CN106503025A - Method and system is recommended in a kind of application - Google Patents
Method and system is recommended in a kind of application Download PDFInfo
- Publication number
- CN106503025A CN106503025A CN201510568010.XA CN201510568010A CN106503025A CN 106503025 A CN106503025 A CN 106503025A CN 201510568010 A CN201510568010 A CN 201510568010A CN 106503025 A CN106503025 A CN 106503025A
- Authority
- CN
- China
- Prior art keywords
- application
- eigenvalue
- app
- dimension
- applications
- 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
Abstract
A kind of application recommendation method and a kind of application commending system is embodiments provided, is related to field of computer technology.Methods described, including：Obtain the user behavior record related to application and each application is analyzed, obtain the first application and alternatively gather；According to the alternative set of the described first application, obtain and recommend set of applications；The each application that recommends in set of applications is shown.The embodiment of the present invention combines the access behavior of user, and relatively objective statistics obtains the recommendation list of the APP of more high-quality, and reduces human cost, improves recommendation efficiency.
Description
Technical field
The present invention relates to field of computer technology, more particularly to a kind of application recommendation method and a kind of application
Commending system.
Background technology
Mobile terminal with mobile phone, panel computer as representative develops rapidly in recent years, CPU and storage device
Greatly improve in performance, corresponding, the application on mobile terminal is also increasingly enriched, use
Family interface is also more and more friendly.
A lot of functions of mobile terminal, the application being required for above mobile terminal (APP) have been come
Into.So how to find suitable software, find game interested, be that mobile phone users are being used
Problems faced when mobile terminal.Thus, various application management platforms arise at the historic moment, such as：Search dog
Mobile phone assistant, 360 mobile phone assistant, 91 mobile phone assistants, pea pods mobile phone assistant, Semen setariae application market,
Fructus Mali pumilae APP STORE etc..Above-mentioned application management platform is intended to help mobile phone users more easily
It was found that, find, download, install, update APP.These application management platforms both may be mounted at PC
Or on MAC, it is also possible to it is mounted directly on mobile terminals.Application management platform after start-up can
Show the APP for organizing to user, and in application management platform, there are a kind of two grades of APP and recommend
The page, to show to user or recommend new restocking or high-quality APP, the APP recommends the exhibition of the page
Show the displaying order between content, APP, can directly affects user needed for the page quickly finds
High-quality APP, further have impact on user using application management platform direct experience and APP
Download, streamflow and commercial value on application management platform.
But, in first technology, APP recommends the APP shown in the page and its order, mainly by people
Work editor, examination ＆ verification, sequence.Such as recommend in page, to show 50 high-quality APP, these APP in APP
Need to be screened inside 300,000 APP from APP storehouses one by one；And 50 high-quality APP will
It is ranked up according to quality order from high to low.If being done by manually completely, need daily to consume greatly
The human resourcess of amount, efficiency are very low.And human-edited, examination ＆ verification, the mode of sequence, subjectivity are strong,
Cannot objectively ensure that the APP for recommending is high-quality APP that user is inclined to use.
Content of the invention
In view of the above problems, it is proposed that the embodiment of the present invention overcomes the problems referred to above or extremely to provide one kind
A kind of application recommendation method for partially solving the above problems and a kind of corresponding application commending system.
In order to solve the above problems, the invention discloses a kind of application recommendation method, including：
Obtain the user behavior record related to application and each application is analyzed, obtain the first application standby
Selected works are closed；
According to the alternative set of the described first application, obtain and recommend set of applications；
The each application that recommends in set of applications is shown.
Preferably, the acquisition user behavior related to application records and each application is analyzed, and obtains
The step of the first application is alternative to gather is obtained, including：
For each application, from the user behavior record related to the application, extract at least one and tie up
The eigenvalue of degree；The eigenvalue of at least one dimension includes：The eigenvalue of the first value relevance, use
Family accesses the eigenvalue of behavior dimension；
According to the eigenvalue for extracting at least one dimension and the weight of respective dimensions, the application is calculated
Total characteristic value；
By N forward for the sequence of total characteristic value1Individual application adds the alternative set of the first application.
Preferably, apply for each, in recording from related to the application user behavior, extract to
The step of eigenvalue of a few dimension, includes：
For each application, calculate between the first value data of the application and the dispensed amount of the application
Ratio, using the ratio as the first value relevance eigenvalue.
Preferably, the eigenvalue of the user access activity dimension includes：Browse dimension eigenvalue, under
Carry the eigenvalue of dimension.
Preferably, when the eigenvalue of the user access activity dimension is the eigenvalue for browsing dimension, institute
State for each application, from the user behavior record related to the application, extract at least one dimension
Eigenvalue the step of include：
For each application, count described and apply browsed in application summary displayed page first to browse
Number of times, the second number of visits browsed in result of page searching, and in application details displayed page
The 3rd number of visits for being browsed；
By first number of visits, the second number of visits and the 3rd number of visits respectively in connection with the place page
Weight, calculate and browse the eigenvalue of dimension.
Preferably, when the eigenvalue of the user access activity dimension is the eigenvalue for downloading dimension, institute
State for each application, from the user behavior record related to the application, extract at least one dimension
Eigenvalue the step of include：
For each application, first download time that applies in application summary displayed page is counted,
The second download time being downloaded in result of page searching, and in being downloaded in application details displayed page
The 3rd download time being downloaded；
By first download time, the second download time and the 3rd download time respectively in connection with the place page
Weight, calculate download dimension eigenvalue.
Preferably, before the step of according to the alternative set of the described first application, obtaining recommendation set of applications,
Also include：
The application for obtaining at least one application management platform is recommended the application in the page and each application is carried out
Analysis, obtains the alternative set of the second application；
According to the alternative set of the described first application, obtaining the step of recommending set of applications is：
Alternative for described first application set and the alternative set of the second application are merged, is obtained and is recommended application
Set.
Preferably, the application right in the page is recommended in the application for obtaining at least one application management platform
Each application is analyzed, and obtaining includes the step of the second application is alternative to gather：
For the page is recommended in the application of at least one application management platform, corresponding application Sorted list is obtained
Table；
When it is two or more to apply sorted lists, each application sorted lists is carried out fusion meter
Calculate, obtaining includes N2Second application of individual application is alternative to gather, wherein N2It is less than or equal to all
The summation of number of applications in list of application.
Preferably, described when it is two or more to apply sorted lists, by each application Sorted list
Table carries out fusion calculation, and obtaining includes N2The step of second application of individual application is alternatively gathered includes：
From each application sorted lists, P forward application of selected and sorted；
For each application, from the user behavior record related to the application, extract at least one and tie up
The eigenvalue of degree；The eigenvalue of at least one dimension includes：The eigenvalue of the first value relevance, use
Family accesses the eigenvalue of behavior dimension；
According to eigenvalue and the weight of respective dimensions of each dimension, the total characteristic value of the application is calculated；
By N forward for the sequence of total characteristic value2Individual application adds the alternative set of the second application.
Preferably, described alternative for described first application set and the alternative set of the second application are merged,
Obtaining the step of recommending set of applications includes：
The front M of set of applications will be recommended1Individual position, forward from selected and sorted in the alternative set of the second application
M1Individual application is filled；
The remaining M of set of applications will be recommended2Individual position, selects sequence forward from the alternative set of the first application
Corresponding M2Individual application is filled；Wherein M1With M2Sum is to recommend application in set of applications
Quantity.
Preferably, after recommendation set of applications is obtained, also include：
For each application that recommends in set of applications, the application is stamped point according to tag along sort model
Class label；And/or
The usage behavior of each function of applying is analyzed according to each user, builds application class label mould
Type.
Present invention also offers a kind of application commending system, including：
The alternative set generation module of first application, records and right for obtaining the user behavior related to application
Each application is analyzed, and obtains the first application and alternatively gathers；
Recommend set of applications acquisition module, for according to the alternative set of the described first application, acquisition is recommended should
With set；
Display module, for each application that recommends in set of applications to be shown.
