CN107220269A - A kind of personalized recommendation method for geographical position sensitivity app - Google Patents
A kind of personalized recommendation method for geographical position sensitivity app Download PDFInfo
- Publication number
- CN107220269A CN107220269A CN201610817958.9A CN201610817958A CN107220269A CN 107220269 A CN107220269 A CN 107220269A CN 201610817958 A CN201610817958 A CN 201610817958A CN 107220269 A CN107220269 A CN 107220269A
- Authority
- CN
- China
- Prior art keywords
- app
- mrow
- user
- susceptibility
- matrix
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 230000035945 sensitivity Effects 0.000 title claims abstract description 17
- 239000011159 matrix material Substances 0.000 claims abstract description 40
- 238000012545 processing Methods 0.000 claims abstract description 5
- 238000000354 decomposition reaction Methods 0.000 claims description 19
- 238000005457 optimization Methods 0.000 claims description 6
- 238000011156 evaluation Methods 0.000 claims description 3
- 230000006872 improvement Effects 0.000 abstract description 3
- 238000005516 engineering process Methods 0.000 description 7
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 239000000686 essence Substances 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000035800 maturation Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 241000894007 species Species 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
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
- G06F16/9537—Spatial or temporal dependent retrieval, e.g. spatiotemporal queries
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention relates to a kind of personalized recommendation method for geographical position sensitivity app, specific implementation is mainly included the following steps that:The acquisition and processing of step 1, initial data；Step 2, city correspondence app susceptibilitys are calculated；Step 3, rating matrix are predicted and provide app recommendations.Calculated by the susceptibility to geographical position sensing app, the matrix disassembling method with geographic location weight passes through the technological improvement in terms of the two so that the prediction of recommendation score is more accurate.
Description
Technical field
The present invention relates to a kind of personalized recommendation method for geographical position sensitivity app.
Background technology
With the popularization of smart mobile phone, mobile phone application app quantity is also more and more, and species is more and more numerous and diverse.In order to more preferable
Understanding current mobile phone app application state of development, the mobile phone app of research three main flows instantly provides shop google online
Play, iOS app store and Amazon app store.It is the maximum offers of mobile phone app that these app provide shop online
Shop, for their advantage and disadvantage, it is each has something to recommend him to say.But, the characteristics of they but have one jointly is exactly：Carry
The mobile phone app enormous amounts of confession, number of users is huge, covers all parts of the world, including each stratum.Also, their mobile phone
App quantity is still in the trend constantly risen (such as Fig. 1).The app that current App store are included has exceeded 1,200,000, attraction
More than 280000 developer (the same developer can develop multiple applications).Google play as Android platform offer
Business, user and in the market are more advantageous, and the app included at present alreadys exceed 1,430,000, and platform has attracted 390,000 developers.
Amazon is then 290,000 and 4.8 ten thousand.
In face of the app of magnanimity, with the mobile phone app of a application of function, remain to find out tens moneys up to a hundred.Moreover, different
Mobile phone app has even more very different in different characteristics, quality and Consumer's Experience.This just to user bring one it is very serious
Problem:User on earth will how from the app of magnanimity application, find it is a be adapted to oneself and high-quality app
This is accomplished by providing the user a kind of recommendation mechanisms, and according to the personality feature of user, living environment, age-sex's feature is very
It is that user automatically finds out his suitable high-quality app applications as weather time etc., the app of this namely personalization is applied
Recommend.In view of this there is provided an algorithm so that this recommendation becomes precisely reliable, so as to strengthen the experience of user so that use
Family can save substantial amounts of time and energy from the app of magnanimity provides shop, look for most suitable app.
Prior art：Mobile phone application app recommended technology, most of is to inherit to recommend in computer pc ends commodity, film etc.
