candidate, and utilize candidate to recommend media information to replace eliminating recommendation media information.Illustrate in Fig. 5 to eliminate and recommend media information list.
In fact, can specifically implement by various ways media information sort method, device and the media information recommending system that embodiment of the present invention proposes.
Such as, the application programming interfaces of certain specification can be followed, media information sort method is written as the plug-in card program be installed in PC, mobile terminal etc., also can be encapsulated as application program and download use voluntarily for user.When being written as plug-in card program, the multiple card format such as ocx, dll, cab can be implemented as.Also can implement by the concrete technology such as Flash plug-in unit, RealPlayer plug-in unit, MMS plug-in unit, MIDI staff plug-in unit, ActiveX plug-in unit the user credit degree defining method that embodiment of the present invention proposes.
The media information sort method that embodiment of the present invention is proposed by the storing mode that can be stored by instruction or instruction set is stored on various storage medium.These storage mediums include, but are not limited to: floppy disk, CD, DVD, hard disk, flash memory, USB flash disk, CF card, SD card, mmc card, SM card, memory stick (MemoryStick), xD card etc.
In addition, the media information sort method that embodiment of the present invention can also be proposed is applied in the storage medium based on flash memory (Nandflash), such as USB flash disk, CF card, SD card, SDHC card, mmc card, SM card, memory stick, xD card etc.
To sum up, in embodiments of the present invention, receive the positive weights trigger data produced based on triggering positive weights assembly, from described positive weights trigger data, determine the useful certificate of media information; Receiving based on triggering the reply trigger data of replying assembly and producing, from described reply trigger data, determining the reply data of described media information; Determine the attribute data of described media information; Based on media information useful certificate, reply the ranking value of data and media information attribute data determination media information; Based on ranking value, media information is sorted, as can be seen here, no longer according to the quantity on top, comment is sorted, but based on media information useful certificate, reply the ranking value that data and media information attribute data comprehensively determine media information, thus can effectively prevent cheating, improve sequence accuracy.And, after application embodiment of the present invention, regularly can determine to eliminate and recommend media information, thus effectively reduce the impact of Matthew effect.
In addition, after application embodiment of the present invention, no longer determine the credit rating of user according to the individual subscriber behavior such as login duration, dispatch frequency of user.On the contrary, to serve as a fill-in and the credit rating at very useful family determines to send the credit rating of the user of media information according to the history at very useful family, very useful family, taken into full account the behavior of other user, the user credit degree accuracy rate therefore determined is high.In addition, embodiment of the present invention avoids the defect according to easily there is cheating in individual subscriber behavior decision user credit degree, thus further increases user credit degree accuracy rate.
Above, be only preferred embodiment of the present invention, be not intended to limit protection scope of the present invention.Within the spirit and principles in the present invention all, any amendment done, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (18)
1. a media information sort method, is characterized in that, the method comprises:
Receive the positive weights trigger data produced based on triggering positive weights assembly, from described positive weights trigger data, determine the useful certificate of media information; Receiving based on triggering the reply trigger data of replying assembly and producing, from described reply trigger data, determining the reply data of described media information; Determine the attribute data of described media information;
Attribute data based on the useful certificate of described media information, the reply data of described media information and described media information determines the ranking value of described media information;
Based on described ranking value, media information is sorted.
2. media information sort method according to claim 1, it is characterized in that, the method comprises further:
To the ranking value execution time attenuation operations of described media information, the ranking value after acquisition time decay;
Describedly sequence carried out to media information be:
Based on the ranking value after time decay, media information is sorted.
3. media information sort method according to claim 2, it is characterized in that, the useful certificate of described media information comprises the credit rating Pi of the user pushing up described media information, described reply data comprise the number of users P replying this media information, and described media information attribute data is the length value C of this media information; The described ranking value execution time attenuation operations to described media information comprises: with time attenuation function (t
comm+ 2)
gdecay the ranking value of described media information;
t (i) is the ranking value after described time decay, wherein:
Work as t
iwith
between mistiming when being greater than predetermined value, t
commfor this predetermined value; Work as t
iwith
between mistiming when being not more than this predetermined value, t
commfor t
iwith
difference; t
ifor the issuing time of this media information;
the issuing time mean value of all media informations under theme corresponding to this media information; N is the number of users pushing up described media information; G is the parameter pre-set.
4. media information sort method according to claim 1, it is characterized in that, the method comprises further:
The media information of predetermined number is selected, using as recommendation media information according to described ranking results;
Show described recommendation media information.
5. media information sort method according to claim 4, it is characterized in that, the method comprises further:
Determine in predetermined time interval, the useful certificate of described recommendation media information;
Media information is recommended according to determining to eliminate based on serving as a fill-in of described recommendation media information;
Select candidate to recommend media information according to described ranking results, and utilize described candidate to recommend media information to replace described superseded recommendation media information.
6. media information sort method according to claim 5, is characterized in that, described based on described recommendation media information serve as a fill-in according to determine eliminate recommend media information comprise:
When t is-1, the superseded value t of calculated recommendation media information i, determines that this recommendation media information i recommends media information for eliminating;
n is the number of recommending media information.
7. media information sort method according to claim 3, is characterized in that, described in
Up (i) is for have to the media information that user i sends all very useful family pushing work; P (j) performs to the media information that user i sends the credit rating pushing the very useful family j of work; N (j) is that the history of very useful family j is served as a fill-in, and E is the escape factor pre-set.
8. media information sort method according to claim 7, it is characterized in that, the method comprises further:
When the iteration convergence condition pre-set does not meet, iteratively upgrade the credit rating Pi of described user i; Wherein said iteration convergence condition comprises: the difference of the front and back iteration result of the credit rating Pi of described user i is less than the first threshold value pre-set, and/or described iterations reaches the second threshold value pre-set.
9. a media information collator, is characterized in that, comprising:
Data determination unit, for receiving the positive weights trigger data produced based on triggering positive weights assembly, determines the useful certificate of media information from described positive weights trigger data; Receiving based on triggering the reply trigger data of replying assembly and producing, from described reply trigger data, determining the reply data of described media information; Determine the attribute data of described media information;
Sorting values determination unit, for determining the ranking value of described media information based on the attribute data of the useful certificate of described media information, the reply data of described media information and described media information;
Sequencing unit, for sorting to media information based on described ranking value.
10. media information collator according to claim 9, is characterized in that, comprises attenuation processing unit further, for the ranking value execution time attenuation operations to described media information, and the ranking value after acquisition time decay;
Described sequencing unit, for sorting to media information based on the ranking value after time decay.
11. media information collators according to claim 10, it is characterized in that, the useful certificate of described media information comprises the credit rating Pi of the user pushing up described media information, described reply data comprise the number of users P replying this media information, and described media information attribute data is the length value C of this media information; The described ranking value execution time attenuation operations to described media information comprises: with time attenuation function (t
comm+ 2)
gdecay the ranking value of described media information;
t (i) is the ranking value after described time decay, wherein:
Work as t
iwith
between mistiming when being greater than predetermined value, t
commfor this predetermined value; Work as t
iwith
between mistiming when being not more than this predetermined value, t
commfor t
iwith
difference; Wherein t
ifor the issuing time of this media information;
the issuing time mean value of all media informations under theme corresponding to this media information; N is the number of users pushing up described media information; G is the parameter pre-set.
12. media information collators according to claim 10, it is characterized in that, also comprise and recommend media information determining unit, for selecting the media information of predetermined number according to described ranking results, using as recommendation media information, and show described recommendation media information.
13. media information collators according to claim 12, is characterized in that,
Recommend media information determining unit, also for determining the useful certificate of described recommendation media information in predetermined time interval; Media information is recommended according to determining to eliminate based on serving as a fill-in of described recommendation media information; Select candidate to recommend media information according to described ranking results, and utilize described candidate to recommend media information to replace described superseded recommendation media information.
14. media information collators according to claim 13, is characterized in that, described recommendation media information determining unit, for the superseded value t of calculated recommendation media information i, determine that this recommendation media information recommends media information for eliminating when t is-1;
n is the number of recommending media information.
15. 1 kinds of media information recommending systems, is characterized in that, comprising:
Issue client terminal, for sending media information to server;
Server, for issuing described media information, receiving the positive weights trigger data produced based on triggering positive weights assembly, determining the useful certificate of media information from described positive weights trigger data; Receiving based on triggering the reply trigger data of replying assembly and producing, from described reply trigger data, determining the reply data of described media information; Determine the attribute data of described media information; Attribute data based on the useful certificate of described media information, the reply data of described media information and described media information determines the ranking value of described media information; Based on described ranking value, media information is sorted.
16. media information recommending systems according to claim 15, is characterized in that,
Server, also for the ranking value execution time attenuation operations to described media information, the ranking value after acquisition time decay, and based on the ranking value after time decay, media information is sorted.
17. media information recommending systems according to claim 16, is characterized in that,
Server, also for selecting the media information of predetermined number according to described ranking results, using as recommendation media information; Show described recommendation media information.
18. media information recommending systems according to claim 17, is characterized in that,
Server, also for determining the useful certificate of described recommendation media information in predetermined time interval; Media information is recommended according to determining to eliminate based on serving as a fill-in of described recommendation media information; Select candidate to recommend media information according to described ranking results, and utilize described candidate to recommend media information to replace described superseded recommendation media information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410192707.7A CN105095258B (en) | 2014-05-08 | 2014-05-08 | A kind of media information sort method, device and media information recommending system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410192707.7A CN105095258B (en) | 2014-05-08 | 2014-05-08 | A kind of media information sort method, device and media information recommending system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105095258A true CN105095258A (en) | 2015-11-25 |
| CN105095258B CN105095258B (en) | 2019-06-21 |
Family
ID=54575712
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410192707.7A Active CN105095258B (en) | 2014-05-08 | 2014-05-08 | A kind of media information sort method, device and media information recommending system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105095258B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105740473A (en) * | 2016-03-14 | 2016-07-06 | 腾讯科技（深圳）有限公司 | User-generated content display method and device |
| CN105898336A (en) * | 2015-12-15 | 2016-08-24 | 乐视网信息技术（北京）股份有限公司 | Hot review determining method, hot review display system and server |
| CN107885808A (en) * | 2017-11-03 | 2018-04-06 | 四川文轩教育科技有限公司 | Shared resource file anti-cheating method |
| CN109213908A (en) * | 2018-08-01 | 2019-01-15 | 浙江工业大学 | A kind of academic meeting paper supplying system based on data mining |
| CN110087118A (en) * | 2019-04-26 | 2019-08-02 | 北京达佳互联信息技术有限公司 | Comment on message treatment method, device, terminal, server and medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2002330245A (en) * | 2001-05-02 | 2002-11-15 | Dainippon Printing Co Ltd | Super facsimile system |
| CN102449658A (en) * | 2009-03-18 | 2012-05-09 | 踏途音乐公司 | Entertainment server and associated social networking services |
| CN102970352A (en) * | 2012-11-08 | 2013-03-13 | 北京小米科技有限责任公司 | Method and device for providing information |
| CN103034706A (en) * | 2012-12-07 | 2013-04-10 | 合一网络技术(北京)有限公司 | Video recommended list generating device and method based on information network |
| CN103164414A (en) * | 2011-12-09 | 2013-06-19 | 腾讯科技（深圳）有限公司 | Multimedia file comment display method and multimedia file comment display system |
| CN103678623A (en) * | 2013-12-18 | 2014-03-26 | 乐视网信息技术（北京）股份有限公司 | Method for displaying browser page and terminal for displaying browser page |
-
2014
- 2014-05-08 CN CN201410192707.7A patent/CN105095258B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2002330245A (en) * | 2001-05-02 | 2002-11-15 | Dainippon Printing Co Ltd | Super facsimile system |
| CN102449658A (en) * | 2009-03-18 | 2012-05-09 | 踏途音乐公司 | Entertainment server and associated social networking services |
| CN103164414A (en) * | 2011-12-09 | 2013-06-19 | 腾讯科技（深圳）有限公司 | Multimedia file comment display method and multimedia file comment display system |
| CN102970352A (en) * | 2012-11-08 | 2013-03-13 | 北京小米科技有限责任公司 | Method and device for providing information |
| CN103034706A (en) * | 2012-12-07 | 2013-04-10 | 合一网络技术(北京)有限公司 | Video recommended list generating device and method based on information network |
| CN103678623A (en) * | 2013-12-18 | 2014-03-26 | 乐视网信息技术（北京）股份有限公司 | Method for displaying browser page and terminal for displaying browser page |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105898336A (en) * | 2015-12-15 | 2016-08-24 | 乐视网信息技术（北京）股份有限公司 | Hot review determining method, hot review display system and server |
| CN105740473A (en) * | 2016-03-14 | 2016-07-06 | 腾讯科技（深圳）有限公司 | User-generated content display method and device |
| CN107885808A (en) * | 2017-11-03 | 2018-04-06 | 四川文轩教育科技有限公司 | Shared resource file anti-cheating method |
| CN107885808B (en) * | 2017-11-03 | 2021-03-30 | 四川文轩教育科技有限公司 | Shared resource file anti-cheating method |
| CN109213908A (en) * | 2018-08-01 | 2019-01-15 | 浙江工业大学 | A kind of academic meeting paper supplying system based on data mining |
| CN110087118A (en) * | 2019-04-26 | 2019-08-02 | 北京达佳互联信息技术有限公司 | Comment on message treatment method, device, terminal, server and medium |
| CN110087118B (en) * | 2019-04-26 | 2022-01-21 | 北京达佳互联信息技术有限公司 | Comment message processing method, comment message processing device, comment message processing terminal, comment message processing server and comment message processing medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105095258B (en) | 2019-06-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110825957B (en) | Deep learning-based information recommendation method, device, equipment and storage medium | |
| US20160098735A1 (en) | Marketing channel attribution | |
| CN103049452A (en) | Method and device for performing application sequencing based on estimated download rate | |
| CN105095258A (en) | Media information sorting method and apparatus and media information recommendation system | |
| CN103502899A (en) | Dynamic predictive modeling platform | |
| CN103020846A (en) | Mobile application pushing method and system | |
| CN109242654A (en) | A kind of item recommendation method and system | |
| CN116541610B (en) | Training method and device for recommendation model | |
| CN111125521A (en) | Information recommendation method, device, equipment and storage medium | |
| CN111460290A (en) | Information recommendation method, device, equipment and storage medium | |
| JP2011227721A (en) | Interest extraction device, interest extraction method, and interest extraction program | |
| CN113254472A (en) | Parameter configuration method, device, equipment and readable storage medium | |
| CN112669078A (en) | Behavior prediction model training method, device, equipment and storage medium | |
| CN116629937A (en) | Marketing strategy recommendation method and device | |
| CN106790350A (en) | A kind of information push-delivery apparatus, server and method | |
| CN110516162B (en) | Information recommendation method, device, equipment and storage medium | |
| CN115935185A (en) | Training method and device for recommendation model | |
| CN110910201A (en) | Information recommendation control method and device, computer equipment and storage medium | |
| CN111680213A (en) | Information recommendation method, data processing method and device | |
| CN111768241A (en) | Interactive qualification grade determining method, device, equipment and storage medium | |
| CN105095257A (en) | User credit determining method and device and media information recommendation system | |
| CN114925275A (en) | Product recommendation method and device, computer equipment and storage medium | |
| CN111507471B (en) | Model training method, device, equipment and storage medium | |
| CN115185606A (en) | Method, device, equipment and storage medium for obtaining service configuration parameters | |
| CN110808899B (en) | Content sharing method, device, client, server and system |
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 |