The embodiment of the present invention includes advantages below：
The embodiment of the present invention passes through dividing for user behavior record on a large scale for APP displayed pages
Analysis, the quality of overall merit APP are alternatively gathered so as to build the first application.Then should from first
Applied according to the demand screening for recommending set of applications with alternatively set, satisfactory application is added
Enter and recommend set of applications, then can recommend the application pipe that applies in the present invention in set of applications
The application of platform is recommended to be shown in the page.The embodiment of the present invention combines the access row of user
Obtain the recommendation list of the APP of more high-quality for, relatively objective statistics, and reduce manpower into
This, improves recommendation efficiency.
Description of the drawings
Fig. 1 is the step of embodiment of the method is recommended in a kind of application of present invention flow chart；
Figure 1A, 1B, 1C are that a kind of APP of the invention shows example；
Fig. 2 is the step of embodiment of the method is recommended in another kind of application of present invention flow chart；
Fig. 3 is the step of embodiment of the method is recommended in another kind of application of present invention flow chart；
Fig. 4 is a kind of structured flowchart of application commending system embodiment of the present invention；
Fig. 5 is the structured flowchart of another kind of application commending system embodiment of the present invention
Fig. 6 is the structural frames of another kind of application commending system embodiment of the present invention.
Specific embodiment
Understandable for enabling the above objects, features and advantages of the present invention to become apparent from, below in conjunction with the accompanying drawings
The present invention is further detailed explanation with specific embodiment.
One of the core concepts of the embodiments of the present invention is, by being directed to APP displayed pages on a large scale
User behavior record analysis, the quality of overall merit APP alternatively collects so as to build the first application
Close.Then can be from the alternative set of the first application according to the demand screening application for recommending set of applications, will
Satisfactory application is added recommends set of applications, then can will recommend applying at this in set of applications
The application of the application management platform of invention is recommended to be shown in the page.The embodiment of the present invention can be combined
The access behavior of user, such as browses, downloads the behavior of application in the dissimilar page, relatively objective
Statistics obtain more high-quality APP recommendation list, and reduce human cost, improve recommendation
Efficiency.
Embodiment one
With reference to Fig. 1, the step of embodiment of the method is recommended in a kind of application of present invention flow chart, tool is shown
Body may include steps of：
Step 110, obtains the user behavior record related to application and each application is analyzed, obtain
First application is alternative to gather.
In embodiments of the present invention, user behavior related to application on a large scale can be passed through to record, such as
Number of visits of the user to certain application place page, to download time that certain is applied etc..
It is appreciated that the embodiment of the present invention can obtain the user behavior from the server of each website
Record.In general, user behavior record is stored in the access log of server, the present invention is implemented
Example can get the user behavior record related to application by the access log.
User behavior record is analyzed, high-quality APP is obtained, then by these high-quality APP
Add the alternative set of the first application.In embodiments of the present invention, high-quality APP can be from multiple sides
Face is weighed, such as number of visits, the aspect such as download time, the income that brings to application management platform.On
State the higher expression APP more high-quality of several aspect quantity.
Apply in an application management platform in the embodiment of the present invention, and meeting in application management platform itself
There is substantial amounts of APP, then when producing recommendation set of applications, in needing to this application management platform
APP is recommended, thus, obtain the user behavior record related to application and each application is analyzed
When, it is to be analyzed for the application in the set of applications that specifies.The specified set of applications can include
In this application management platform all application, or including the certain applications in this application management platform,
The new application for adding this application management platform in such as a period of time recently.
Optionally, step 110 includes：
Sub-step A10, for each application, from the user behavior record related to the application, carries
Take the eigenvalue of at least one dimension；The eigenvalue of at least one dimension includes：First value relevance
Eigenvalue, the eigenvalue of user access activity dimension.
Wherein, the user access activity dimensional characteristics value includes：Browse the eigenvalue of dimension, download dimension
The eigenvalue of degree.The eigenvalue for browsing dimension can be understood as user in each webpage show certain
The tolerance of the pageview of APP.The eigenvalue for downloading dimension can be understood as user to certain APP
Download tolerance.
The eigenvalue of multiple dimensions can be extracted from recording with user behavior in embodiments of the present invention, than
Such as the eigenvalue of the first value relevance, the eigenvalue of dimension is browsed, download the eigenvalue of dimension.Described
The income that the eigenvalue of one value relevance can be brought by certain APP unit dispensed amount.Described browse
The eigenvalue of dimension can be understood as tolerance of the user to the pageview of an APP of displaying in each webpage.
The eigenvalue for downloading dimension can be understood as tolerance of the user to the download of an APP.The use
Family access behavior dimensional characteristics value can be understood as user access activity generation flow, such as pageview,
Download etc..
Optionally, sub-step A10 includes：
Sub-step A101, for each application, the first value data for calculating the application are answered with described
Ratio between dispensed amount, using the ratio as the first value relevance eigenvalue.
In embodiments of the present invention, described first the business flowing water that data can be APP is worth, namely should
First value data can be this application management platform by the distribution total revenue data that obtain of the APP.
That is this application management platform has distributed a large amount of APP, APP developer then these dispensed amounts to this application pipe
Platform is paid with certain price, and price of this payment is exactly the APP under the dispensed amount
Business flowing water total revenue in other words.It is possible to obtain the first value data of each APP, may be used also
To obtain dispensed amounts of each APP in this application management platform, such as APPI, the APPI
First value data can be designated as " APPITotal revenue ", its dispensed amount in application platform can be remembered
For " APPIDispensed amount ", then can calculate eigenvalue by formula (1)：
SCOREP(APPI)=APPITotal revenue/APPIDispensed amount (1)
Wherein SCOREP(APPI) it is certain APPIEigenvalue.
Formula (1) calculates APPIThe income brought of unit dispensed amount, i.e. this application management is flat
Platform often distributes an APPIThe income for obtaining.APPIThe income brought of unit dispensed amount be the bigger the better.
Certainly, in the embodiment of the present invention, can obtain in section at the same time, total receipts of each APP
Benefit, and the dispensed amount of each APP.Such as obtain in nearest two week, the total revenue of each APP,
And the dispensed amount of each APP.Such that it is able to more accurately extract the first value dimension of each APP
The eigenvalue of degree, more accurately to recommend APP.
Optionally, when the eigenvalue of the user access activity dimension is the eigenvalue for browsing dimension, son
Step A10 includes：
Sub-step A102, for each application, counts described and applies quilt in application summary displayed page
The first number of visits for browsing, the second number of visits browsed in result of page searching, and in application
The 3rd number of visits browsed in details displayed page.
In embodiments of the present invention, can for each application, according to described apply clear in all kinds of webpages
Look at number of times, calculate browse dimension browse eigenvalue.Above-mentioned all kinds of webpages include showing APP summaries letter
The APP schematic representations webpages of breath, may illustrate the summary info of multiple APP simultaneously in the page,
As shown in Figure 1A, which is that the part APP shown in the APP schematic representations pages shows example.Above-mentioned
All kinds of webpages include searching for certain APP in searched page for APP result of page searching, i.e. user
The page of acquisition, such as Figure 1B, its are that user searches for " PDF readers " in searched page and obtains
The few examples of APP search results pages, user can browse APP from the page, certainly, can also wrap
Include the search results pages of other forms.Above-mentioned all kinds of webpages may also include application details displayed page, that is, use
Family clicks the page for introducing application concrete condition, illustrates that user compares care, such as Fig. 1 C to the application,
It illustrates the example of certain application details displayed page.
Wherein, for the mobile terminals such as mobile phone, when search dog mobile phone assistant is opened on mobile phone, this should
With during management platform application, application management platform have " fine work ", " application ", " game ", " video ",
" management ", each channel such as " recommendation ", " classification ", " seniority among brothers and sisters ", " special topic ", " information ", in other words
Each page, click on wherein any one, can all jump to corresponding channel in other words in the page, each
Channel (page) is all an APP displayed page.Aforementioned in embodiments of the present invention " TAB page " can
To be interpreted as each channel above-mentioned of application management platform.
Certainly, the embodiment of the present invention extracts user behavior note in by the user access logses of server
During record, to user access each network address, can judge its type.In embodiments of the present invention, can be with
Network address for each website arranges the regular expression for judging website type in advance, each in the website to judge
Which kind of type individual webpage belongs to, so as to be counted.Wherein, for regular expression：Of the invention real
Apply in example, the searched page of such as various search engines, the network address that can be directed to the search engine page is just arranged
Then expression formula, is then the corresponding type of webpage of each regular expression, on matching regular expressions then
Confirm that network address belongs to the type of webpage, do not match, be not belonging to the type of webpage.Search page such as search dog
The regular expression of the network address in face for bSOSO.COM b, mate soso.com characters between two " b "
String, matches, then searched page of the network address for search dog search is described, then judges each net of corresponding website
Page belongs to result of page searching type.
Sub-step A103, by first number of visits, the second number of visits and the 3rd number of visits point
Not Jie He the place page weight, calculate and browse the eigenvalue of dimension.
In embodiments of the present invention, all types of pages can be respectively and gives different weights, such as be applied
The weight of the schematic representations page is Λ 1, and result of page searching is Λ 2, and application details displayed page is Λ 3.
So the present invention can count described and apply browsed in application summary displayed page first to browse
Number of times C1(APPI), the second number of visits C browsed in result of page searching2(APPI), and answering
With the 3rd number of visits C browsed in details displayed page3(APPI).
The eigenvalue for browsing dimension can be calculated by formula (2) so：
SCOREBROWSING(APPI)=Λ 1*C1(APPI)+Λ2*C2(APPI)+Λ3*C3(APPI) formula (2)
Wherein SCOREBROWSING(APPI) it is APPIThe eigenvalue for browsing dimension.
Optionally, when the user access activity dimensional characteristics value is the eigenvalue for downloading dimension, sub-step
Rapid A10 includes：
Sub-step A104, for each application, counts described and applies the of application summary displayed page
One download time, the second download time being downloaded in result of page searching, and show in application details
The 3rd download time being downloaded in being downloaded in the page.
In embodiments of the present invention, can apply for each, according to the download channel of the application,
Calculate download dimension browses eigenvalue.The download channel can include：By showing APP summaries
The channel of the APP schematic representations webpages of information, shows the channel that page is downloaded by Search Results, and logical
Cross the channel that application details show that page is downloaded.
In embodiments of the present invention, aforementioned APP schematic representations webpage and application details show page, search
As a result show page, can be the page in application management platform.Download channel so in other pages,
Such as other software download channel, such as current application management platform be search dog mobile phone assistant, Yong Hutong
The APP for crossing search dog input method downloads page-downloading APP channels then for downloading in other pages.
Above-mentioned all kinds of webpages include the APP schematic representations webpages for showing APP summary infos, in the page
The summary info of multiple APPs is illustrated simultaneously may, as shown in Figure 1A, which is APP schematic representations
The part APP shown in the page shows example.Above-mentioned all kinds of webpages are included for APP search results pages
The page that certain APP is obtained is searched in face, i.e. user in searched page, and such as Figure 1B, its are that user exists
The few examples of the APP search results pages that search " PDF readers " is obtained, user in searched page
APP can be browsed from the page, certainly, may also include the search results pages of other forms.Above-mentioned each
Class webpage may also include application details displayed page, i.e. user and click the page for introducing application concrete condition
Face, illustrates that user compares care to the application, and such as Fig. 1 C it illustrates certain application details and show page
The example in face.
Sub-step A105, by first download time, the second download time and the 3rd download time point
Not Jie He the place page weight, calculate download dimension eigenvalue.
In embodiments of the present invention, all types of pages can be respectively and gives different weights, such as be applied
The weight of the schematic representations page is λ 1, and the weight of result of page searching is λ 2, applies details displayed page
Weight be λ 3.
So the present invention can statistics application quilt in application summary displayed page and application details displayed page
The first download time D that downloads1(APPI), the second download time being downloaded in result of page searching
D2(APPI), and the 3rd download time being downloaded in being downloaded in application details displayed page
D3(APPI).
Formula (3) can so be passed through and calculate the eigenvalue for downloading dimension：
SCOREDOWNLOADING(APPI)=λ 1*D1(APPI)+λ2*D2(APPI)+λ3*D3(APPI)
... formula (3)
Wherein SCOREDOWNLOADING(APPI) it is APPIDownload dimension eigenvalue.
Certainly, the embodiment of the present invention can also obtain other download times for downloading channels, such as other
Management download time of the distribution platform to the APP, then arranges weight for the management distribution platform, will
The weight is multiplied with number of times, is subsequently adding above-mentioned formula (3).
Sub-step A11, according to eigenvalue and the weight of respective dimensions of each dimension, calculates the application
Total characteristic value.
So the present invention can be according to the eigenvalue of the first value relevance, the eigenvalue for browsing dimension, download
The eigenvalue of dimension, calculates the total characteristic value of each APP.
Such as aforementioned APPI, can be that each characteristic dimension arranges a weight in the embodiment of the present invention, than
Such as：The weight of the eigenvalue of the first value relevance is λP, browse dimension eigenvalue weight λL, download
Weight λ of the eigenvalue of dimensionD, then APPITotal characteristic value can be calculated by formula (4)：
SCORETOTAL(APPI)=λP*SCOREP(APPI)+λL*SCOREBROWSING(APPI)+
λD*SCOREDOWNLOADING(APPI) ... formula (4)
Certainly, the embodiment of the present invention can also first by SCOREBROWSINGAnd SCOREDOWNLOADINGAltogether
Calculate the First Eigenvalue SCORE that user integrally accesses behaviorUSERS(APPI), the First Eigenvalue can be with
Judge the APP whether required for most of users.Then with SCOREP(APPI) judge the APP
Whether there is larger commercial value.Both revest weight respectively, in conjunction with calculating total characteristic value
SCORETOTAL(APPI).
SCORE is calculated by formula (5)USERS(APPI)：
SCOREUSERS(APPI)=λL1*SCOREBROWSING(APPI)+λD1*SCOREDOWNLOADING(APPI)
... formula (5)
Wherein λL1For SCOREBROWSING(APPI) weight, wherein λD1For SCOREDOWNLOADING(APPI)
Weight.
Then SCORE is calculated by formula (6)TOTAL(APPI)：
SCORETOTAL(APPI)=λP1*SCOREP(APPI)+λU*SCOREUSERS(APPI)
... formula (6)
Wherein λP1For SCOREP(APPI) weight, wherein λUFor SCOREUSERS(APPI) power
Weight.
Sub-step A12, by N forward for the sequence of total characteristic value1Individual application adds the alternative set of the first application.
So, then the SCORE of each APP is calculatedTOTAL(APPI) after, you can sequence is leaned on
Front N1The first application of individual addition is alternative to gather.
Certainly, in the embodiment of the present invention, can in webpage is calculated all APP SCORETOTAL
(APPI) after, by the APP in the APP storehouses of application management platform, according to the SCORETOTAL(APPI)
It is ranked up, then the forward N of selected and sorted1Individual APP adds the alternative set of the first application.The sequence is leaned on
Front N1Individual APP is high-quality APP.Wherein N1It is the integer more than or equal to 0, its size value
Can determine according to the actual requirements.
Step 120, according to the alternative set of the described first application, obtains and recommends set of applications.
Then, according to recommendation set of applications demand L APP, the past N1Individual APP continues screening L
APP.Wherein, N1>L.
The screening is such as directly by N1Sort forward in individual APP L adds and recommends set of applications.
Step 130, each application that recommends in set of applications is shown.
Then, it is recommended that each application in set of applications, you can according to sequence, the relevant information of application is put
Entering application recommends the relevant position of the page to be shown, there is provided browse/download to user.