Technology, but also there is oneself exclusive characteristic.Mobile phone application app recommended technology is started to walk soon at present, not as pc ends are recommended
Technology maturation, but be due to develop on the basis of pc recommended technologies, the development of the recommended technology of mobile phone terminal app applications is suitable
Rapidly.At present, main recommendation method includes：Recommend based on commending contents, collaborative filtering recommending, based on correlation rule, be based on
Effectiveness is recommended, knowledge based is recommended and combined recommendation.These algorithms all comparative maturities, but all respectively have its good and bad (such as table 1).Especially
It is that have obvious region (i.e. when a app is applied：When this app is specific to the application somewhere developed),
These traditional corresponding algorithms often these softwares to falling into oblivion, but in fact these app for local people,
Their most suitable app applications may be only.
Table 1
The content of the invention
It is an object of the invention to overcome the shortcoming and deficiency of prior art, it is proposed that one kind is sensitive for geographical position
App personalized recommendation method, is calculated, the matrix with geographic location weight point by the susceptibility to geographical position sensing app
Solution method, passes through the technological improvement in terms of the two so that the prediction of recommendation score is more accurate.
To achieve the above object, the technical solution adopted in the present invention is：
A kind of personalized recommendation method for geographical position sensitivity app, it is characterised in that comprise the following steps:
The acquisition and processing of step 1, initial data；
Step 2, city correspondence app susceptibilitys are calculated；
Step 3, rating matrix are predicted and provide app recommendations.
Further, the step 1 is using the download, unloading, update status of the app to draw use by counting user
Scoring of the family for certain a app.
Further, step 2 specifically includes following steps：
The geographical location sen-sitivity of App correspondences is calculated；
Matrix decomposition with susceptibility weight.
Further, the geographical location sen-sitivity of App correspondences, which is calculated, specifically includes following steps：
1) collect and handle original user traffic data, count each city for the use per a app
Amount；
2) using tf-idf algorithms to step 1) matrix that draws carries out preliminary treatment, obtains original city and app
Susceptibility relation；
3) a confidence level function is defined, the susceptibility relation come between further Optimizing City and app.
Further, the matrix decomposition with susceptibility weight, which is specifically included, uses decomposition formula：
Then decomposed and reconstituted Ri, j pass through optimization object function as predicted value so that predicted value is approached as far as possible
Actual value, object function：
In the matrix disassembling method with geographical position susceptibility, geographical position sensitivity is added when decomposing
The weight of degree, object function is just changed into：
N in formula, M represent N number of user and M app respectively,What is represented is user i positions for appj
Susceptibility.ri,jWhat is represented is evaluations point of the user i for app j.
Further, optimization method preferably is least square method or stochastic gradient descent.
Further, step 3 is specifically included carrys out the structure of prediction steps 1 using the matrix decomposition with city app susceptibility weights
User-the App built rating matrix, removes to predict the missing values in rating matrix with the method for matrix decomposition, has finally drawn
Whole rating matrix, by the rating matrix of the user-app, for some user with regard to the row of an app scoring can be provided
The high app of scoring is recommended user by sequence, system automatically.
The beneficial effects of the invention are as follows：Calculated by the susceptibility to geographical position sensing app, with geographic location weight
Matrix disassembling method, passes through the technological improvement in terms of the two so that the prediction of recommendation score is more accurate.
Brief description of the drawings
Fig. 1 is mobile phone app volume trends figures in 2016；
Fig. 2 is a kind of personalized recommendation method flow chart for geographical position sensitivity app；
Fig. 3 is No. 802 obvious abnormal conditions schematic diagram of app；
Fig. 4 is the matrix decomposition figure with susceptibility weight.
Embodiment
Technical solution of the present invention is further illustrated below in conjunction with the accompanying drawings：
A kind of personalized recommendation method for geographical position sensitivity app, flow chart are as shown in Fig. 2 specific implementation is main
Comprise the following steps:
The acquisition and processing of step 1, initial data；
Step 2, city correspondence app susceptibilitys are calculated；
Step 3, rating matrix are predicted and provide app recommendations.
Wherein, the acquisition and processing of step 1, initial data：
App datas on flows are used by built-in mobile phone software collection cellphone subscriber, the target data of collection mainly includes：With
Family, the geographical position of user, the uninstall situation of user, user daily using mobile phone app situation (flow that uses, when
Between, number of times etc.) and user download the situation for updating software.The data collected by our embedded softwares, build use
Family-App grade forms are shown in Table 2, and foregoing city is for App sensitivity matrix.
| Download | Unloading | Update | Score |
| 1 | 0 | 0 | 1 |
| 1 | 1 | 0 | 2 |
| 1 | 0 | 0 | 3 |
| 1 | 1 | 1 | 4 |
| 1 | 0 | 1 | 5 |
Table 2
As shown in table 2 above, user is used under the app for certain a app scoring by counting user
Carry, unloading, update status is come the scoring that draws.
Step 2, city correspondence app susceptibilitys are calculated, and app sensitivity matrixes can be drawn by the following method：
First, App correspondences geographical position (mainly in units of city) susceptibility is calculated：
1) collect and handle original user traffic data, count each city for the use per a app
Amount.The result formats of statistics are probably such as table 3.This data mode of table 3, exactly meets the data structure of tf-idf algorithms
's.
| Tengxun qq | Shenzhen bus | Visit | ||
| Beijing | 2332323 | 33 | - | - |
| Shanghai | 3433434 | 1 | - | - |
| Shenzhen | 2233232 | 343444 | - | - |
2) using tf-idf algorithms to step 1) matrix that draws carries out preliminary treatment, obtains original city and app
Susceptibility relation.Value in matrix is between 0,1.Such as Beijing is to the susceptibility of first item software Tengxun qq in sectional drawing
For 0.269.Here, illustratively tf-idf algorithms.
TF-IDF (term frequency-inverse document frequency) be it is a kind of be used for information retrieval with
The conventional weighting technique of data mining.It is a kind of statistical method.Main thought is：If some word or phrase are in an article
The frequency TF of appearance is high, and seldom occurs in other articles, then it is assumed that this word or phrase have good class discrimination
Ability, is adapted to classification.TFIDF is actually：TF*IDF, TF word frequency (Term Frequency), the reverse document-frequencies of IDF
(Inverse Document Frequency).TF represents the frequency that entry occurs in document d.IDF main thought is：Such as
Document of the fruit comprising entry t is fewer, that is, n smaller, and IDF is bigger, then illustrates that entry t has good class discrimination ability.
If the number of files comprising entry t is m in a certain class document C, and total number of documents of other classes comprising t is k, it is clear that all to include
T number of files n=m+k, when m is big, n is also big, and the IDF obtained according to IDF formula value can be small, just illustrates entry t
Class discrimination is indifferent.If but in fact, an entry is frequently occurred in the document of a class, illustrating the entry
The feature of the text of this class can be represented very well, and such entry should assign higher weight to them, and choosing is used as
The Feature Words of the class text are to distinguish and other class documents.Here it is IDF weak point is in the given file of portion, word
Frequently (term frequency, TF) refers to the frequency that some given word occurs in this document.This numeral is pair
The normalization of word number (term count), to prevent it to be inclined to long file.(same word may be than short in long file
File has higher word number, whether important but regardless of the word.) for the word in a certain specific file, its weight
The property wanted is represented by：
In above equation molecule is the occurrence number of the word hereof, and denominator is then that all words go out hereof
Occurrence number sum.Reverse document-frequency (inverse document frequency, IDF) is a word general importance
Measurement.The IDF of a certain particular words, can file by general act number divided by comprising the word number, then will obtain
Business, which takes the logarithm, to be obtained：
Wherein
|D|：Total number of files in corpus：Number of files (number of files i.e.) comprising word is not if the word exists
In corpus, may result in denominator is zero, therefore is generally used
1+|{d∈D：t∈d}|
It is used as denominator.
Then TF and IDF product are calculated again.
tfidfi,j=tfi,j×idfi
High term frequencies in a certain specific file, and low document-frequency of the word in whole file set, can
To produce the TF-IDF of high weight.Therefore, TF-IDF tends to filter out common word, retains important word.
Come to use in city and app relation based on this principle analogy.Tfidf, which is calculated, can just illustrate, if some app
It is higher in some city usage amount, but when other city usage amounts are low, then the tfidf results calculated are just
Can be big, also just illustrate susceptibility of the city to certain app.
3) a confidence level function is defined, the susceptibility relation come between further Optimizing City and app.In step 2) calculate
Method has calculated city and the direct susceptibilitys of app by tfidf, but reality is when be Data Collection, at some
The data volume that big city is collected much will be greater than some relatively backward cities.Data volume is fewer, the city that the data reflect
Relation between app is just inaccurate all the more.In other words, data volume is fewer, and the sensitive relations drawn just may not be accurate.This
In by analyze data, define a confidence level function.
Function value sends out confidence level of f (x) value with regard to some city between 0,1, and x represents all app in the city and averagely used
Family is measured, and customer volume is bigger, and confidence level is higher.
The susceptibility defined by algorithm, is found through experiments that some certain cities are more sensitive for some app.As schemed
Shown in 3, No. 802 app has obvious abnormal conditions.By network check No. 802 app be Shenzhen a seating it is public
The app of automobile.This app is basic only to be used in several cities on Shenzhen and periphery, with obvious region characteristic.So deep
Ditch between fields is high for the susceptibility of this software.This explanation, the susceptibility algorithm of definition is correct reliable.
2nd, the matrix decomposition with susceptibility weight.
Matrix disassembling method is the more traditional method of commending system, will be disassembled to matrix as the product of several matrixes.
Common are three kinds of 1) triangle decomposition method (Triangular Factorization), 2) QR decomposition methods (QR
Factorization), 3) singular value decomposition method (Singular Value Decompostion).The method that the present invention is used exists
Improved on the basis of QR decomposition methods.Decomposition formula：
Exploded view is as shown in Figure 4.
Decomposed and reconstituted Ri,jIt is used as predicted value.Then optimization object function is passed through so that predicted value is approached very as far as possible
Real value.Common optimization method is：Least square method and stochastic gradient descent.
Object function：
In the matrix disassembling method with geographical position susceptibility, geographical position sensitivity is added when decomposing
The weight of degree.Object function is just changed into：
N in formula, M represent N number of user and M app respectively,What is represented is user i positions for appj
Susceptibility.ri,jWhat is represented is evaluations point of the user i for app j.
Step 3, rating matrix are predicted and provide app recommendations：
Predicted using the matrix decomposition with city app susceptibility weights, the scoring square for the user-App that step 1 is built
Battle array.Remove to predict the missing values in rating matrix with the method for matrix decomposition.Finally draw complete rating matrix.By this
User-app rating matrix, high app will be scored certainly for some user with regard to that can provide the sequence of an app scoring, system
It is dynamic to recommend user.
Above-described embodiment is preferably embodiment, but embodiments of the present invention are not by above-described embodiment of the invention
Limitation, other any Spirit Essences without departing from the present invention and the change made under principle, modification, replacement, combine, simplification,
Equivalent substitute mode is should be, is included within protection scope of the present invention.
Claims (7)
1. a kind of personalized recommendation method for geographical position sensitivity app, it is characterised in that comprise the following steps:
The acquisition and processing of step 1, initial data；
Step 2, city correspondence app susceptibilitys are calculated；
Step 3, rating matrix are predicted and provide app recommendations.
2. according to the method described in claim 1, it is characterised in that the step 1 is used under the app by counting user
Carry, unloading, update status draw scoring of the user for certain a app.
3. according to the method described in claim 1, it is characterised in that step 2 specifically includes following steps：
The geographical location sen-sitivity of App correspondences is calculated；
Matrix decomposition with susceptibility weight.
4. method according to claim 3, it is characterised in that：The geographical location sen-sitivity of App correspondences calculates specific bag
Include following steps：
1) collect and handle original user traffic data, count each city for the usage amount per a app；
2) using tf-idf algorithms to step 1) matrix that draws carries out preliminary treatment, obtains original city and app sensitivity
Degree relation；
3) a confidence level function is defined, the susceptibility relation come between further Optimizing City and app.
5. method according to claim 3, it is characterised in that the matrix decomposition with susceptibility weight, which is specifically included, adopts
Use decomposition formula：
Ri,j=Ui TVj；
Decomposed and reconstituted Ri,jAs predicted value, then pass through optimization object function so that predicted value approaching to reality as far as possible
Value, object function：
<mrow>
<mi>L</mi>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>N</mi>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>j</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>M</mi>
</munderover>
<msup>
<mrow>
<mo>(</mo>
<msub>
<mi>r</mi>
<mrow>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
</mrow>
</msub>
<mo>-</mo>
<msubsup>
<mi>U</mi>
<mi>i</mi>
<mi>T</mi>
</msubsup>
<msub>
<mi>V</mi>
<mi>j</mi>
</msub>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
</mrow>
In the matrix disassembling method with geographical position susceptibility, geographical position susceptibility is added when decomposing
Weight, object function is just changed into：
<mrow>
<mi>L</mi>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>N</mi>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>j</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>M</mi>
</munderover>
<msubsup>
<mi>W</mi>
<mrow>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
</mrow>
<mrow>
<mi>l</mi>
<mi>o</mi>
<mi>c</mi>
<mi>a</mi>
<mi>t</mi>
<mi>i</mi>
<mi>o</mi>
<mi>n</mi>
</mrow>
</msubsup>
<msup>
<mrow>
<mo>(</mo>
<msub>
<mi>r</mi>
<mrow>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
</mrow>
</msub>
<mo>-</mo>
<msubsup>
<mi>U</mi>
<mi>i</mi>
<mi>T</mi>
</msubsup>
<msub>
<mi>V</mi>
<mi>j</mi>
</msub>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>.</mo>
</mrow>
N in formula, M represent N number of user and M app, W respectivelylocation i,jWhat is represented is user i positions for appj
Susceptibility.ri,jWhat is represented is evaluations point of the user i for app j.
6. method according to claim 5, optimization method preferably is least square method or stochastic gradient descent.
7. according to the method described in claim 1, it is characterised in that step 3 is specifically included using band city app susceptibility weights
Matrix decomposition come prediction steps 1 structure user-App rating matrix, gone to predict scoring square with the method for matrix decomposition
Missing values in battle array, finally draw complete rating matrix, by the rating matrix of the user-app, for some user
The high app of scoring is recommended user by the sequence with regard to that can provide an app scoring, system automatically.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610817958.9A CN107220269B (en) | 2016-09-12 | 2016-09-12 | Personalized recommendation method for geographic position sensitive app |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610817958.9A CN107220269B (en) | 2016-09-12 | 2016-09-12 | Personalized recommendation method for geographic position sensitive app |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107220269A true CN107220269A (en) | 2017-09-29 |
| CN107220269B CN107220269B (en) | 2020-10-30 |
Family
ID=59928153
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610817958.9A Active CN107220269B (en) | 2016-09-12 | 2016-09-12 | Personalized recommendation method for geographic position sensitive app |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107220269B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107967330A (en) * | 2017-11-27 | 2018-04-27 | 中山大学 | Method is recommended based on the mobile application for downloading behavioral data and vectorial representative learning |
| CN110119465A (en) * | 2019-05-17 | 2019-08-13 | 哈尔滨工业大学 | Merge the mobile phone application user preferences search method of LFM latent factor and SVD |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011064675A1 (en) * | 2009-11-30 | 2011-06-03 | France Telecom | Method and system to recommend applications from an application market place |
| CN103617075A (en) * | 2013-12-04 | 2014-03-05 | 百度在线网络技术（北京）有限公司 | Application program recommending method, system and server |
| CN104813316A (en) * | 2012-06-04 | 2015-07-29 | 苹果公司 | Mobile device with localized app recommendations |
| CN105227445A (en) * | 2015-10-23 | 2016-01-06 | 中国联合网络通信集团有限公司 | Platform is recommended in application recommend method and application |
-
2016
- 2016-09-12 CN CN201610817958.9A patent/CN107220269B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011064675A1 (en) * | 2009-11-30 | 2011-06-03 | France Telecom | Method and system to recommend applications from an application market place |
| CN104813316A (en) * | 2012-06-04 | 2015-07-29 | 苹果公司 | Mobile device with localized app recommendations |
| CN103617075A (en) * | 2013-12-04 | 2014-03-05 | 百度在线网络技术（北京）有限公司 | Application program recommending method, system and server |
| CN105227445A (en) * | 2015-10-23 | 2016-01-06 | 中国联合网络通信集团有限公司 | Platform is recommended in application recommend method and application |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107967330A (en) * | 2017-11-27 | 2018-04-27 | 中山大学 | Method is recommended based on the mobile application for downloading behavioral data and vectorial representative learning |
| CN110119465A (en) * | 2019-05-17 | 2019-08-13 | 哈尔滨工业大学 | Merge the mobile phone application user preferences search method of LFM latent factor and SVD |
| CN110119465B (en) * | 2019-05-17 | 2023-06-13 | 哈尔滨工业大学 | Mobile phone application user preference retrieval method integrating LFM potential factors and SVD |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107220269B (en) | 2020-10-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109359244B (en) | Personalized information recommendation method and device | |
| CN103914478B (en) | Webpage training method and system, webpage Forecasting Methodology and system | |
| CN108287864B (en) | Interest group dividing method, device, medium and computing equipment | |
| CN105808526B (en) | Commodity short text core word extracting method and device | |
| CN109299994B (en) | Recommendation method, device, equipment and readable storage medium | |
| Shinde et al. | Hybrid personalized recommender system using centering-bunching based clustering algorithm | |
| CN107862022B (en) | Culture resource recommendation system | |
| CN106682686A (en) | User gender prediction method based on mobile phone Internet-surfing behavior | |
| US20140108190A1 (en) | Recommending product information | |
| CN110163476A (en) | Project intelligent recommendation method, electronic device and storage medium | |
| CN109190023A (en) | The method, apparatus and terminal device of Collaborative Recommendation | |
| CN109102371A (en) | Method of Commodity Recommendation, device, computer equipment and storage medium | |
| CN106651542A (en) | Goods recommendation method and apparatus | |
| CN106844407B (en) | Tag network generation method and system based on data set correlation | |
| CN104239338A (en) | Information recommendation method and information recommendation device | |
| CN109165975A (en) | Label recommendation method, device, computer equipment and storage medium | |
| Manley et al. | A review of machine learning and big data applications in addressing ecosystem service research gaps | |
| CN109460519B (en) | Browsing object recommendation method and device, storage medium and server | |
| CN105956083A (en) | Application software classification system, application software classification method and server | |
| CN104317784A (en) | Cross-platform user identification method and cross-platform user identification system | |
| CN109271806A (en) | Research on Privacy Preservation Mechanism based on user behavior | |
| CN103150667A (en) | Individualization recommending method based on mechanical structure | |
| CN104615631A (en) | Information recommendation method and device | |
| Zandian et al. | Feature extraction method based on social network analysis | |
| CN106452809A (en) | Data processing method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |