CN104361062A - Associated information recommendation method and device - Google Patents
Associated information recommendation method and device Download PDFInfo
- Publication number
- CN104361062A CN104361062A CN201410610726.7A CN201410610726A CN104361062A CN 104361062 A CN104361062 A CN 104361062A CN 201410610726 A CN201410610726 A CN 201410610726A CN 104361062 A CN104361062 A CN 104361062A
- Authority
- CN
- China
- Prior art keywords
- node
- information node
- weighted value
- hops
- level message
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
- G06F16/78—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/7867—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using information manually generated, e.g. tags, keywords, comments, title and artist information, manually generated time, location and usage information, user ratings
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/951—Indexing; Web crawling techniques
Abstract
The invention provides an associated information recommendation method and device. The method comprises the steps of obtaining m first weight values corresponding to m second-level information nodes associated with a first-level information node, wherein the m first weight values are obtained through click jump behaviors of a user, the first-level information node is an information node which is used by the user at current, and the m is a positive integer; determining associated information to be recommended to the user according to the m first weight values and preset n recommended information nodes, wherein the n is a positive integer. The associated information recommendation method and device have the advantages that since the click jump behaviors of the user are referred through the m first weight values, the problem of content focusing for recommending node information to the user through the contents of information nodes, uploaders or writers of the information nodes, types of the information nodes and the like is avoided; since the m first weight values are obtained by referring to the click jump behaviors of the user, the effect of recommending associated information to users through a big data concept is realized.
Description
Technical field
The present invention relates to field of computer technology, particularly relate to a kind of recommend method and device of related information.
Background technology
At computer technology and internet arena, make one from search needs, to display carrying, arrive the closed-loop circulation experience that potential demand excites again, the recommendation of related information is then the committed step exciting potential demand, and it can shorten user's step-length and screening cost substantially, simultaneously, substantially flow can also be deposited in specific product system, thus build user's viscosity.
When carrying out the recommendation of related information, video is specially for the information of recommending, when prior art recommends video by content-based association to user, mainly from video, consider the similarity of video information, comprising: the information such as the scoring of video title (title), video presentation, video uploader, video, video click volume.Inventor finds, content-based recommendation mode, the associated video of recommendation is all close in content, therefore can produce the problem of Content aggregation, cannot excite the point of interest of user.
Summary of the invention
Embodiments of the invention provide a kind of recommend method and device of related information, avoid, by the content focus issues to user's recommended node information such as type of the content of information node, the uploader of information node or author, information node, exciting the point of interest of user.
For achieving the above object, embodiments of the invention adopt following technical scheme:
A recommend method for related information, the method comprises:
Obtain the m corresponding with m the second-level message node that first order information node is associated first weighted value, wherein, described m the first weighted value is got by the click redirect behavior of user, and described first order information node is the information node of the current use of described user, and m is positive integer;
Be the related information that described user determines to recommend according to described m the first weighted value and n the recommendation information node preset, wherein, n is positive integer.
A recommendation apparatus for related information, this device comprises:
First acquisition module, for m the first weighted value that m the second-level message node obtained with first order information node is associated is corresponding, wherein, described m the first weighted value is got by the click redirect behavior of user, described first order information node is the information node of the current use of described user, and m is positive integer;
Recommending module, for being the related information that described user determines to recommend according to described m the first weighted value and n default recommendation information node, wherein, n is positive integer.
The recommend method of the related information that the embodiment of the present invention provides and device, the click redirect behavior of user is with reference to by m the first weighted value, avoid by the content focus issues to user's recommended node information such as type of the content of information node, the uploader of information node or author, information node, obtain manner due to m the first weighted value is the click redirect behavior with reference to user, achieves and recommends related information by large data thought to user.
Accompanying drawing explanation
The schematic flow sheet of the recommend method of the related information that Fig. 1 provides for the embodiment of the present invention one.
The schematic flow sheet of the recommend method of the related information that Fig. 2 provides for the embodiment of the present invention two.
Fig. 3 is the schematic diagram of related information digraph embodiment illustrated in fig. 2.
The schematic flow sheet of the recommend method of the related information that Fig. 4 provides for the embodiment of the present invention three.
Fig. 5 is the schematic diagram of related information digraph embodiment illustrated in fig. 4.
The structural representation of the recommendation apparatus of the related information that Fig. 6 provides for the embodiment of the present invention four.
The structural representation of the recommendation apparatus of the related information that Fig. 7 provides for the embodiment of the present invention five.
Embodiment
Below in conjunction with accompanying drawing, the recommend method of embodiment of the present invention related information and device are described in detail.
Embodiment one:
The schematic flow sheet of the recommend method of the related information that Fig. 1 provides for the embodiment of the present invention one; As shown in Figure 1, the embodiment of the present invention comprises the steps:
Step 101, obtain the m corresponding with m the second-level message node that first order information node is associated first weighted value, wherein, m the first weighted value is got by the click redirect behavior of user, first order information node is the information node of the current use of user, and m is positive integer;
Step 102, be the related information that user determines to recommend according to m the first weighted value and n default recommendation information node, wherein, n is positive integer.
In embodiments of the present invention, first order information node to be A, m second-level message node be B1, B2 ..., Bm, A, B1, B2 ..., incidence relation between Bm represented by weight, weight is larger, shows that correlation degree is stronger.Weighted value can be got in conjunction with the click redirect behavior of user.
The recommend method of the related information that the embodiment of the present invention provides, the click redirect behavior of user is with reference to by m the first weighted value, avoid by the content focus issues to user's recommended node information such as type of the content of information node, the uploader of information node or author, information node, obtain manner due to m the first weighted value is the click redirect behavior with reference to user, achieves and recommends related information by large data thought to user.
Embodiment two:
The schematic flow sheet of the recommend method of the related information that Fig. 2 provides for the embodiment of the present invention two, Fig. 3 is the schematic diagram of related information digraph embodiment illustrated in fig. 2; The embodiment of the present invention is greater than or equal to the information node recommended to user number n for the number m of second-level message node carries out exemplary illustration, and as shown in Figure 2, the embodiment of the present invention comprises the steps:
Step 201, obtains the click number of hops of user at first order information node, wherein, clicks number of hops and comprises the first number of hops and the second number of hops.
Step 202, obtains m the first weighted value of m the second-level message node be associated with first order information node according to the first number of hops and the second number of hops.
Step 203, sorts m the first weighted value, obtains ranking results.
Step 204, according to ranking results from m second-level message node for user recommends n related information node.
Be described below in conjunction with Fig. 3, wherein, first order information node is A information node, second-level message node be B1, B2 ..., Bm information node, relative to the third level information node C1, C2 of A information node, further, C1 information node is the second-level message node of B1 information node, and C2 information node is the second-level message node of B2 information node.In embodiments of the present invention, with A information node, B1, B2 ..., to be specially video film be that example carries out exemplary illustration for Bm information node, C1 information node, C2 information node.
As shown in Figure 3, have A, B1, B2 ..., multiple video film such as Bm, C1, C2, A, B1, B2 ..., associated by oriented arrow between Bm, C1, C2, incidence relation each other can be represented by weight, weight is larger, shows that correlation degree is stronger.In step 201, the click redirect behavior of user by getting in history log, by obtaining the click redirect behavior of user, can count the click number of hops of user at A information node, wherein, click number of hops and comprise the first number of hops and the second number of hops, the click redirect behavior of user mainly contains two kinds of sources, one is initiatively click redirect by user, this initiatively the clicks redirect adopting consecutive click chemical reaction A video that is user within the time interval of a setting and B2 (or B3, Bm) behavior, the first number of hops can be obtained by the number of times of statistics active redirect, another is the redirect undertaken by the associated recommendation of video, belong to the redirect having guiding, namely, for when seeing A video, the video of associated recommendation is B1, B2, Bm, then jump to B1 by statistics by A video, B2, the click number of hops of Bm can obtain the second number of hops.
In step 202., equation can be passed through
calculate m the first weighted value, wherein, N (A, B) represent from first order information node A to B1, B2, first number of hops of Bm, N (A, *) represent from first order information node A to B1, B2, Bm altogether m second-level message node the first number of hops with, M (A, B) represent from first order information node A to B1, B2, second number of hops of Bm, M (A, *) represent from first order information node A to B1, B2, Bm altogether m second-level message node the second number of hops with, α and β is weight proportion, in one embodiment, alpha+beta=1.
Specifically, A and B1, B2 ..., weight weight between Bm
1(A, B1), weight
1(A, B2) ..., weight
1(A, Bm).In order to make A information node and B1, B2 ..., weight between Bm information node calculating more accurate, can also based on the similarity calculating method of term vector, obtain A information node respectively with B1, B2 ..., title between Bm information node and description similarity weight
c(A, B1), weight
c(A, B2) ..., weight
c(A, Bm).Be added by the first weighted value is weighted with above-mentioned similarity, thus obtain A and B1, B2 ..., more accurate weighted value between Bm, that is:
weight(A,B1)＝γ
1*weight
1(A,B1)+γ
2*weight
C(A,B1)，
weight(A,B2)＝γ
1*weight
1(A,B2)+γ
2*weight
C(A,B2)，
…
weight(A,Bm)＝γ
1*weight
1(A,Bm)+γ
2*weight
C(A,Bm)。
Wherein, γ
1with γ
2for weighting coefficient, γ
1+ γ
2=1.Further degree, can participate in follow-up recommendation process using this more accurate weighted value as the first weighted value, thus makes the information node recommended closer to user's.
In step 203 and step 204, due to m the first weighted value represent B1, B2 ..., correlation degree between Bm information node and A information node, weighted value larger expression relevance is stronger, therefore by weight
1(A, B1), weight
1(A, B2) ..., weight
1(A, Bm) descendingly to sort, obtain ranking results, such as, m=5, n=3, need to recommend 3 videos to user from 5 videos, ranking results is B4, B2, B3, B5, B1, according to ranking results from B1, B2 ..., in a B5 information node for user recommends B4, B2, B3 information node, the relevance of this B4, B2, B3 information node and A information node is stronger.
The recommend method of the related information that the embodiment of the present invention provides, recommended the information node be associated with first order information node to user in the click number of hops of first order information node by user, owing to being associated with the behavior of user when clicking and jump number of times, therefore the embodiment of the present invention avoids the content by information node, the uploader of information node or author, the problem that the content recommendation that the type of information node etc. produce focuses on, because weighted value obtain manner can produce association between information node by the click number of hops of user, achieve and recommend related information by large data thought to user, in addition, for the second-level message node participating in weighted value calculating, in information node recommendation process after this, without the need to again participating in the calculating of weighted value, only need the weighted value calculating the information node newly increased, therefore improve the real-time that information node is recommended, substantially reduce computation period.
Above-mentionedly be specially video film for information node and carry out exemplary illustration, those skilled in the art it will also be appreciated that, information node can also be keyword, such as user inputs " Liu Dehua " in a search engine, then according to the invention described above embodiment, can also recommend to user " schoolmate, Guo Fucheng " etc.
Embodiment three:
The schematic flow sheet of the recommend method of the related information that Fig. 4 provides for the embodiment of the present invention three, Fig. 5 is the schematic diagram of related information digraph embodiment illustrated in fig. 4; The embodiment of the present invention is less than the information node recommended to user number n for the number m of second-level message node carries out exemplary illustration, and as shown in Figure 4, the embodiment of the present invention comprises the steps:
Step 301, obtains the click number of hops of user at first order information node, wherein, clicks number of hops and comprises the first number of hops and the second number of hops.
Step 302, obtains m the first weighted value of m the second-level message node be associated with first order information node according to the first number of hops and the second number of hops.
Step 303, obtains r the third level information node be associated with m second-level message node, and calculates r the second weighted value between r third level information node and m second-level message node, and r is positive integer;
Step 304, obtains the multiple recommendations between m second-level message node and each self-corresponding multiple third level information node according to m the first weighted value and individual second weighted value of r;
Step 305, is defined as n-m third level information node of user's recommendation according to multiple recommendation.
Step 306, recommends user by m second-level message node and n-m third level information node.
Be described below in conjunction with Fig. 5, wherein, first order information node is A information node, second-level message node is B1, B2 information node, relative to the third level information node C1, C2, C3, D1, D2 of A information node, further, C1, C2, C3 information node is the second-level message node of B1 information node, and D1, D2 information node is the second-level message node of B2 information node.In embodiments of the present invention, be specially video film for A information node, C1, C2, C3, D1, D2 information node and carry out exemplary illustration.
As shown in Figure 3, have multiple video film such as A, B1, B2, C1, C2, C3, D1, D2, associated between A, B1, B2, C1, C2, C3, D1, D2 by oriented arrow, incidence relation each other can be represented by weight, weight is larger, shows that correlation degree is stronger.
The specific implementation process of step 301 and step 302 with reference to the description of the step 201 in above-described embodiment two and step 202, can not repeat them here.And calculated by this step 2 the first weighted values are respectively: weight
1(A, B1), weight
1(A, B2).
In step 303, as shown in Figure 5, need to recommend 3 videos to user from 5 videos, now, m=2, n=5, m<n, second-level message Node B 1, B2 meets recommended requirements not, now, B1 is being recommended to user, on the basis of B2 information node, also need traverse node B1, the associated nodes of B2 (as shown in Figure 5, r=5), therefrom select the information node of high weight, namely from the third level information node C1 of second-level message Node B 1, C2, the third level information node D1 of C3 and second-level message Node B 2, D2 recommends the information node of 3 high weights again to user.
Calculate 5 the second weighted values between 5 third level information node C1, C2, C3, D1, D2 and B1, B2 information nodes, the calculating of the second weighted value see the computing method of the weighted value in above-described embodiment two, can not repeat them here.5 the second weighted values calculated are respectively weight
2(B1, C1), weight
2(B1, C2), weight
2(B1, C3), weight
2(B2, D1), weight
2(B2, D2).
In step 304, multiple recommendation can be calculated by following equation to comprise:
Rec
value=δ * weight
1* weight
2, wherein, δ is decay factor, weight
1represent the first weighted value of first order information node and second-level message node, weight
2represent the second weighted value of second-level message node and third level information node.
Specifically, in embodiments of the present invention, 5 recommendations are specific as follows:
rec
value1＝δ*weight
1(A,B1)*weight
2(B1,C1)，
rec
value2＝δ*weight
1(A,B1)*weight
2(B1,C2)，
rec
value3＝δ*weight
1(A,B1)*weight
2(B1,C3)，
rec
value4＝δ*weight
1(A,B2)*weight
2(B2,D1)，
rec
value5＝δ*weight
1(A,B2)*weight
2(B2,D2)。
Wherein, δ is decay factor.
In step 305 and step 306, can right to choose weight values is larger from multiple recommendation information node, such as, in above-mentioned 5 recommendations, if recommendation rec
value5, rec
value3, rec
value2before coming, then corresponding for recommendation information node D2, C3, C2 are recommended user together with information node B1, B2.
When user jumps to the recommended third level information node D2, C3, C2 by the click of first order information node, information node A is shown as and information node D2, C3, C2 set up directed edge in the digraph shown in Fig. 5, and using the above-mentioned recommendation calculated as the initial weight on directed edge.
In another embodiment, if the number n recommended is 10, now above-mentioned second-level message node is 2, third level information node is 5, whole information nodes is also not enough to recommend 10 information nodes to user, in the case, remaining difference (10-2-5=3) adopts random mode of recommending to produce recommended node.
The recommend method of the related information that the embodiment of the present invention provides, recommended the information node be associated with first order information node to user in the click number of hops of first order information node by user, owing to being associated with the behavior of user when clicking and jump number of times, therefore the embodiment of the present invention avoids the content by information node, the uploader of information node or author, the problem that the content recommendation that the type of information node etc. produce focuses on, because weighted value obtain manner can produce association between information node by the click number of hops of user, achieve and recommend related information by large data thought to user, in addition, for the second-level message node participating in weighted value calculating, in information node recommendation process after this, without the need to again participating in the calculating of weighted value, only need the weighted value calculating the information node newly increased, therefore improve the real-time that information node is recommended, substantially reduce computation period.
On the basis of above-described embodiment two and embodiment three, the embodiment of the present invention also comprises the steps:
First, first order information node and the title of m second-level message node and the similarity of description is calculated; Secondly, according to the first initial weighted value between similarity determination first order information node and m second-level message node.
In one embodiment, in the digraph represented by above-mentioned Fig. 3 or Fig. 5, can also carry out Modling model by the mode of graph model to the relation between recommended information node, the weight on each limit in digraph can be calculated by the click behavior of user.On this basis, by scheming upper random walk (random walk) method, and then adjust the weight on each limit, and get the second-level message node be more associated with first order information node.
In one embodiment, when click behavior due to the click behavior or user that do not have user is less, first can produce initial digraph by the similarity of the title between information node and description, thus, the weight on the limit in initial digraph is calculated by the relevance degree of video.Because the quantity of video on network is large, therefore can not calculate by the calculating between any two videos the weight getting initial limit.In one embodiment, simhash method can be passed through, first find out each information node (such as, video) candidate association information node, again by the similarity calculating method based on term vector, the title (title) calculating two information nodes and the similarity described, thus obtain the weight on the limit of initial digraph.
In another embodiment, if information node be an isolated node point (namely, not with other any information node incidence relation), due in the digraph shown in Fig. 3 or Fig. 5, isolated point does not have out-degree, therefore cannot do correlation recommendation for isolated node, therefore, in one embodiment, by the random mode producing video, isolated node can be activated.If the random video produced, before next time is random, redirect is clicked by user, then because the weight that be calculated video association by the embodiment of the present invention is understood in the click redirect behavior of user, if the random video produced all was not clicked by user, so when next round upgrades, for this isolated node produces the recommendation video of association again at random.
Above-mentionedly be specially video film for information node and carry out exemplary illustration, those skilled in the art it will also be appreciated that, information node can also be keyword, such as user inputs " Liu Dehua " in a search engine, then according to the invention described above embodiment, can also recommend to user " schoolmate, Guo Fucheng " etc.
Embodiment four:
The structural representation of the recommendation apparatus of the related information that Fig. 6 provides for the embodiment of the present invention four; As shown in Figure 6, the embodiment of the present invention comprises:
First acquisition module 41, for m the first weighted value that m the second-level message node obtained with first order information node is associated is corresponding, wherein, described first order information node is the information node of the current use of user, and m is positive integer;
Recommending module 42, for being the related information that described user determines to recommend according to described m the first weighted value and n default recommendation information node, wherein, n is positive integer.
Detailed description in the embodiment of the present invention and Advantageous Effects refer to above-described embodiment one, do not repeat them here.
Embodiment five:
The structural representation of the recommendation apparatus of the related information that Fig. 7 provides for the embodiment of the present invention five; As shown in Figure 7, on the basis of above-described embodiment four, the first acquisition module 41 in the embodiment of the present invention comprises:
First acquiring unit 411, for obtaining the click number of hops of user at described first order information node, described click number of hops comprises the first number of hops and the second number of hops;
Second acquisition unit 412, for obtaining m the first weighted value of m the second-level message node be associated with described first order information node according to described first number of hops and described second number of hops.
In one embodiment, second acquisition unit 412 calculates described m the first weighted value by following equation and comprises:
wherein, N (a, b) the first number of hops from the information node B described first order information node A to a described m second-level message node is represented, N (a, *) represent from the first number of hops of described first order information node A to a described m second-level message node with, M (a, b) the second number of hops from described first order information node A to described information node B is represented, M (a, *) represent from the second number of hops of described first order information node A to a described m second-level message node with, α and β is weight proportion.
In one embodiment, if described m is greater than or equal to described n, recommending module 42 comprises:
First sequencing unit 421, for being sorted by described m the first weighted value, obtains ranking results;
First recommendation unit 422, for recommending n related information node for described user according to described ranking results from described m second-level message node.
In another embodiment, if described m is less than described n, recommending module 42 comprises:
3rd acquiring unit 423, for obtaining and described m the r that second-level message node is associated third level information node, r is positive integer;
4th acquiring unit 424, for obtaining the multiple recommendations between described m second-level message node and each self-corresponding multiple third level information node according to described m the first weighted value;
Second recommendation unit 425, for being defined as n-m the information node that described user recommends according to described multiple recommendation.
In one embodiment, the 4th acquiring unit 424 is comprised by the described multiple recommendation of following equation calculating:
rec
value＝δ*weight
1*weight
2，
Wherein, δ is decay factor, weight
1represent the first weighted value of described first order information node and described second-level message node, weight
2represent the second weighted value of described second-level message node and described third level information node.
In one embodiment, described device also comprises:
Computing module 43, for calculating described first order information node and described m the title of second-level message node and the similarity of description;
Second acquisition module 44, for determining the first initial weighted value between described first order information node and described m second-level message node according to described similarity.
Detailed description in the embodiment of the present invention and Advantageous Effects refer to above-described embodiment two and embodiment three, do not repeat them here.
The above; be only the specific embodiment of the present invention, but protection scope of the present invention is not limited thereto, is anyly familiar with those skilled in the art in the technical scope that the present invention discloses; change can be expected easily or replace, all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection domain of described claim.
Claims (14)
1. a recommend method for related information, is characterized in that, described method comprises:
Obtain the m corresponding with m the second-level message node that first order information node is associated first weighted value, wherein, described m the first weighted value is got by the click redirect behavior of user, and described first order information node is the information node of the current use of described user, and m is positive integer;
Be the related information that described user determines to recommend according to described m the first weighted value and n the recommendation information node preset, wherein, n is positive integer.
2. method according to claim 1, is characterized in that, the step of m the first weighted value that described acquisition is corresponding with m the second-level message node that first order information node is associated comprises:
Obtain the click number of hops of user at described first order information node, described click number of hops comprises the first number of hops and the second number of hops;
M the first weighted value of m the second-level message node be associated with described first order information node is obtained according to described first number of hops and described second number of hops.
3. method according to claim 2, it is characterized in that, in the step of described m the first weighted value obtaining m the second-level message node be associated with described first order information node according to described first number of hops and described second number of hops, calculate described m the first weighted value by following equation:
4. method according to claim 1, is characterized in that, if described m is greater than or equal to described n, described is that described user determines that the step of the related information recommended comprises according to described m the first weighted value and n default recommendation information node:
Described m the first weighted value is sorted, obtains ranking results;
From described m second-level message node, n related information node is recommended for described user according to described ranking results.
5. method according to claim 1, is characterized in that, if described m is less than described n, described is that described user determines that the step of the related information recommended comprises according to described m the first weighted value and n default recommendation information node:
Obtain and described m the r that second-level message node is associated third level information node, and calculate r the second weighted value between described r third level information node and described m second-level message node, r is positive integer;
The multiple recommendations between described m second-level message node and each self-corresponding multiple third level information node are obtained according to described m the first weighted value and individual second weighted value of described r;
N-m information node of described user recommendation is defined as according to described multiple recommendation.
6. method according to claim 5, it is characterized in that, describedly obtain in the step of the multiple recommendations between described m second-level message node and each self-corresponding third level information node according to described m the first weighted value, by the described multiple recommendation of following equation calculating:
rec
value＝δ*weight
1*weight
2，
Wherein, δ is decay factor, weight
1represent the first weighted value of described first order information node and described second-level message node, weight
2represent the second weighted value of described second-level message node and described third level information node.
7., according to the arbitrary described method of claim 1-6, it is characterized in that, described method also comprises:
Calculate described first order information node and described m the title of second-level message node and the similarity of description;
The the first initial weighted value between described first order information node and described m second-level message node is determined according to described similarity.
8. a recommendation apparatus for related information, is characterized in that, described device comprises:
First acquisition module, for m the first weighted value that m the second-level message node obtained with first order information node is associated is corresponding, wherein, described m the first weighted value is got by the click redirect behavior of user, described first order information node is the information node of the current use of described user, and m is positive integer;
Recommending module, for being the related information that described user determines to recommend according to described m the first weighted value and n default recommendation information node, wherein, n is positive integer.
9. device according to claim 8, is characterized in that, described first acquisition module comprises:
First acquiring unit, for obtaining the click number of hops of user at described first order information node, described click number of hops comprises the first number of hops and the second number of hops;
Second acquisition unit, for obtaining m the first weighted value of m the second-level message node be associated with described first order information node according to described first number of hops and described second number of hops.
10. device according to claim 9, is characterized in that, described second acquisition unit calculates described m the first weighted value by following equation:
11. devices according to claim 9, is characterized in that, if described m is greater than or equal to described n, described recommending module comprises:
First sequencing unit, for being sorted by described m the first weighted value, obtains ranking results;
First recommendation unit, for recommending n related information node for described user according to described ranking results from described m second-level message node.
12. devices according to claim 9, is characterized in that, if described m is less than described n, described recommending module comprises:
3rd acquiring unit, for obtaining and described m the r that second-level message node is associated third level information node, and calculate r the second weighted value between described r third level information node and described m second-level message node, r is positive integer;
4th acquiring unit, for obtaining the multiple recommendations between described m second-level message node and each self-corresponding multiple third level information node according to described m the first weighted value and individual second weighted value of described r;
Second recommendation unit, for being defined as n-m the information node that described user recommends according to described multiple recommendation.
13. devices according to claim 12, is characterized in that, described 4th acquiring unit calculates described multiple recommendation by following equation:
rec
value＝δ*weight
1*weight
2，
Wherein, δ is decay factor, weight
1represent the first weighted value of described first order information node and described second-level message node, weight
2represent the second weighted value of described second-level message node and described third level information node.
14. according to the arbitrary described device of claim 9-13, and it is characterized in that, described device also comprises:
Computing module, for calculating described first order information node and described m the title of second-level message node and the similarity of description;
Second acquisition module, for determining the first initial weighted value between described first order information node and described m second-level message node according to described similarity.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410610726.7A CN104361062B (en) | 2014-11-03 | 2014-11-03 | A kind of recommendation method and device of related information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410610726.7A CN104361062B (en) | 2014-11-03 | 2014-11-03 | A kind of recommendation method and device of related information |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104361062A true CN104361062A (en) | 2015-02-18 |
| CN104361062B CN104361062B (en) | 2017-10-31 |
Family
ID=52528323
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410610726.7A Active CN104361062B (en) | 2014-11-03 | 2014-11-03 | A kind of recommendation method and device of related information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104361062B (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105488154A (en) * | 2015-11-28 | 2016-04-13 | 小米科技有限责任公司 | Theme application recommendation method and device |
| CN105630766A (en) * | 2015-12-22 | 2016-06-01 | 北京奇虎科技有限公司 | Multi-news correlation calculation method apparatus |
| CN105867897A (en) * | 2015-12-07 | 2016-08-17 | 乐视网信息技术（北京）股份有限公司 | Page redirection analysis method and apparatus |
| CN105989154A (en) * | 2015-03-03 | 2016-10-05 | 华为技术有限公司 | Similarity measurement method and equipment |
| CN106686063A (en) * | 2016-12-07 | 2017-05-17 | 乐视控股（北京）有限公司 | Information recommendation method and apparatus, and electronic device |
| CN106919582A (en) * | 2015-12-24 | 2017-07-04 | 阿里巴巴集团控股有限公司 | The association of network articles and related information statistical method and device |
| CN108737902A (en) * | 2018-05-22 | 2018-11-02 | 北京奇艺世纪科技有限公司 | A kind of update method of associated video, device and electronic equipment |
| US10217025B2 (en) | 2015-12-22 | 2019-02-26 | Beijing Qihoo Technology Company Limited | Method and apparatus for determining relevance between news and for calculating relevance among multiple pieces of news |
| CN117058061A (en) * | 2023-10-12 | 2023-11-14 | 广东工业大学 | Multi-focus image fusion method and related device based on target detection |
| CN117194459A (en) * | 2023-09-22 | 2023-12-08 | 天翼爱音乐文化科技有限公司 | Operation and maintenance knowledge base updating method, system, device and medium based on operation and maintenance event |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101079062A (en) * | 2007-06-22 | 2007-11-28 | 腾讯科技（深圳）有限公司 | Web page importance evaluation method and system |
| CN101520878A (en) * | 2009-04-03 | 2009-09-02 | 华为技术有限公司 | Method, device and system for pushing advertisements to users |
| CN102054004A (en) * | 2009-11-04 | 2011-05-11 | 清华大学 | Webpage recommendation method and device adopting same |
| CN103530339A (en) * | 2013-10-08 | 2014-01-22 | 北京百度网讯科技有限公司 | Mobile application information push method and device |
| CN104123321A (en) * | 2013-04-28 | 2014-10-29 | 百度在线网络技术（北京）有限公司 | Method and device for determining recommended pictures |
-
2014
- 2014-11-03 CN CN201410610726.7A patent/CN104361062B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101079062A (en) * | 2007-06-22 | 2007-11-28 | 腾讯科技（深圳）有限公司 | Web page importance evaluation method and system |
| CN101520878A (en) * | 2009-04-03 | 2009-09-02 | 华为技术有限公司 | Method, device and system for pushing advertisements to users |
| CN102054004A (en) * | 2009-11-04 | 2011-05-11 | 清华大学 | Webpage recommendation method and device adopting same |
| CN104123321A (en) * | 2013-04-28 | 2014-10-29 | 百度在线网络技术（北京）有限公司 | Method and device for determining recommended pictures |
| CN103530339A (en) * | 2013-10-08 | 2014-01-22 | 北京百度网讯科技有限公司 | Mobile application information push method and device |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10579703B2 (en) | 2015-03-03 | 2020-03-03 | Huawei Technologies Co., Ltd. | Similarity measurement method and device |
| CN105989154A (en) * | 2015-03-03 | 2016-10-05 | 华为技术有限公司 | Similarity measurement method and equipment |
| CN105989154B (en) * | 2015-03-03 | 2020-07-14 | 华为技术有限公司 | Similarity measurement method and equipment |
| CN105488154A (en) * | 2015-11-28 | 2016-04-13 | 小米科技有限责任公司 | Theme application recommendation method and device |
| CN105867897A (en) * | 2015-12-07 | 2016-08-17 | 乐视网信息技术（北京）股份有限公司 | Page redirection analysis method and apparatus |
| US10217025B2 (en) | 2015-12-22 | 2019-02-26 | Beijing Qihoo Technology Company Limited | Method and apparatus for determining relevance between news and for calculating relevance among multiple pieces of news |
| CN105630766A (en) * | 2015-12-22 | 2016-06-01 | 北京奇虎科技有限公司 | Multi-news correlation calculation method apparatus |
| CN105630766B (en) * | 2015-12-22 | 2018-11-06 | 北京奇虎科技有限公司 | Correlation calculations method and apparatus between more news |
| CN106919582A (en) * | 2015-12-24 | 2017-07-04 | 阿里巴巴集团控股有限公司 | The association of network articles and related information statistical method and device |
| CN106686063A (en) * | 2016-12-07 | 2017-05-17 | 乐视控股（北京）有限公司 | Information recommendation method and apparatus, and electronic device |
| CN108737902A (en) * | 2018-05-22 | 2018-11-02 | 北京奇艺世纪科技有限公司 | A kind of update method of associated video, device and electronic equipment |
| CN108737902B (en) * | 2018-05-22 | 2021-03-02 | 北京奇艺世纪科技有限公司 | Associated video updating method and device and electronic equipment |
| CN117194459A (en) * | 2023-09-22 | 2023-12-08 | 天翼爱音乐文化科技有限公司 | Operation and maintenance knowledge base updating method, system, device and medium based on operation and maintenance event |
| CN117058061A (en) * | 2023-10-12 | 2023-11-14 | 广东工业大学 | Multi-focus image fusion method and related device based on target detection |
| CN117058061B (en) * | 2023-10-12 | 2024-01-30 | 广东工业大学 | Multi-focus image fusion method and related device based on target detection |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104361062B (en) | 2017-10-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104361062A (en) | Associated information recommendation method and device | |
| CN103514304B (en) | Project recommendation method and device | |
| CN103209342B (en) | A kind of introduce video popularity and the collaborative filtered recommendation method of user's interests change | |
| CN104331459B (en) | A kind of network resource recommended method and device based on on-line study | |
| CN107451894B (en) | Data processing method, device and computer readable storage medium | |
| Chang et al. | Using groups of items for preference elicitation in recommender systems | |
| CN102495864A (en) | Collaborative filtering recommending method and system based on grading | |
| CN105005582A (en) | Recommendation method and device for multimedia information | |
| WO2018121700A1 (en) | Method and device for recommending application information based on installed application, terminal device, and storage medium | |
| Guo et al. | A social influence approach for group user modeling in group recommendation systems | |
| EP2860672A2 (en) | Scalable cross domain recommendation system | |
| CN103744966A (en) | Item recommendation method and device | |
| CN103886090A (en) | Content recommendation method and device based on user favorites | |
| CN105095256A (en) | Information push method and apparatus based on similarity degree between users | |
| CN104199896A (en) | Video similarity determining method and video recommendation method based on feature classification | |
| US9881091B2 (en) | Content item audience selection | |
| CN110209946B (en) | Social and community-based product recommendation method, system and storage medium | |
| CN108446297B (en) | Recommendation method and device and electronic equipment | |
| CN109800328B (en) | Video recommendation method, device thereof, information processing equipment and storage medium | |
| Rokach et al. | Initial profile generation in recommender systems using pairwise comparison | |
| CN110046278B (en) | Video classification method and device, terminal equipment and storage medium | |
| Nargundkar et al. | InfluenceRank: A machine learning approach to measure influence of Twitter users | |
| KR20170120661A (en) | Algorithmic identification of appropriate content for children without human intervention | |
| CN110727872A (en) | Method and device for mining ambiguous selection behavior based on implicit feedback | |
| US20140258400A1 (en) | Content item audience selection |
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 |