The embodiment of the present invention passes through the analysis for recording on a large scale for the user behavior of APP displayed pages,
The quality of overall merit APP, alternatively gathers so as to build the first application.Then can be from the first application
According to the demand screening application for recommending set of applications in alternative set, satisfactory application is added recommendation
Set of applications, then can by recommend set of applications in apply the present invention application management platform should
With recommend the page in be shown.The embodiment of the present invention combines the access behavior of user, such as in difference
The behavior of application is browsed, is downloaded in type page, and relatively objective statistics obtains the APP's of more high-quality
Recommendation list, and human cost is reduced, improve recommendation efficiency.
Embodiment two
With reference to Fig. 2, the step of embodiment of the method is recommended in a kind of application of present invention flow chart, tool is shown
Body may include steps of：
Step 210, obtains the user behavior record related to application and each application is analyzed, obtain
First application is alternative to gather.
Step 220, obtains the application that applies in the recommendation page of at least one application management platform and to each
Application is analyzed, and obtains the alternative set of the second application.
In actual applications, in network, the possibility of existing application management platform also has application to recommend the page,
The page is recommended in application in other words, and which possesses certain ginseng generally by technical staff human-edited
Examine value.So the present invention may specify that the page is recommended in the specified application of crawl other application management platform, i.e.,
The page is recommended in the application that crawl is specified, and parses application therein, and then each application is analyzed,
The second application can be obtained alternatively to gather.
Certainly, the embodiment of the present invention can obtain the application recommendation page of one or more application management platform
Application in face.
Optionally, step 220 includes：
Sub-step B10, for the page is recommended in the application of at least one application management platform, obtains corresponding
Application sorted lists.
Application can be specified to recommend the page and parsed using the crawl of webpage capture device in the present invention, be obtained
Must to should the page application sorted lists.
Sub-step B11, when it is two or more to apply sorted lists, by each application Sorted list
Table carries out fusion calculation, and obtaining includes N2Second application of individual application is alternative to gather, wherein N2Be less than
Or the summation equal to number of applications in all list of applications.
Due to because of there is multiple applications that specifies to recommend the page, then accordingly there are multiple application Sorted lists
Table, then multiple application sorted lists can be merged by the present invention, remove sequence APP rearward, obtain
Arriving includes N2Second application of individual high-quality APP is alternative to gather.
Optionally, sub-step B11 includes：
Sub-step B1101, from each application sorted lists, P forward application of selected and sorted.
Wherein, front P cumulative quantity of multiple application sorted lists is needed more than the alternative set of the second application
Quantity N that asks2Individual.Such as application sorted lists have 5, then the number of APP is 5*P.
Sub-step B1102, for each application, from the user behavior record related to the application,
Extract the eigenvalue of at least one dimension；The eigenvalue of at least one dimension includes：First is worth dimension
The eigenvalue of degree, user access activity dimensional characteristics value；
Sub-step B1103, according to eigenvalue and the weight of respective dimensions of each dimension, calculates the application
Total characteristic value.
In embodiments of the present invention, the implementation procedure of sub-step B1102 and sub-step B1103, with enforcement
Sub-step A10 referred in example one is similar with the implementation procedure of sub-step A11, and here is no longer gone to live in the household of one's in-laws on getting married and chatted.
By sub-step B1102 and B1103, its total characteristic value can be calculated to each APP.
Sub-step B1104, by N forward for the sequence of total characteristic value2Individual application adds the alternative collection of the second application
Close.
Then the N for from each APP, selecting the sequence of total characteristic value forward2Individual application adds the second application standby
Selected works are closed.
Optionally, the method for the fusion in embodiments of the present invention can adopt BORDA COUNT (ripples
Reach counting method) method merged, sub-step B11 may include：
Sub-step B1111 is in each application sorted lists, according to the application location, right
Each application marking.
The length of the results list of hypothesis crawl is N, and for the APP for making number one, which is in the list
In score be designated as N；For coming deputy APP, its in the list must be divided into N-1；……
The APP for rolling into last place must be divided into 1.For the APP in all lists, repeat above-mentioned marking
Process.
Sub-step B1112, for same application, applies described in each application sorted lists
Marking is added up, and obtains the PTS of the application；
Marking of each APP in each list is accumulated.
Sub-step B1113, PTS is sorted forward N2Individual application adds the alternative set of the second application.
APP is ranked up from high to low according to accumulation marking, then PTS can be sorted forward
N2Individual application adds the alternative set of the second application.Wherein N2It is the integer more than or equal to 0, its size
Can arrange according to the actual requirements.
Certainly, in the embodiment of the present invention, also other blending algorithms can be adopted, the embodiment of the present invention is not to which
It is any limitation as.
Step 230, alternative for the described first application set and the alternative set of the second application is merged, is obtained
Set of applications must be recommended.
After so obtaining that two applications are alternative and gathering, the present invention need from two applications are alternative gather in select
Select suitable APP to be put in recommendation set of applications, the recommendation set of applications can be put into application and recommend the page
It is shown.So the present invention can the alternative set of first application and the alternative set of the second application melted
Close, optimum multiple APP are added to recommendation set of applications.
Optionally, sub-step 230 includes：
Sub-step C11, will recommend the front M of set of applications1Individual position, from the alternative set of the second application
The forward M of selected and sorted1Individual application is filled.
Sub-step C12, will recommend the remaining M of set of applications2Position, from the alternative set of the first application
The corresponding M for selecting sequence forward2Individual application is filled；Wherein M1With M2Sum collects for recommending application
The quantity that applies in conjunction.
For M before recommendation set of applications1Individual position, the front M in being gathered by the second application is alternative1Individual APP
Filling；Recommend set of applications M1M after position2Individual position, in being gathered by the first application is alternative
APP sorts according to total characteristic value, is filled from high to low.This process is until fill up recommendation set of applications
Position till.Wherein M1、M2Value can set according to actual needs, typically 3 or 4,
It is usually no more than 10.
Step 240, each application that recommends in set of applications is shown.
The embodiment of the present invention passes through the analysis for recording on a large scale for the user behavior of APP displayed pages,
The quality of overall merit APP, alternatively gathers so as to build the first application.Then from specified application
Recommend in the page, the page is recommended in the application of existing application management platform in such as network, obtain above-mentioned each
Individual application is recommended the application in the page and each application is analyzed, and builds the alternative set of the second application, so
Alternative for the first application set and the alternative set fusion of the second application are obtained final recommendation application collection again afterwards
Close, applying in set of applications then can will be recommended to recommend to show in the page in application.The present invention is implemented
Example combines the access behavior of user, such as browses, downloads the behavior of application in the dissimilar page,
And the application recommending data in the page is recommended in existing specified application, relatively objective statistics obtains more excellent
The recommendation list of the APP of matter, and human cost is reduced, improve recommendation efficiency.
Embodiment three
With reference to Fig. 3, the step of embodiment of the method is recommended in a kind of application of present invention flow chart, tool is shown
Body may include steps of：
Step 310, obtains the user behavior record related to application and each application is analyzed, obtain
First application is alternative to gather.
Step 320, obtains the application that applies in the recommendation page of at least one application management platform and to each
Application is analyzed, and obtains the alternative set of the second application.
Step 330, alternative for the described first application set and the alternative set of the second application is merged, is obtained
Set of applications must be recommended.
Step 340, for each application that recommends in set of applications, according to tag along sort model to described
Tag along sort is stamped in application.
Optionally, can also include after step 330：
Step 300, is analyzed to the usage behavior of each function of applying according to each user, builds application
Tag along sort model.
Optionally, can also include simultaneously：
Step 340 and step 300.
In the embodiment of the present invention, the advance structure APP tag along sort models of meeting, the APP tag along sort models
The usage behavior of each function of applying is analyzed by each user, builds application class label model.
In actual applications, each APP generally has multiple subdivision functional modules, generally using option etc.
Form is shown that the such as APP of game class, in addition to game entry entrance, also has corresponding community respectively
The functional modules such as forum, attack strategy exchange.
So, for each APP, the embodiment of the present invention can detect that each has downloaded the terminal of the APP
In, then usage behavior of the user to each function of the APP build APP according to the usage behavior
Tag along sort model.
Such as, if detecting certain function of a large number of users using APP, it is considered that user is inclined
To with using APP the function, then can stamp tag along sort corresponding with the function for the APP.
Such as aforementioned APP " fast tooth " a large number of users that detects has downloaded the APP, and the user more than 70%
Interactive game module therein is used, then tag along sort " trip can be fitted according to above-mentioned user behavior
Play is popular hey ", and the tag along sort is corresponding with " fast tooth ".
In embodiments of the present invention, for " white collar clan ", " military fan ", " financing intelligent ", " knapsack
The tag along sort such as visitor ", " without classification ", can regard tag along sort as a classification, then pass through machine
Maximum entropy classifiers in study build APP tag along sort models, i.e. maximum entropy tag along sort model.
The ultimate principle of maximum entropy classifiers is under qualificationss, seeks to meet available sample set entropy maximum
A kind of probability distribution, the building process of its model approximately as：
1st, a series of classification are preset.Such as " white collar clan ", " military fan ", " financing intelligent ", " back of the body
The classification such as bag visitor ", " without classification ", each classification is exactly a tag along sort.
2nd, a series of APP is selected, the using for each functional module for gathering each each APP of terminal-pair is gone
For.
3rd, the usage behavior of each functional module is analyzed, the corresponding contingency tables of each above-mentioned each APP are determined
Sign, using the above-mentioned APP for stamping tag along sort as training set.
4th, the feature of each APP in training set, such as skill APP name, APP belonging to are extracted
Key word, the author of APP, the sectional drawing of APP in art classification, APP description informations etc..Wherein
Technique classification belonging to APP such as tool-class, system tool class, input method class etc..
5th, using the features described above of APP as input value, in being input into into maximum entropy tag along sort model, most
Big entropy tag along sort model obtains the first classification results after calculating.
6th, to train the tag along sort of each APP in set as sample value, by the first classification results with
The sample value is compared, if error is more than threshold value, the adjustment maximum entropy classification of combined training algorithm
The model parameter of label model.Circulation is trained, until between the first classification results and the sample value
Error be less than threshold value, then maximum entropy tag along sort model construction finish.
So, in step 340, for each application that recommends in set of applications, according to tag along sort
Model stamps tag along sort to the application to be included：
Recommend each APP in set for application, extract its feature.APP names as the aforementioned,
The key word in technique classification, APP description informations, the author of APP, APP belonging to APP cut
Figure etc..
Then the feature of each APP is input into aforementioned maximum entropy tag along sort model, obtains classification results,
Tag along sort is beaten to the APP according to classification results.
If classification results certain specific tag along sort corresponding, stamps the tag along sort for the APP.
If the classification results of classification results corresponding " without classification ", do not beat tag along sort for the APP.
Certainly, in the embodiment of the present invention, APP tag along sorts model can also pass through Naive Bayes Classification
The modes such as device, support vector machine classifier, deep neural network, random forest are constructed, and the present invention is not right
Which is any limitation as.
Certainly, in the embodiment of the present invention, it is also possible to directly gather using in the user terminal for each APP
Family is analyzed to the usage behavior of each functions of APP, determines the tag along sort of each APP.Such as, right
Downloaded by user in a large number within a period of time recently in APP " fast tooth ", although which is that a file is passed
The software of defeated class, but detect 70% user and used interactive game module therein, then basis is upper
State user behavior and fit " game is popular hey " tag along sort, and " fast tooth " is enclosed this point
Class label.
In embodiments of the present invention, after recommendation set of applications is obtained, in order to make user more quick
Solution APP is and for the reasons why giving one recommendation of user, can also be certain when actual displaying
A little APP have stamped some tag along sorts, such as：" white collar clan ", " military fan ", " financing intelligent ", " back of the body
Bag visitor ", etc..Above-mentioned tag along sort substantially characterizes the user crowd that the APP is suitable for, if just
Good active user is a certain class people therein, then can greatly promote clicking rate of the user to the APP.?
In the embodiment of the present invention, for recommend set of applications in each application, for recommend set of applications in
Each application, stamps tag along sort to the application.
Step 350, each application that recommends in set of applications is shown.
By abovementioned steps, when showing, you can showing that summary, download link of APP etc. are basic
Outside information, show its tag along sort, facilitate user quickly to understand APP, and enrich and push away to user
The content that recommends.
The embodiment of the present invention can pass through the user behavior on a large scale for APP displayed pages and record
Analysis, the quality of overall merit APP alternatively gathers so as to build the first application.Then can be with
Recommend in the page from specified application displayed page application, existing application management in such as network is put down
The page is recommended in the application displayed page application of platform, obtains each application displayed page application above-mentioned and recommends
Application in the page is simultaneously analyzed to each application, is built the alternative set of the second application, then again will
The alternative set of first application and the alternative set fusion of the second application, obtain final recommendation set of applications,
And for recommending the tagged tag along sort of the application in set of applications, set of applications then will be recommended
In apply and recommend to show in the page in application displayed page application.The embodiment of the present invention combines use
The access behavior at family, such as browses, downloads the behavior of application, Yi Jiyi in the dissimilar page
The application recommending data in the page, relatively objective system are recommended in the specified application displayed page application having
Meter obtains the recommendation list of the APP of more high-quality, and reduces human cost, improves recommendation effect
Rate, and labeling label is easy to user quickly to understand each function of applying and stressed, and enriches
The content of displaying.
It should be noted that for embodiment of the method, in order to be briefly described, therefore which is all expressed as one it is
The combination of actions of row, but those skilled in the art should know, and the embodiment of the present invention is not by described
Sequence of movement restriction because according to the embodiment of the present invention, some steps can adopt other orders or
Person is carried out simultaneously.Secondly, those skilled in the art should also know, embodiment described in this description
Preferred embodiment is belonged to, necessary to the involved action not necessarily embodiment of the present invention.
Example IV
With reference to Fig. 4, a kind of structured flowchart of application commending system embodiment of the present invention is shown, specifically
Can include such as lower module：
The alternative set generation module 410 of first application, for obtaining the user behavior record related to application
And each application is analyzed, obtain the alternative set of the first application.
Recommend set of applications acquisition module 420, for according to the alternative set of the described first application, being pushed away
Recommend set of applications.
Display module 430, for each application that recommends in set of applications to be shown.
Optionally, the alternative set generation module 410 of the first application includes：
First initial characteristicses word computing module, for for each application, from the use related to the application
In the behavior record of family, the eigenvalue of at least one dimension is extracted；The eigenvalue bag of at least one dimension
Include：The eigenvalue of the first value relevance, the eigenvalue of user access activity dimension.
First total characteristic value computing module, for the eigenvalue according to each dimension and the weight of respective dimensions,
Calculate the total characteristic value of the application.
First alternative set constructing module, for sorting forward N total characteristic value1Individual application adds the
One application is alternative to gather.
Optionally, the first initial characteristicses word computing module includes：
First value relevance characteristic value calculating module, for for each application, calculating the of the application
Ratio between the dispensed amount of monovalence Value Data and the application, using the ratio as the first value relevance
Eigenvalue.
Optionally, the user access activity dimensional characteristics value includes：Browse eigenvalue, the download of dimension
The eigenvalue of dimension.
Optionally, when the user access activity dimensional characteristics value is the eigenvalue for browsing dimension, first
Initial characteristicses word computing module includes：
For each application, count described and apply browsed in application summary displayed page first to browse
Number of times, the second number of visits browsed in result of page searching, and in application details displayed page
The 3rd number of visits for being browsed.
Characteristic value calculating module is browsed, for by first download time, the second download time and the 3rd
Weight of the download time respectively in connection with the place page, calculates the eigenvalue for downloading dimension.
Optionally, when the user access activity dimensional characteristics value is the eigenvalue for downloading dimension, first
Initial characteristicses word computing module includes：
Page-downloading number of times statistical module, for for each application, counting described application in application summary
First download time of displayed page, the second download time being downloaded in result of page searching, and
The 3rd download time being downloaded in being downloaded in application details displayed page.
Download features value computing module, for by first download time, the second download time and the 3rd
Weight of the download time respectively in connection with the place page, calculates the eigenvalue for downloading dimension.
Embodiment five
With reference to Fig. 5, a kind of structured flowchart of application commending system embodiment of the present invention is shown, specifically
Can include such as lower module：
The alternative set generation module 510 of first application, for obtaining the user behavior record related to application
And each application is analyzed, obtain the alternative set of the first application.
The alternative set generation module 520 of second application, for obtaining answering at least one application management platform
It is analyzed with the application that recommends in the page and to each application, obtains the alternative set of the second application；
Recommend set of applications acquisition module 530, specifically include：
Fusion Module 531, for carrying out alternative for the described first application set and the alternative set of the second application
Fusion, obtains and recommends set of applications.
Display module 540, for each application that recommends in set of applications to be shown.
Optionally, the alternative set generation module 520 of second application includes：
Sorted lists acquisition module, recommends the page for the application at least one application management platform,
Obtain corresponding application sorted lists.
List Fusion Module, for when it is two or more to apply sorted lists, each being applied
Sorted lists carry out fusion calculation, and obtaining includes N2Second application of individual application is alternative to gather, wherein N2
It is less than or equal to the summation of number of applications in all list of applications.
Optionally, the list Fusion Module includes：
Application fetches module, for applying in sorted lists from each, P forward application of selected and sorted；
Second initial characteristicses word computing module, for for each application, from the use related to the application
In the behavior record of family, the eigenvalue of at least one dimension is extracted；The eigenvalue bag of at least one dimension
Include：The eigenvalue of the first value relevance, user access activity dimensional characteristics value.
Second total characteristic value computing module, for the eigenvalue according to each dimension and the weight of respective dimensions,
Calculate the total characteristic value of the application.
Second alternative set constructing module, for sorting forward N total characteristic value2Individual application adds the
Two applications are alternative to gather.
Optionally, the list Fusion Module includes：
First scoring modules, in each application sorted lists, according to the application location,
Each application is given a mark.
Marking accumulation module, for being directed to same application, applies described in each application sorted lists
In marking added up, obtain the PTS of the application.
Sequence constructing module, for sorting forward N PTS2Individual application adds the second application alternative
Set.
Optionally, the Fusion Module 531 includes：
First filling module, for the front M by set of applications is recommended1Individual position, alternative from the second application
The forward M of selected and sorted in set1Individual application is filled.
Second filling module, for the remaining M by set of applications is recommended2Position, alternative from the first application
The corresponding M for selecting sequence forward in set2Individual application is filled, wherein M1With M2Sum is recommendation
The quantity that applies in set of applications.
Embodiment six
With reference to Fig. 6, a kind of structured flowchart of application commending system embodiment of the present invention is shown, specifically
Can include such as lower module：
The alternative set generation module 610 of first application, for obtaining the user behavior record related to application
And each application is analyzed, obtain the alternative set of the first application.
The alternative set generation module 620 of second application, for obtaining answering at least one application management platform
It is analyzed with the application that recommends in the page and to each application, obtains the alternative set of the second application.
Recommend set of applications acquisition module 630, specifically include：
Fusion Module 631, for carrying out alternative for the described first application set and the alternative set of the second application
Fusion, obtains and recommends set of applications.
Sort module 640, for for each application that recommends in set of applications, according to tag along sort mould
Type stamps tag along sort to the application.
Display module 650, for each application that recommends in set of applications to be shown.
Optionally, also include：
Tag along sort model construction module, for entering to the usage behavior of each function of applying according to each user
Row analysis, builds application class label model.
For system embodiment, due to itself and embodiment of the method basic simlarity, so the comparison of description
Simply, related part is illustrated referring to the part of embodiment of the method.
Each embodiment in this specification is described by the way of going forward one by one, and each embodiment is stressed
Be all difference with other embodiment, between each embodiment identical similar part mutually referring to
?.
Those skilled in the art are it should be appreciated that the embodiment of the embodiment of the present invention can be provided as method, be
System or computer program.Therefore, the embodiment of the present invention can adopt complete hardware embodiment, completely
Software implementation or with reference to software and hardware in terms of embodiment form.And, the embodiment of the present invention
Can adopt and storage can be used to be situated between in one or more computers for wherein including computer usable program code
The upper computer journey that implements of matter (including but not limited to disk memory, CD-ROM, optical memory etc.)
The form of sequence product.
In a typical configuration, the computer equipment includes one or more processors
(CPU), input/output interface, network interface and internal memory.Internal memory potentially includes computer-readable medium
In volatile memory, the shape such as random access memory (RAM) and/or Nonvolatile memory
Formula, such as read only memory (ROM) or flash memory (FLASH RAM).Internal memory is computer-readable medium
Example.Computer-readable medium includes permanent and non-permanent, removable and non-removable media
Information Store can be realized by any method or technique.Information can be computer-readable instruction,
Data structure, the module of program or other data.The example of the storage medium of computer includes, but
It is not limited to phase transition internal memory (PRAM), static RAM (SRAM), dynamic randon access
Memorizer (DRAM), other kinds of random access memory (RAM), read only memory
(ROM), Electrically Erasable Read Only Memory (EEPROM), fast flash memory bank or other in
Deposit technology, read-only optical disc read only memory (CD-ROM), digital versatile disc (DVD) or other
Optical storage, magnetic cassette tape, tape magnetic rigid disk storage other magnetic storage apparatus or any its
His non-transmission medium, can be used to store the information that can be accessed by a computing device.According to herein
Define, computer-readable medium does not include the computer readable media (TRANSITORY of non-standing
MEDIA), such as data signal and the carrier wave of modulation.
The embodiment of the present invention is with reference to method according to embodiments of the present invention, terminal unit (system) and meter
The flow chart and/or block diagram of calculation machine program product is describing.It should be understood that can be by computer program instructions
Each flow process and/or square frame and flow chart and/or square frame in flowchart and/or block diagram
The combination of flow process and/or square frame in figure.Can provide these computer program instructions to general purpose computer,
The processor of special-purpose computer, Embedded Processor or other programmable data processing terminal equipments is producing
One machine so that by the computing device of computer or other programmable data processing terminal equipments
Instruction produce for realizing in one flow process of flow chart or one square frame of multiple flow processs and/or block diagram or
The system of the function of specifying in multiple square frames.
These computer program instructions may be alternatively stored in and computer or other programmable datas can be guided to process
In the computer-readable memory that terminal unit is worked in a specific way so that be stored in the computer-readable
Instruction in memorizer is produced and includes the manufacture of instruction system, and instruction system realization is in flow chart one
The function of specifying in flow process or one square frame of multiple flow processs and/or block diagram or multiple square frames.
These computer program instructions can also be loaded into computer or other programmable data processing terminals set
Standby upper so that on computer or other programmable terminal equipments, execution series of operation steps is in terms of producing
The process that calculation machine is realized, the instruction so as to execute on computer or other programmable terminal equipments provide use
In realization in one flow process of flow chart or one square frame of multiple flow processs and/or block diagram or multiple square frames
The step of function of specifying.
Although having been described for the preferred embodiment of the embodiment of the present invention, those skilled in the art are once
Basic creative concept is known, then other change and modification can be made to these embodiments.So,
Claims are intended to be construed to include preferred embodiment and fall into the institute of range of embodiment of the invention
Have altered and change.
Finally, in addition it is also necessary to explanation, herein, such as first and second or the like relational terms
Be used merely to an entity or operation is made a distinction with another entity or operation, and not necessarily require
Or imply between these entities or operation, there is any this actual relation or order.And, art
Language " including ", "comprising" or its any other variant are intended to including for nonexcludability, so that
A series of process, method, article or terminal unit including key elements not only includes those key elements, and
Also include other key elements being not expressly set out, or also include for this process, method, article or
The intrinsic key element of person's terminal unit.In the absence of more restrictions, by sentence " including one
Individual ... " key element that limits, it is not excluded that at the process, method, article or the end that include the key element
Also there is other identical element in end equipment.
Above to a kind of application recommendation method provided by the present invention and a kind of application commending system, carry out
It is discussed in detail, specific case used herein is set forth to the principle of the present invention and embodiment,
The explanation of above example is only intended to help and understands the method for the present invention and its core concept；Meanwhile, right
In one of ordinary skill in the art, according to the thought of the present invention, in specific embodiment and range of application
On will change, in sum, this specification content should not be construed as limiting the invention.
Claims (12)
1. one kind applies recommendation method, it is characterised in that include：
Obtain the user behavior record related to application and each application is analyzed, obtain the first application standby
Selected works are closed；
According to the alternative set of the described first application, obtain and recommend set of applications；
The each application that recommends in set of applications is shown.
2. method according to claim 1, it is characterised in that the acquisition is related to application
User behavior records and each application is analyzed, and obtains the step of the first application is alternative to gather, including：
For each application, from the user behavior record related to the application, extract at least one and tie up
The eigenvalue of degree；The eigenvalue of at least one dimension includes：The eigenvalue of the first value relevance, use
Family accesses the eigenvalue of behavior dimension；
According to the eigenvalue for extracting at least one dimension and the weight of respective dimensions, the application is calculated
Total characteristic value；
By N forward for the sequence of total characteristic value1Individual application adds the alternative set of the first application.
3. method according to claim 2, it is characterised in that for each application, from institute
State in the related user behavior record of application, include the step of the eigenvalue for extracting at least one dimension：
For each application, calculate between the first value data of the application and the dispensed amount of the application
Ratio, using the ratio as the first value relevance eigenvalue.
4. method according to claim 2, the eigenvalue of the user access activity dimension include：
Browse dimension eigenvalue, download dimension eigenvalue.
5. method according to claim 4, it is characterised in that
When the eigenvalue of the user access activity dimension is the eigenvalue for browsing dimension, described for every
Individual application, from the user behavior record related to the application, extracts the eigenvalue of at least one dimension
The step of include：
For each application, count described and apply browsed in application summary displayed page first to browse
Number of times, the second number of visits browsed in result of page searching, and in application details displayed page
The 3rd number of visits for being browsed；
By first number of visits, the second number of visits and the 3rd number of visits respectively in connection with the place page
Weight, calculate and browse the eigenvalue of dimension.
6. method according to claim 4, it is characterised in that when the user access activity is tieed up
When the eigenvalue of degree is the eigenvalue for downloading dimension, described for each application, from related to the application
User behavior record in, extract at least one dimension eigenvalue the step of include：
For each application, first download time that applies in application summary displayed page is counted,
The second download time being downloaded in result of page searching, and in being downloaded in application details displayed page
The 3rd download time being downloaded；
By first download time, the second download time and the 3rd download time respectively in connection with the place page
Weight, calculate download dimension eigenvalue.
7. method according to claim 1, it is characterised in that standby according to the described first application
Selected works are closed, and before obtaining the step of recommending set of applications, are also included：
The application for obtaining at least one application management platform is recommended the application in the page and each application is carried out
Analysis, obtains the alternative set of the second application；
According to the alternative set of the described first application, obtaining the step of recommending set of applications is：
Alternative for described first application set and the alternative set of the second application are merged, is obtained and is recommended application
Set.
8. method according to claim 7, it is characterised in that the acquisition at least one is applied
The application of management platform is recommended the application in the page and each application is analyzed, and obtains the second application alternative
The step of set, includes：
For the page is recommended in the application of at least one application management platform, corresponding application Sorted list is obtained
Table；
When it is two or more to apply sorted lists, each application sorted lists is carried out fusion meter
Calculate, obtaining includes N2Second application of individual application is alternative to gather, wherein N2It is less than or equal to all
The summation of number of applications in list of application.
9. method according to claim 7, it is characterised in that described when application sorted lists be
Two or more when, each application sorted lists is carried out fusion calculation, obtaining includes N2Individual should
The step of second application is alternatively gathered includes：
From each application sorted lists, P forward application of selected and sorted；
For each application, from the user behavior record related to the application, extract at least one and tie up
The eigenvalue of degree；The eigenvalue of at least one dimension includes：The eigenvalue of the first value relevance, use
Family accesses the eigenvalue of behavior dimension；
According to eigenvalue and the weight of respective dimensions of each dimension, the total characteristic value of the application is calculated；
By N forward for the sequence of total characteristic value2Individual application adds the alternative set of the second application.
10. method according to claim 7, it is characterised in that described will be standby for the described first application
Selected works are closed and the alternative set of the second application is merged, and obtaining the step of recommending set of applications includes：
The front M of set of applications will be recommended1Individual position, forward from selected and sorted in the alternative set of the second application
M1Individual application is filled；
The remaining M of set of applications will be recommended2Individual position, selects sequence forward from the alternative set of the first application
Corresponding M2Individual application is filled；Wherein M1With M2Sum is to recommend application in set of applications
Quantity.
11. methods according to claim 7, it is characterised in that obtain recommend set of applications it
Afterwards, also include：
For each application that recommends in set of applications, the application is stamped point according to tag along sort model
Class label；And/or
The usage behavior of each function of applying is analyzed according to each user, builds application class label mould
Type.
12. a kind of application commending systems, it is characterised in that include：
The alternative set generation module of first application, records and right for obtaining the user behavior related to application
Each application is analyzed, and obtains the first application and alternatively gathers；
Recommend set of applications acquisition module, for according to the alternative set of the described first application, acquisition is recommended should
With set；
Display module, for each application that recommends in set of applications to be shown.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510568010.XA CN106503025B (en) | 2015-09-08 | 2015-09-08 | Application recommendation method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510568010.XA CN106503025B (en) | 2015-09-08 | 2015-09-08 | Application recommendation method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106503025A true CN106503025A (en) | 2017-03-15 |
| CN106503025B CN106503025B (en) | 2021-02-12 |
Family
ID=58286950
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510568010.XA Active CN106503025B (en) | 2015-09-08 | 2015-09-08 | Application recommendation method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106503025B (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105224614A (en) * | 2015-09-17 | 2016-01-06 | 北京金山安全软件有限公司 | Application program classification display method and device |
| CN107330719A (en) * | 2017-06-09 | 2017-11-07 | 上海新概念保险经纪有限公司 | A kind of insurance products recommend method and system |
| CN107704494A (en) * | 2017-08-24 | 2018-02-16 | 上海斐讯数据通信技术有限公司 | A kind of user information collection method and system based on application software |
| CN107844530A (en) * | 2017-10-17 | 2018-03-27 | 广东欧珀移动通信有限公司 | A kind of data processing method, device, server and computer-readable recording medium |
| CN107885572A (en) * | 2017-12-11 | 2018-04-06 | 广东欧珀移动通信有限公司 | Classification card generation method, system, server and computer-readable recording medium |
| WO2018223993A1 (en) * | 2017-06-08 | 2018-12-13 | 广州优视网络科技有限公司 | Application search method, device and server |
| CN109144721A (en) * | 2018-07-18 | 2019-01-04 | Oppo广东移动通信有限公司 | Resource ordering method, resource display method, relevant apparatus and storage medium |
| CN109657141A (en) * | 2018-12-13 | 2019-04-19 | 上海二三四五网络科技有限公司 | A kind of control method and control device for recommending application to user in mobile phone assistant system |
| CN110633420A (en) * | 2019-10-14 | 2019-12-31 | 北京代码乾坤科技有限公司 | Game content recommendation method and device, storage medium, processor and electronic device |
| CN110765352A (en) * | 2019-10-11 | 2020-02-07 | 上海上湖信息技术有限公司 | User interest identification method and device |
| WO2020087386A1 (en) * | 2018-10-31 | 2020-05-07 | 深圳市欢太科技有限公司 | Content recommendation method and device, mobile terminal, and server |
| TWI706358B (en) * | 2018-10-08 | 2020-10-01 | 合隆毛廠股份有限公司 | Information recommendation system and method |
| CN111880872A (en) * | 2020-06-28 | 2020-11-03 | 华为技术有限公司 | Method, terminal device, server and system for managing application program APP |
| CN117454018A (en) * | 2023-12-22 | 2024-01-26 | 深圳市康莱米电子股份有限公司 | Education platform implementation method and system based on tablet personal computer |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101887390A (en) * | 2010-06-23 | 2010-11-17 | 宇龙计算机通信科技(深圳)有限公司 | Method and device for evaluating rating of application software |
| CN103514496A (en) * | 2012-06-21 | 2014-01-15 | 腾讯科技（深圳）有限公司 | Method and system for processing recommended target software |
| CN103578007A (en) * | 2012-07-20 | 2014-02-12 | 三星电子（中国）研发中心 | Mixed recommendation system and method for intelligent device |
| CN103930871A (en) * | 2011-05-09 | 2014-07-16 | 谷歌公司 | Recommending applications for mobile devices based on installation histories |
| CN104090967A (en) * | 2014-07-16 | 2014-10-08 | 北京智谷睿拓技术服务有限公司 | Application program recommending method and device |
| CN104137138A (en) * | 2011-12-29 | 2014-11-05 | 三星电子株式会社 | Method, server, and terminal for recommending an application based on application usage |
| CN104239571A (en) * | 2014-09-30 | 2014-12-24 | 北京奇虎科技有限公司 | Method and device for application recommendation |
| CN104360858A (en) * | 2014-11-12 | 2015-02-18 | 华为技术有限公司 | Method and device for calculating hotness of application |
| CN104991914A (en) * | 2015-06-23 | 2015-10-21 | 腾讯科技（深圳）有限公司 | Application recommendation method and server |
-
2015
- 2015-09-08 CN CN201510568010.XA patent/CN106503025B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101887390A (en) * | 2010-06-23 | 2010-11-17 | 宇龙计算机通信科技(深圳)有限公司 | Method and device for evaluating rating of application software |
| CN103930871A (en) * | 2011-05-09 | 2014-07-16 | 谷歌公司 | Recommending applications for mobile devices based on installation histories |
| CN104137138A (en) * | 2011-12-29 | 2014-11-05 | 三星电子株式会社 | Method, server, and terminal for recommending an application based on application usage |
| CN103514496A (en) * | 2012-06-21 | 2014-01-15 | 腾讯科技（深圳）有限公司 | Method and system for processing recommended target software |
| CN103578007A (en) * | 2012-07-20 | 2014-02-12 | 三星电子（中国）研发中心 | Mixed recommendation system and method for intelligent device |
| CN104090967A (en) * | 2014-07-16 | 2014-10-08 | 北京智谷睿拓技术服务有限公司 | Application program recommending method and device |
| CN104239571A (en) * | 2014-09-30 | 2014-12-24 | 北京奇虎科技有限公司 | Method and device for application recommendation |
| CN104360858A (en) * | 2014-11-12 | 2015-02-18 | 华为技术有限公司 | Method and device for calculating hotness of application |
| CN104991914A (en) * | 2015-06-23 | 2015-10-21 | 腾讯科技（深圳）有限公司 | Application recommendation method and server |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105224614A (en) * | 2015-09-17 | 2016-01-06 | 北京金山安全软件有限公司 | Application program classification display method and device |
| WO2018223993A1 (en) * | 2017-06-08 | 2018-12-13 | 广州优视网络科技有限公司 | Application search method, device and server |
| CN107330719A (en) * | 2017-06-09 | 2017-11-07 | 上海新概念保险经纪有限公司 | A kind of insurance products recommend method and system |
| CN107704494A (en) * | 2017-08-24 | 2018-02-16 | 上海斐讯数据通信技术有限公司 | A kind of user information collection method and system based on application software |
| CN107844530A (en) * | 2017-10-17 | 2018-03-27 | 广东欧珀移动通信有限公司 | A kind of data processing method, device, server and computer-readable recording medium |
| CN107885572A (en) * | 2017-12-11 | 2018-04-06 | 广东欧珀移动通信有限公司 | Classification card generation method, system, server and computer-readable recording medium |
| CN109144721A (en) * | 2018-07-18 | 2019-01-04 | Oppo广东移动通信有限公司 | Resource ordering method, resource display method, relevant apparatus and storage medium |
| TWI706358B (en) * | 2018-10-08 | 2020-10-01 | 合隆毛廠股份有限公司 | Information recommendation system and method |
| WO2020087386A1 (en) * | 2018-10-31 | 2020-05-07 | 深圳市欢太科技有限公司 | Content recommendation method and device, mobile terminal, and server |
| CN109657141A (en) * | 2018-12-13 | 2019-04-19 | 上海二三四五网络科技有限公司 | A kind of control method and control device for recommending application to user in mobile phone assistant system |
| CN110765352B (en) * | 2019-10-11 | 2022-11-11 | 上海上湖信息技术有限公司 | User interest identification method and device |
| CN110765352A (en) * | 2019-10-11 | 2020-02-07 | 上海上湖信息技术有限公司 | User interest identification method and device |
| CN110633420A (en) * | 2019-10-14 | 2019-12-31 | 北京代码乾坤科技有限公司 | Game content recommendation method and device, storage medium, processor and electronic device |
| CN111880872A (en) * | 2020-06-28 | 2020-11-03 | 华为技术有限公司 | Method, terminal device, server and system for managing application program APP |
| CN117454018A (en) * | 2023-12-22 | 2024-01-26 | 深圳市康莱米电子股份有限公司 | Education platform implementation method and system based on tablet personal computer |
| CN117454018B (en) * | 2023-12-22 | 2024-03-08 | 深圳市康莱米电子股份有限公司 | Education platform implementation method and system based on tablet personal computer |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106503025B (en) | 2021-02-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106503025A (en) | Method and system is recommended in a kind of application | |
| US20210271975A1 (en) | User tag generation method and apparatus, storage medium, and computer device | |
| CN103324665B (en) | Hot spot information extraction method and device based on micro-blog | |
| CN108363821A (en) | A kind of information-pushing method, device, terminal device and storage medium | |
| CN110532451A (en) | Search method and device for policy text, storage medium, electronic device | |
| CN106372249B (en) | A kind of clicking rate predictor method, device and electronic equipment | |
| CN101283353B (en) | The system and method for relevant documentation is found by analyzing tags | |
| CN105989004A (en) | Information releasing pretreatment method and device | |
| CN106339502A (en) | Modeling recommendation method based on user behavior data fragmentation cluster | |
| CN111708740A (en) | Mass search query log calculation analysis system based on cloud platform | |
| CN104199833B (en) | The clustering method and clustering apparatus of a kind of network search words | |
| CN104899273A (en) | Personalized webpage recommendation method based on topic and relative entropy | |
| US20150205580A1 (en) | Method and System for Sorting Online Videos of a Search | |
| CN106445963B (en) | Advertisement index keyword automatic generation method and device of APP platform | |
| CN104602042A (en) | User behavior based label setting method | |
| KR101566616B1 (en) | Advertisement decision supporting system using big data-processing and method thereof | |
| CN106610970A (en) | Collaborative filtering-based content recommendation system and method | |
| CN102542066B (en) | Video clustering method, ordering method, video searching method and corresponding devices | |
| CN110134845A (en) | Project public sentiment monitoring method, device, computer equipment and storage medium | |
| CN110175264A (en) | Construction method, server and the computer readable storage medium of video user portrait | |
| CN106570031A (en) | Service object recommending method and device | |
| KR100954842B1 (en) | Method and System of classifying web page using category tag information and Recording medium using by the same | |
| CN112749330B (en) | Information pushing method, device, computer equipment and storage medium | |
| CN111079009A (en) | User interest detection method and system for government map service | |
| CN106095939A (en) | The acquisition methods of account authority and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |