CN107590400A - A kind of recommendation method and computer-readable recording medium for protecting privacy of user interest preference - Google Patents
A kind of recommendation method and computer-readable recording medium for protecting privacy of user interest preference Download PDFInfo
- Publication number
- CN107590400A CN107590400A CN201710707014.0A CN201710707014A CN107590400A CN 107590400 A CN107590400 A CN 107590400A CN 201710707014 A CN201710707014 A CN 201710707014A CN 107590400 A CN107590400 A CN 107590400A
- Authority
- CN
- China
- Prior art keywords
- user
- video
- topic
- privacy
- similarity
- 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.)
- Pending
Links
Abstract
The present invention discloses a kind of recommendation method for protecting privacy of user interest preference; including the recommendation method for indirect access attack protection privacy of user and the user terminal rendering method of personal interest preference is preserved, learns the privacy correlation of the topic according to the rating information of video topic;The weight of different topics in user behavior record, and the user's similarity recorded according to the weight amendment based on user behavior are sexually revised using the privacy correlation of the video topic;According to input parameter of the user's similarity as difference secret protection index Mechanism Model, user's similarity is recalculated, calculates video fraction, and select the high user of similarity as neighbours;The recommendation list of video is sent to the user terminal;The video fraction of the user terminal foundation individual subscriber topic preference amendment recommendation is simultaneously ranked up presentation.The present invention attacks for indirect access, and can protect user's sensitivity interest preference, and do not sacrifice recommendation performance.
Description
Technical field
The present invention relates to information security field, more particularly, to a kind of recommendation side for protecting privacy of user interest preference
Method and computer-readable recording medium.
Background technology
With the fast development of video website, website user's privacy faces increasingly severe disclosure risk, especially right and wrong
Directly access attack and bring huge challenge to privacy of user protection work.Existing K arest neighbors (kNN, k-
NearestNeighbor in) attacking, the information of targeted customer can be obtained among the content recommendation from commending system to its neighbour
Take, the open output without directly accessing commending system.Specifically, KNN attackers are knowing target user part information
After the basic recommendation mechanism of commending system, the known record of targeted customer can be replicated completely and establishes several Virtual User,
Each Virtual User forms neighbours with high probability and targeted customer and other Virtual User, and obtains recommendation based on neighbours
Breath.Obtain new of Virtual User and the record in Given information is not included in, is exactly the record of targeted customer.This attack ratio
Directly access attack to be more difficult to prevent, and be not easy to be noticeable.
The method of the existing protection privacy of user based on user behavior data, all to sacrifice the accuracy recommended as generation
Valency, carry out balance selection.However, the degree of privacy of the interest preference of user is not consistent, interest of the sensitivity with privacy is inclined
The leakage that good leakage is significantly larger than popular preference influences to caused by user.How in guarding website user privacy information it is same
When keep system recommendation performance turn into it is in the urgent need to address the problem of.
Accordingly, it is desirable to provide a kind of recommendation method for the indirect protection privacy of user interest preference for accessing attack.
The content of the invention
It is an object of the invention to provide a kind of guarantor for the indirect differentiation interest preference sensitivity for accessing attack
Protect the recommendation method of privacy of user.The method can protect user's sensitivity interest preference, and not sacrifice recommendation performance.
To reach above-mentioned purpose, the present invention uses following technical proposals:
A kind of recommendation method for indirect access attack protection privacy of user, the recommendation method include:S101:Root
Learn the privacy correlation of the topic according to the rating information of video topic;S103:Utilize the privacy correlation of the video topic
Change the weight of different topics in user behavior record, the higher weight of topic privacy correlation is lower, according to the weight
Correct user's similarity based on user behavior record;S105:According to user's similarity as difference secret protection index
The input parameter of Mechanism Model, user's similarity is recalculated, calculate video fraction, and select the high user of similarity to make
For neighbours;S107:The recommendation list of video is sent to user terminal.
Preferably, the video topic is related to that sensitive rank is higher, and the privacy correlation is higher.
Preferably, the difference secret protection index Mechanism Model also includes Laplace mechanism, on the video fraction
The Laplace noises that average value is zero are added, video fraction are recalculated, for reducing the order of the recommendation list to user
The leakage of interest preference.
A kind of user terminal rendering method for preserving personal interest preference, according to above-mentioned of individual subscriber topic preference amendment
The video fraction of one scheme recommendation is simultaneously ranked up presentation.
Preferably, the video fraction is arranged using descending, final choice Top-k recommendation results.
A kind of recommendation apparatus for indirect access attack protection privacy of user, the recommendation apparatus include:Learn mould
Block, for learning the privacy correlation of the topic according to the rating information of video topic;Correcting module, for according to the video
The privacy correlation of topic sexually revises the weight of different topics in user behavior record, and the higher weight of topic privacy correlation is more
User's similarity that is low, being recorded according to the weight amendment based on user behavior;Selecting module, for similar according to the user
The input parameter as difference secret protection index Mechanism Model is spent, recalculates user's similarity, calculates video fraction,
And the high user of similarity is selected as neighbours;And sending module, for the recommendation list of video to be sent into user terminal.
Preferably, add module of making an uproar, the difference secret protection index Mechanism Model also includes Laplace mechanism, for
The Laplace noises that average value is zero are added on the video fraction, recalculate video fraction, for reducing the recommendation row
Leakage of the order of table to user interest preference.
A kind of user terminal for preserving personal interest preference, for according to any of the above-described side of individual subscriber topic preference amendment
The video fraction of case recommendation is simultaneously ranked up.
A kind of recommendation method for protecting privacy of user interest preference, it is characterised in that the recommendation method includes:
S101:Learn the privacy correlation of the topic according to the rating information of video topic;
S103:The weight of different topics in user behavior record, words are sexually revised using the privacy correlation of the video topic
It is lower to inscribe the higher weight of privacy correlation, the user's similarity recorded according to the weight amendment based on user behavior;
S105:According to input parameter of the user's similarity as difference secret protection index Mechanism Model, count again
User's similarity is calculated, calculates video fraction, and select the high user of similarity as neighbours;
S107:The recommendation list of video is sent to user terminal;
S109:The user terminal is according to video fraction described in individual subscriber topic preference and is ranked up presentation;
S111:The user terminal is arranged using descending the video fraction, final choice Top-k recommendation results.
A kind of computer-readable medium, it has program stored therein, when computing device described program, to cause:According to regarding
The rating information of frequency topic learns the privacy correlation of the topic;User's row is sexually revised using the privacy correlation of the video topic
For the weight of different topics in record, topic privacy correlation is higher, and weight is lower, and user's row is based on according to the weight amendment
For user's similarity of record;According to input parameter of the user's similarity as difference secret protection index Mechanism Model,
User's similarity is recalculated, calculates video fraction, and select the high user of similarity as neighbours;By the recommendation of video
List is sent to user terminal;The user terminal is according to video fraction described in individual subscriber topic preference and is ranked up presentation;Institute
State user terminal to arrange the video fraction using descending, final choice Top-k recommendation results.
Beneficial effects of the present invention are as follows:
Technical scheme of the present invention uses difference secret protection, distinguishing protection difference for indirect access attack
The interest preference of sensitivity, while protection user's sensitivity interest preference can be realized, recommendation performance is not sacrificed, have more preferable
Practicality.
Brief description of the drawings
The embodiment of the present invention is described in further detail below in conjunction with the accompanying drawings.
Fig. 1 shows the structured flowchart of the recommendation method of protection privacy of user interest preference of the present invention;
Fig. 2 shows flow chart of the present invention for the recommendation method of indirect access attack protection privacy of user;
Fig. 3 shows block diagram of the present invention for the recommendation apparatus of indirect access attack protection privacy of user;
Fig. 4 shows the flow chart of the recommendation method of protection privacy of user interest preference of the present invention.
Embodiment
In order to illustrate more clearly of the present invention, the present invention is done further with reference to preferred embodiments and drawings
It is bright.Similar part is indicated with identical reference in accompanying drawing.It will be appreciated by those skilled in the art that institute is specific below
The content of description is illustrative and be not restrictive, and should not be limited the scope of the invention with this.
By taking the protection of Internet film database (IMDb) video website user interest as an example, illustrate a kind of protection privacy of user
The recommendation method of interest preference.Fig. 1 shows the structured flowchart of the recommendation method of protection privacy of user interest preference of the present invention,
Comprising two parts, for the indirect recommendation method for accessing attack protection privacy of user and the user for preserving personal interest preference
Hold rendering method.
In a specific embodiment, Fig. 2 shows the present invention pushing away for indirect access attack protection privacy of user
The block diagram of method is recommended, wherein recommending method to include step S101:Learn the privacy of the topic according to the rating information of video topic
Correlation, the video topic are the keyword or phrase that can summarize video content in video name, and the topic privacy is related
Property the degree of correlation of other people know or other people inconvenience are known personal information is reluctant for video topic and party.
In another specific embodiment, in a certain topic of IMDb site searches, the film sensitivity rank found is
NC-17 levels (17 years old and following spectators forbid watching --- and the film of the rank is positioned people's film, and minor is firmly banned
Only watch) quantity it is more, i.e., the sensitive rank that video topic is related to is higher, then the privacy correlation of the topic is higher.
Step S103:The power of different topics in user behavior record is sexually revised using the privacy correlation of the video topic
Weight, the higher weight of topic privacy correlation is lower, similar based on the user that user behavior records according to the weight amendment
Degree.User's similarity is the close degree of user's viewing behavior record, and it is higher to record more consistent similarity;Topic privacy phase
The higher weight of closing property is lower, it is therefore intended that user's Similarity Measure based on user behavior record is depended on low privacy phase
Topic of sex is closed, that is, obscures the behavior record of user, reduces the possibility of privacy leakage.
Step S105:According to input parameter of the user's similarity as difference secret protection index Mechanism Model, weight
User's similarity is newly calculated, calculates video fraction, and select the high user of similarity as neighbours.The video fraction is
Come out using the Similarity Measure between user and all neighbours for having seen a certain video, that is, seen the phase of the neighbours of this video
Like the cumulative of degree and.Using user's similarity of low privacy as input parameter, difference secret protection index Mechanism Model is utilized
Calculated, it is therefore intended that neighbours are selected based on popular preference, therefore selected neighbours will not obtain the privacy of targeted customer
Record;This method can be effectively prevented from indirect access and attack, so as to protect privacy of user.
Step S107:The recommendation list of video is sent to user terminal, indirect access attack, low privacy will be directed to
, the video recommendations list based on popular preference be sent to user terminal.
In another specific embodiment, video recommendations order also can inadvertently reveal the interest preference of user, and
And compared with individually recommending a certain video, movie collection is recommended to need to add the order that more noise disturbances recommend output.Institute
So that to reduce leakage of the recommendation list order to user interest preference, difference secret protection index Mechanism Model also includes
Laplace mechanism, the Laplace noises that average value is zero are added on the video fraction, recalculate the video fraction,
For reducing leakage of the order to user interest preference of the recommendation list, the sensitive interest that this process is used to increase user is inclined
Good protection.
In another specific embodiment, the rendering method of the user terminal of personal interest preference is preserved, according to user
The video fraction of the personal above-mentioned recommendation of topic preference amendment is simultaneously ranked up presentation.User terminal is received for non-straight receiving
The recommendation list of attack protection privacy of user is asked, according to the personal interest preference for being stored in user terminal, to regarding in recommendation list
Frequency division number is modified and presentation of sorting.
In another specific embodiment, the rendering method for preserving the user terminal of personal interest preference can also be right
The video fraction is arranged using descending, final choice Top-k recommendation results;Search can be found out in massive video data
Frequency comes the video of predetermined quantity above.
In another specific embodiment, Fig. 3 shows the present invention for indirect access attack protection privacy of user
Recommendation apparatus block diagram, the recommendation apparatus includes:
Study module, for learning the privacy correlation of the topic according to the rating information of video topic.The video words
The keyword or phrase of video content can be summarized in entitled video name, the topic privacy correlation is video topic and works as thing
People is reluctant the degree of correlation for the personal information that other people know or other people inconvenience are known.
Correcting module, for sexually revising different topics in user behavior record according to the privacy correlation of the video topic
Weight, topic privacy correlation is higher, and weight is lower, corrects user's similarity based on user behavior record.The user is similar
Spend for the close degree of user's viewing behavior record, it is higher to record more consistent similarity;The higher weight of topic privacy correlation is more
It is low, it is therefore intended that user's Similarity Measure based on user behavior record is depended on low privacy correlation topic of sex, i.e. mould
The behavior record of user is pasted, reduces the possibility of privacy leakage.
Selecting module, for user's similarity according to the amendment as the defeated of difference secret protection index Mechanism Model
Enter parameter, recalculate user's similarity, calculate video fraction, and select the high user of similarity as neighbours;It is described
Video fraction is come out using the Similarity Measure between user and all neighbours for having seen a certain video, that is, has seen this video
Neighbours similarity cumulative and.Using user's similarity of low privacy as input parameter, referred to using difference secret protection
Number Mechanism Model is calculated, it is therefore intended that selects neighbours based on popular preference, therefore selected neighbours will not obtain target
The privacy record of user;This method can be effectively prevented from indirect access and attack, so as to protect privacy of user.
And sending module, for the recommendation list of video to be sent into the user terminal, indirect access will be directed to
Attack, low privacy, the video recommendations list based on popular preference be sent to user terminal.
In another specific embodiment, to reduce leakage of the recommendation list order to user interest preference, using adding
Make an uproar module, difference secret protection index Mechanism Model also includes Laplace mechanism, is for adding average value on video fraction
Zero Laplace noises, the video fraction is recalculated, for reducing the order of the recommendation list to user interest preference
Leakage, this process be used for increase user sensitive interest preference protection.
In another specific embodiment, the user terminal of personal interest preference is preserved, for being talked about according to individual subscriber
Inscribe the video fraction of the above-mentioned recommendation of preference amendment and be ranked up, receive hidden for indirect access attack protection user
During the recommendation list of private, according to the personal interest preference for being stored in user terminal, the video fraction in recommendation list is modified
And sort.
In another specific embodiment, Fig. 4 shows the recommendation of protection privacy of user interest preference of the present invention
The flow chart of method, a kind of recommendation method for protecting privacy of user interest preference, it is characterised in that the recommendation method includes:
Step S101:Learn the privacy correlation of the topic according to the rating information of video topic, the video topic is
Can summarize the keyword or phrase of video content in video name, the topic privacy correlation be video topic with party not
It is willing to the degree of correlation for the personal information that other people know or other people inconvenience are known.
Step S103:The power of different topics in user behavior record is sexually revised using the privacy correlation of the video topic
Weight, the higher weight of topic privacy correlation is lower, similar based on the user that user behavior records according to the weight amendment
Degree.User's similarity is the close degree of user's viewing behavior record, and it is higher to record more consistent similarity;Topic privacy phase
The higher weight of closing property is lower, it is therefore intended that user's Similarity Measure based on user behavior record is depended on low privacy phase
Topic of sex is closed, that is, obscures the behavior record of user, reduces the possibility of privacy leakage.
Step S105:According to input parameter of the user's similarity as difference secret protection index Mechanism Model, weight
User's similarity is newly calculated, calculates video fraction, and select the high user of similarity as neighbours.The video fraction is
Come out using the Similarity Measure between user and all neighbours for having seen a certain video, that is, seen the phase of the neighbours of this video
Like the cumulative of degree and.Using user's similarity of low privacy as input parameter, difference secret protection index Mechanism Model is utilized
Calculated, it is therefore intended that neighbours are selected based on popular preference, therefore selected neighbours will not obtain the privacy of targeted customer
Record;This method can be effectively prevented from indirect access and attack, so as to protect privacy of user.
Step S107:The recommendation list of video is sent to user terminal, indirect access attack, low privacy will be directed to
, the video recommendations list based on popular preference be sent to user terminal.
Step S109:The user terminal is according to video fraction described in individual subscriber topic preference and is ranked up presentation.With
Family termination receives the recommendation list for indirect access attack protection privacy of user, according to the personal interest for being stored in user terminal
Preference, is modified and presentation of sorting to the video fraction in recommendation list.
Step S111:The user terminal is arranged using descending the video fraction, final choice Top-k recommendation results,
The video that search rate comes predetermined quantity above can be found out in massive video data.
The step of method or algorithm described by the disclosure of invention can be realized in a manner of hardware or
Realized by the mode of computing device software instruction.Software instruction can be made up of corresponding software module, and software module can
To be stored on RAM memory, flash memory, ROM memory, eprom memory, eeprom memory, register, hard disk, movement
In the storage medium of hard disk, CD-ROM or any other form well known in the art.A kind of exemplary storage medium coupling
To processor, so as to enable a processor to from the read information, and information can be write to the storage medium.Certainly,
Storage medium can also be the part of processor.Processor and storage medium can be located in ASIC.In addition, the ASIC can
With in user equipment.Certainly, processor and storage medium can also be present in user equipment as discrete assembly.
Those skilled in the art are it will be appreciated that in said one or multiple examples, work(described in the invention
It is able to can be realized with hardware, software, firmware or their any combination.When implemented in software, can be by these functions
It is stored in computer-readable medium or is transmitted as one or more instructions on computer-readable medium or code.
Computer-readable medium includes computer-readable storage medium and communication media, and wherein communication media includes being easy to from a place to another
Any medium of one place transmission computer program.It is any that storage medium can be that universal or special computer can access
Usable medium.
Therefore, in another specific embodiment, it is related to a kind of computer-readable medium, it has program stored therein, when
During computing device described program, to cause:
Learn the privacy correlation of the topic according to the rating information of video topic;Utilize the privacy phase of the video topic
The weight for sexually revising different topics in user behavior record is closed, topic privacy correlation is higher, and weight is lower, according to the weight
Correct user's similarity based on user behavior record;According to user's similarity as difference secret protection index mechanism mould
The input parameter of type, user's similarity is recalculated, calculate video fraction, and select the high user of similarity as adjacent
Occupy;The recommendation list of video is sent to user terminal;Video fraction described in the user terminal foundation individual subscriber topic preference is simultaneously
It is ranked up presentation;The user terminal is arranged using descending the video fraction, final choice Top-k recommendation results.
Obviously, the above embodiment of the present invention is only intended to clearly illustrate example of the present invention, and is not pair
The restriction of embodiments of the present invention, all embodiments of the invention can be right in any combination in the case where not violating logic
For those of ordinary skill in the art, can also make on the basis of the above description it is other it is various forms of change or
Change, all embodiments can not be exhaustive here, it is every belong to that technical scheme extends out show and
Row of the change or variation being clear to still in protection scope of the present invention.
Claims (10)
- A kind of 1. recommendation method for indirect access attack protection privacy of user, it is characterised in that the recommendation method bag Include:S101:Learn the privacy correlation of the topic according to the rating information of video topic;S103:The weight of different topics in user behavior record, the words are sexually revised using the privacy correlation of the video topic It is lower to inscribe the higher weight of privacy correlation, the user's similarity recorded according to the weight amendment based on user behavior;S105:According to input parameter of the user's similarity as difference secret protection index Mechanism Model, institute is recalculated User's similarity is stated, calculates video fraction, and select the high user of similarity as neighbours;S107:The recommendation list of video is sent to user terminal.
- 2. recommendation method according to claim 1, it is characterised in that it is higher that the video topic is related to sensitive rank, institute It is higher to state privacy correlation.
- 3. recommendation method according to claim 1, it is characterised in that the difference secret protection index Mechanism Model also wraps Laplace mechanism is included, the Laplace noises that average value is zero are added on the video fraction, recalculates the video point Number, for reducing leakage of the order to user interest preference of the recommendation list.
- A kind of 4. user terminal rendering method for preserving personal interest preference, it is characterised in thatAccording to any one of individual subscriber topic preference amendment claim 1-3 video fractions recommended and it is ranked up and is in It is existing.
- 5. user terminal rendering method according to claim 4, it is characterised in that arranged using descending the video fraction Row, final choice Top-k recommendation results.
- A kind of 6. recommendation apparatus for indirect access attack protection privacy of user, it is characterised in that the recommendation apparatus bag Include:Study module, for learning the privacy correlation of the topic according to the rating information of video topic;Correcting module, for sexually revising the power of different topics in user behavior record according to the privacy correlation of the video topic Weight, the higher weight of topic privacy correlation is lower, similar based on the user that user behavior records according to the weight amendment Degree;Selecting module, for according to input parameter of the user's similarity as difference secret protection index Mechanism Model, weight User's similarity is newly calculated, calculates video fraction, and select the high user of similarity as neighbours;AndSending module, for the recommendation list of video to be sent into user terminal.
- 7. recommendation apparatus according to claim 6, it is characterised in that add module of making an uproar, the difference secret protection index machine Simulation also includes Laplace mechanism, for adding the Laplace noises that average value is zero on the video fraction, again Video fraction is calculated, for reducing leakage of the order to user interest preference of the recommendation list.
- 8. a kind of user terminal for preserving personal interest preference, it is characterised in that for according to individual subscriber topic preference amendment The video fraction of any one of claim 6 to 7 recommendation is simultaneously ranked up.
- A kind of 9. recommendation method for protecting privacy of user interest preference, it is characterised in that the recommendation method includes:S101:Learn the privacy correlation of the topic according to the rating information of video topic;S103:The weight of different topics in user behavior record is sexually revised using the privacy correlation of the video topic, topic is hidden The private higher weight of correlation is lower, the user's similarity recorded according to the weight amendment based on user behavior;S105:According to input parameter of the user's similarity as difference secret protection index Mechanism Model, institute is recalculated User's similarity is stated, calculates video fraction, and select the high user of similarity as neighbours;S107:The recommendation list of video is sent to user terminal;S109:The user terminal is according to video fraction described in individual subscriber topic preference and is ranked up presentation;S111:The user terminal is arranged using descending the video fraction, final choice Top-k recommendation results.
- 10. a kind of computer-readable medium, it has program stored therein, when computing device described program, to cause:Learn the privacy correlation of the topic according to the rating information of video topic;The weight of different topics in user behavior record is sexually revised using the privacy correlation of the video topic, topic privacy is related The higher weight of property is lower, the user's similarity recorded according to the weight amendment based on user behavior;According to input parameter of the user's similarity as difference secret protection index Mechanism Model, the user is recalculated Similarity, video fraction is calculated, and select the high user of similarity as neighbours;The recommendation list of video is sent to user terminal;The user terminal is according to video fraction described in individual subscriber topic preference amendment and is ranked up presentation;The user terminal is arranged using descending the video fraction, final choice Top-k recommendation results.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710707014.0A CN107590400A (en) | 2017-08-17 | 2017-08-17 | A kind of recommendation method and computer-readable recording medium for protecting privacy of user interest preference |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710707014.0A CN107590400A (en) | 2017-08-17 | 2017-08-17 | A kind of recommendation method and computer-readable recording medium for protecting privacy of user interest preference |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107590400A true CN107590400A (en) | 2018-01-16 |
Family
ID=61042253
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710707014.0A Pending CN107590400A (en) | 2017-08-17 | 2017-08-17 | A kind of recommendation method and computer-readable recording medium for protecting privacy of user interest preference |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107590400A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108427891A (en) * | 2018-03-12 | 2018-08-21 | 南京理工大学 | Neighborhood based on difference secret protection recommends method |
CN110134879A (en) * | 2019-03-06 | 2019-08-16 | 辽宁工业大学 | A kind of point of interest proposed algorithm based on difference secret protection |
CN110162691A (en) * | 2018-11-15 | 2019-08-23 | 腾讯科技(深圳)有限公司 | Topic recommendation, method for running, device and machinery equipment in online content service |
CN111506804A (en) * | 2020-03-18 | 2020-08-07 | 上海大犀角信息科技有限公司 | Client content recommendation system and method based on user terminal behaviors |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105138664A (en) * | 2015-09-02 | 2015-12-09 | 中国地质大学(武汉) | Big data recommendation method and system with privacy protection function |
CN106134142A (en) * | 2013-02-08 | 2016-11-16 | 汤姆逊许可公司 | Resist the privacy of the inference attack of big data |
-
2017
- 2017-08-17 CN CN201710707014.0A patent/CN107590400A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106134142A (en) * | 2013-02-08 | 2016-11-16 | 汤姆逊许可公司 | Resist the privacy of the inference attack of big data |
CN105138664A (en) * | 2015-09-02 | 2015-12-09 | 中国地质大学(武汉) | Big data recommendation method and system with privacy protection function |
Non-Patent Citations (1)
Title |
---|
TINGTING FENG等: "A Differential Private Collaborative Filtering Framework Based on Privacy-Relevance of Topics", 《2016 IEEE SYMPOSIUM ON COMPUTERS AND COMMUNICATION (ISCC)》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108427891A (en) * | 2018-03-12 | 2018-08-21 | 南京理工大学 | Neighborhood based on difference secret protection recommends method |
CN108427891B (en) * | 2018-03-12 | 2022-11-04 | 南京理工大学 | Neighborhood recommendation method based on differential privacy protection |
CN110162691A (en) * | 2018-11-15 | 2019-08-23 | 腾讯科技(深圳)有限公司 | Topic recommendation, method for running, device and machinery equipment in online content service |
CN110162691B (en) * | 2018-11-15 | 2023-05-23 | 腾讯科技(深圳)有限公司 | Topic recommendation, operation method, device and machine equipment in online content service |
CN110134879A (en) * | 2019-03-06 | 2019-08-16 | 辽宁工业大学 | A kind of point of interest proposed algorithm based on difference secret protection |
CN111506804A (en) * | 2020-03-18 | 2020-08-07 | 上海大犀角信息科技有限公司 | Client content recommendation system and method based on user terminal behaviors |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107590400A (en) | A kind of recommendation method and computer-readable recording medium for protecting privacy of user interest preference | |
US8478768B1 (en) | Privacy-preserving collaborative filtering | |
US8560528B2 (en) | Data structures for collaborative filtering systems | |
EP3283979B1 (en) | Reader application with a personalized feed and method of providing recommendations while maintaining user privacy | |
US20120036011A1 (en) | Search Personalization Using Identifiers and Authentication State | |
US9171045B2 (en) | Recommending queries according to mapping of query communities | |
Embarak | A method for solving the cold start problem in recommendation systems | |
Papaodyssefs et al. | Web identity translator: Behavioral advertising and identity privacy with wit | |
CN110134879A (en) | A kind of point of interest proposed algorithm based on difference secret protection | |
CN103544313A (en) | Data processing method and device for webpage recommendation | |
CN111753312B (en) | Data processing method, device, equipment and system | |
CN106102059B (en) | For the method and apparatus for the owner for determining hotspot | |
Wang et al. | Toward the new item problem: context-enhanced event recommendation in event-based social networks | |
CN105718573A (en) | Attention relationship extracting and annotating method in view of user interests | |
Xu et al. | Leakage-Abuse Attacks Against Forward and Backward Private Searchable Symmetric Encryption | |
US10191988B2 (en) | System and method for returning prioritized content | |
CN116346307A (en) | Statistical information-based iterative attack method for searchable symmetric encryption | |
Zhang et al. | A practical defense against attribute inference attacks in session-based recommendations | |
Guerraoui et al. | The utility and privacy effects of a click | |
Eldaw et al. | Collective suffix tree-based models for location prediction | |
Yan et al. | DynaEgo: privacy-preserving collaborative filtering recommender system based on social-aware differential privacy | |
Khayati et al. | A practical privacy-preserving targeted advertising scheme for IPTV users | |
Feng et al. | A differential private collaborative filtering framework based on privacy-relevance of topics | |
CN111708946A (en) | Personalized movie recommendation method and device and electronic equipment | |
CN106383910B (en) | Method for determining search term weight, and method and device for pushing network resources |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180116 |
|
RJ01 | Rejection of invention patent application after publication |