CN105138574A - Man-machine interaction based hybrid recommendation system used for recommending travel and leisure place - Google Patents
Man-machine interaction based hybrid recommendation system used for recommending travel and leisure place Download PDFInfo
- Publication number
- CN105138574A CN105138574A CN201510449722.XA CN201510449722A CN105138574A CN 105138574 A CN105138574 A CN 105138574A CN 201510449722 A CN201510449722 A CN 201510449722A CN 105138574 A CN105138574 A CN 105138574A
- Authority
- CN
- China
- Prior art keywords
- user
- man
- recommendation
- machine interaction
- algorithm
- 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
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
The present invention relates to the technical field of Internet travel, and particularly to a man-machine interaction based hybrid recommendation system used for recommending a travel and leisure place. The system comprises: a user interface module, used for presenting information content, acquiring a user instruction and recording a user behavior; a user model module, used for analyzing and modeling the user behavior according to historical user information and behaviors; a recommendation engine module, used for calculating a recommendation result according to a user model through a man-machine interaction scenario algorithm and a hybrid recommendation algorithm; and a recommendation object model module, used for classifying recommendation objects according to a user interest dimension in combination with recommendation results, and for recommending an object and providing recommended content. According to the man-machine interaction based hybrid recommendation system used for recommending the travel and leisure place, the recall rate and the diversity are used as evaluation indicators and used for assessing the recommendation system quality in a man-machine interaction scenario.
Description
Technical field
The present invention relates to internet tourism technical field, being specifically related to a kind of for recommending the mixing commending system based on man-machine interaction on tourism and leisure trip ground.
Background technology
Commending system (Recommendersystems) is used to provide the related advisory such as product, information, service to user, has nowadays been widely used in every commercial fields such as film, music, books.Proposed algorithm is the core of whole commending system.
Proposed algorithm mainly contains three types: based on internal memory, based on model, and a combined method.Based on the algorithm of internal memory, usually use similarity measures to calculate distance between two users or two article, then using nearest individuality as recommendation items; Be then that user, content or other relevant information are carried out statistical study based on model algorithm, create a data model in order to calculate recommendation items; Hybrid algorithm is then the proposed algorithm in conjunction with two or more type, thus obtains the performance than wherein any one is more outstanding.
Existing commending system great majority are by supposing that the user behavior of certain particular type provides recommendation results as prerequisite, and these hypothesis behaviors comprise: browse, mark and sequence independence behavior (TheSequence-IndependentManner).Specifically: navigation patterns refers to which content item user has browsed; Scoring behavior refers to that user is which content item is marked.Sequence independence behavior refers to that recommendation items is unordered and onrelevant, separate.But this kind of commending system seldom considers the impact that in actual application environment, man-machine interaction scene is brought.
Summary of the invention
Solve the problems of the technologies described above, the invention provides a kind of for recommending the mixing commending system based on man-machine interaction on tourism and leisure trip ground, by using recall rate and diversity index as evaluation criterion, employing hybrid algorithm, improves recommendation efficiency.
In order to achieve the above object, the technical solution adopted in the present invention is, a kind of for recommending the mixing commending system based on man-machine interaction on tourism and leisure trip ground, comprising:
Subscriber interface module, for presenting information content, obtains user instruction and recording user behavior,
User model module, for according to user history information and behavior, carries out cunalysis and calculation to user behavior,
Recommended engine module, adopts in conjunction with man-machine interaction reuse algorithm and mixing proposed algorithm according to user model, carries out recommendation results calculating,
Recommended model module, to classify, and recommended provides content recommendation in conjunction with recommendation results according to user interest dimension to recommended.
Further, user model module, for according to user history information and behavior, carries out cunalysis and calculation to user behavior, and user behavior comprises three steps:
Step 1: obtain N number of recommendation items,
Step 2: recommendation items and personal interest are compared,
Step 3: therefrom select a recommendation items meeting personal interest most to browse.
Recommended engine module, adopts in conjunction with man-machine interaction reuse algorithm and mixing proposed algorithm according to user model, carries out recommendation results calculating, specifically comprise the following steps:
Step 1: accept user's request,
Step 2: according to special algorithm by qualified N number of proposed recommendations to user,
Step 3: recording user is selected to be used for further recommending.
Further, described special algorithm comprises mixing proposed algorithm, based on the data characteristic browsed in the matrix of project, can be divided into three phases: the starting stage, transition period and stabilization sub stage.Three characteristics of starting stage are: browsed the matrix of project for empty (null); The project of browsing of login user is for empty; The number of entry that any user browses can not more than TR.This stage mainly uses random algorithm.
At transition stage, the matrix having browsed project has two kinds of features: browsed the matrix of project not for empty; Browsed the quantity >TR>0 of item, but recommended item lazy weight is to start kNN algorithm.In this stage, kNN algorithm is used to recommend a few items, mostly several, is recommended by random algorithm.
In the stabilization sub stage, viewed item quantity is more than TR and be enough to start kNN algorithm.This stage recommends primarily of kNN algorithm, and in order to increase diversity, random algorithm also can be used to recommend a few items simultaneously.
Further, man-machine interaction scene is calculated
legal personmachine interactive algorithm, based on man-machine interaction scene, is divided into user and system two angles.
From user perspective, man-machine interaction scene is as follows:
User, to system input request, obtains N number of content of system recommendation;
When user determines to browse content recommendation, whether the content no matter user browses is based on her interest inventory, and recommended flowsheet all can continue; If user does not like the project of system recommendation, then can exit.
From commending system angle, man-machine interaction scene is as follows:
Commending system accepts the request of user and utilizes hybrid algorithm to recommend N number of item to this user;
If user has browsed wherein one, this selection can be recorded to and browse in a matrix by commending system, otherwise Flow ends.
All users use commending system not have particular order.
The present invention is by adopting technique scheme, and compared with prior art, tool has the following advantages:
1, define " man-machine interaction " (User-RecommenderInteraction) flow process, it comprises commending system and user two class.
2, the mixing commending system based on man-machine interaction that has incremental learning function is established.The system starting stage, when there is not any man-machine intersection record, system will be recommended by random algorithm; In transition period, when namely human-machine interaction data is also very sparse, the mode that system is then mixed mutually by random algorithm and KNN algorithm is recommended; In the stabilization sub stage, after the data namely collected exceed given threshold, then KNN algorithm is mainly relied on to recommend.
3, use recall rate and diversity as evaluation index, be used for assessing the commending system quality under man-machine interaction scene.
Accompanying drawing explanation
Fig. 1 is the structural representation of embodiments of the invention.
Fig. 2 is the user behavior process flow diagram of embodiments of the invention.
Fig. 3 is the process flow diagram of the recommended engine module of embodiments of the invention.
Fig. 4 describes the scene schematic diagram of user u1-u4 and commending system interaction.
Embodiment
Now the present invention is further described with embodiment by reference to the accompanying drawings.
As a specific embodiment, as shown in Figure 1, of the present invention a kind of for recommending the mixing commending system based on man-machine interaction on tourism and leisure trip ground, comprising:
Subscriber interface module, for presenting information content, obtains user instruction and recording user behavior,
User model module, for according to user history information and behavior, carries out cunalysis and calculation to user behavior,
Recommended engine module, adopts in conjunction with man-machine interaction reuse algorithm and mixing proposed algorithm according to user model, carries out recommendation results calculating,
Recommended model module, to classify, and recommended provides content recommendation in conjunction with recommendation results according to user interest dimension to recommended.
The design concept of the present embodiment is as follows:
(1) hybrid algorithm designs in order to the interactive recall rate of equilibrium and diversity.
(2) man-machine interaction model builds based on hybrid algorithm.
(3) tuning can be carried out to algorithm by setup parameter.
Two, man-machine interaction scene
The man-machine interaction scene set forth in the present embodiment refers to a series of interactive action between login user and commending system.Represent with the simplest example, namely after logging in system by user, system can return one or more recommendation items to user, and after user selects one of them or direct shutdown system, epicycle terminates alternately.Man-machine interaction behavior is mainly divided into user behavior and commending system behavior two parts.
1, user behavior flow process,
Shown in figure 2, in the present embodiment, user can be divided into two kinds of main Types according to their feedback: if user only browses recommendation items, be called as and browse user; If user marks to recommendation items, be then referred to as the user that marks.Only use in the present embodiment and browse user as an example.
User operation process settings is three steps:
(1) N number of recommendation items of system is obtained
(2) recommendation items and personal interest are compared
(3) carrying out meeting personal interest most is therefrom selected to browse
The behavior of user can be described more intuitively in Fig. 2.Browse user for one to sign in in system and to obtain N number of recommendation items.User can obtain the item the browsed set of a candidate according to her interest and system recommendation.
When user cannot find her interested content any from recommendation items, just can lose patience and log off.The sum of the item can recommended due to system is limited, and therefore all users finally can exit.
Grey inventory (greylist) is for improving the efficiency of recommendation.Each user has oneself grey inventory.When this user does not like certain recommendation items time, this can join in the grey inventory of user by system, and system can not recommend any item in grey inventory again to this user afterwards.
When user log off, system can obtain the matrix that a user has browsed those recommendation items records.We are referred to as to browse a matrix.It should be noted that and browse the impact that a matrix is not only subject to commending system, also can be subject to the impact of some behavior of user, the time interval of such as twice navigation patterns simultaneously.
Suppose that they would not reuse this system after user exits commending system.But do not get rid of some patient especially users.They can repeatedly access and use commending system.We define such user behavior for paying a return visit (revisit).
2, commending system flow process
The flow process of recommended engine module has detailed description in figure 3.When commending system receives login and the request of user, and recommend N number of item and browse user to this.If user has browsed one of them recommendation items, the selection of user can be recorded to and browse in a matrix by system.Having browsed item number can because mutual increasing and increases, and therefore system has incremental learning (incrementallearning) function simultaneously.If user can not find the item of her interest in N number of recommendation items, just can log off.When all users exit, system stops.
Shown in figure 3, the workflow of recommended engine module has three steps:
(1) user's request is accepted
(2) according to special algorithm by qualified N number of proposed recommendations to user
(3) recording user is selected to be used for further recommending
Described special algorithm comprises mixing proposed algorithm,
In the present embodiment, hybrid algorithm of carrying is made up of random algorithm and kNN algorithm.Random algorithm is used to solve cold start-up problem also for commending system provides diversity.KNN algorithm finds the individual similar user of k by calculating cosine value.Each adjacent user can recommend a series of project.TR is the threshold value in kNN algorithm.By it, commending system can determine whether use kNN algorithm.Based on the principle browsing project matrix, the item number of browsing of user ui is | br (ui) |.We are referred to as kNN switch:
Hybrid algorithm is the core of proposed algorithm.Based on the data characteristic browsed in the matrix of project, three phases can be divided into: the starting stage, transition period and stabilization sub stage.
Three characteristics of starting stage are: browsed the matrix of project for empty (null); The project of browsing of login user is for empty; The number of entry that any user browses can not more than TR.This stage mainly uses random algorithm.
At transition stage, the matrix having browsed project has two kinds of features: browsed the matrix of project not for empty; Browsed the quantity >TR>0 of item, but recommended item lazy weight is to start kNN algorithm.In this stage, kNN algorithm is used to recommend a few items, mostly several, is recommended by random algorithm.
In the stabilization sub stage, viewed item quantity is more than TR and be enough to start kNN algorithm.This stage recommends primarily of kNN algorithm, and in order to increase diversity, random algorithm also can be used to recommend a few items simultaneously.
Algorithm one describes this increment hybrid algorithm.Input includes user ID (uid) and three designated parameter: total recommended item number (N), the proportion (RT) of random recommendation and TR.
Algorithm one: mixing proposed algorithm
This algorithm forms primarily of four steps:
Step 1: in the starting stage, has browsed in the matrix of project without any historical data.Application random algorithm solves cold start-up problem.3-6 in the corresponding algorithm of this step is capable.
Step 2:TR is used to determine whether use kNN algorithm.Calculate the item number Nb that user ui has browsed.If Nb>=TR, start kNN algorithm.8-14 in the corresponding algorithm of this step is capable.
Step 3: when after commending system even running, random algorithm excavates the new point of interest of user by being used for.15-18 in the corresponding algorithm of this step is capable.
Step 4:kNN algorithm is only itemized, even if still have part recommendation items to be based on random algorithm in transition and stabilization sub stage for recommended unit.17-23 in the corresponding algorithm of this step is capable.
By the output of this algorithm stored in internal memory.
2, man-machine interaction reuse algorithm
Man-machine interaction algorithm, based on man-machine interaction scene mentioned above, is divided into user and system two angles.
From user perspective, man-machine interaction scene is as follows:
User, to system input request, obtains N number of content of system recommendation;
When user determines to browse content recommendation, whether the content no matter user browses is based on her interest inventory, and recommended flowsheet all can continue; If user does not like the project of system recommendation, then can exit.
From commending system angle, man-machine interaction scene is as follows:
Commending system accepts the request of user and utilizes hybrid algorithm to recommend N number of item to this user;
If user has browsed wherein one, this selection can be recorded to and browse in a matrix by commending system, otherwise Flow ends.
All users use commending system not have particular order.
Algorithm two describes the flow process of interactive commending system.Input includes three designated parameter: N, RT and TR.
Algorithm two: man-machine interaction reuse algorithm
This algorithm forms primarily of three steps:
Step 1: call algorithm one by internal memory and obtain N number of recommendation items.Eighth row in the corresponding algorithm of this step;
Step 2: if user loses interest in this recommendation items, commending system will stop more recommendation.The 11st row in the corresponding algorithm of this step;
Step 3: if user is interested in this recommendation items, commending system will recording user selection and upgrade browsed a matrix.The 13rd row in the corresponding algorithm of this step.
Because number of users is limited, therefore man-machine interaction flow process finally can terminate.So the recall rate of this interaction and diversity also can be calculated simultaneously.18-19 in the corresponding algorithm of this step is capable.
Three, example
Cite an actual example this commending system is described.
Table one illustrates the example of an interests matrix.4 users and 8 films are had in table.In the starting stage, browse the matrix of project for empty.Parameter N, RT, TR and k in this example corresponding assignment are respectively 3,0.25,1 and 1
Table two: user interest matrix
Fig. 4 illustrates interaction and is recommended in an example in interaction scenarios.Different user random sequence signs in in commending system.We suppose that the order logged in is (u3, u2, u1, u4).
In Fig. 4, (a) describes the scene of user u3 and commending system interaction.Because she is the user that system first logs in, without any adjacent user, therefore system recommends her some films based on random algorithm.We suppose the interactive process always total three-wheel of user u3 and commending system.The recommendation inventory of the first round is (m5, m6, m2).This user is interested is m6, m2}, and m5 is not in her interested collection the inside.It is because this is the film of her first her hobby of coupling that user browses m6.Therefore, 1 is set as in the relevant position of browsing the project in the matrix of project from 0.And m5 is added into inside her grey inventory.The second recommendation inventory of taking turns is (m7, m8, m4).User only have selected and browses m4, and { therefore m7, m8} are added into inside her grey inventory.The recommendation inventory of third round is (m3, m1, m2).This time user selects to browse m2, and { m3, m1} are added in her grey inventory.Because system does not have more film can recommend this user, she will exit commending system.Finally, user u3 has browsed film { m2, m4, m6}, her grey inventory has been { m1, m3, m5, m7, m8}.
In Fig. 4, (b) describes the scene of user u2 and commending system interaction.Interactive process has three-wheel.The recommendation of the first round is based on random algorithm equally.The recommendation inventory of the first round is (m8, m4, m1).User selects to browse m4, and she uninterested m8 joins in her grey inventory by system.Here, due to TR=1, k=1, therefore user u3 just becomes the neighboring user of user u2.Take turns in the recommendation to user u2 second, { m2, m6} have kNN algorithm to recommend to film, and m7 then recommends based on random algorithm.The second recommendation inventory of taking turns is { m2, m6, m7}.But because user u2 only selects to browse m6, therefore (m7, m2) is added in her grey inventory.Because do not have other films can recommend by kNN algorithm, the recommendation of third round be random algorithm.This is taken turns and recommends inventory to be { m1, m3, m5}.User u2 can exit commending system because recommend in inventory without any the film enough allowing her like in third round.
In Fig. 4, (c) describes the scene of user u1 and commending system interaction.This interactive process has six and takes turns.The recommendation of the first round is based on random algorithm equally.The recommendation inventory of the first round is (m3, m1, m7).User selects to browse m2, and she uninterested m3 is added grey inventory by system.Ensuing one to take turns recommendation be based on kNN and random algorithm.At this moment, be Nk=[N* (1-RT)]=[3* (1-0.25)]=2 by the item number that kNN algorithm is recommended.And the item number of recommending at random is Nrd=N-Nk=1.User u3 becomes her neighboring user and recommends { m4, m6}.These two films and the random m7 recommended constitute last recommendation inventory (m7, m4, m6).User have selected and browses m4, and m7 is added in her grey inventory.All based on random algorithm in following several recommendation taken turns, because her another two neighboring user u2 and u3 cannot recommend any New cinema.
In Fig. 4, (d) describes the scene of user u4 and commending system interaction.Owing to recommending film in inventory, { m7, m6, m4} not in her interested inventory, therefore this interactive process only has and takes turns.
When not having more any active ues in commending system when in use, final project matrix of browsing is shown in the diagram in (e).From the viewpoint of this example, if the interest of user is more extensive, system can better be recommended.Finally, interactive recall rate is by being calculated as ir (U, T)=11/15=0.73.The item number of successful referral is 6; Therefore interactive diversity is id (U, T)=6/8=0.75.Due to user login sequence and select the recommendation that finally have impact on afterwards, what therefore we can say this user and commending system facilitates the perfect of machine Active Learning function alternately, is intelligentized.
Four, the present embodiment is summed up
The present embodiment mainly sets forth three partial contents:
1, define " man-machine interaction " (User-RecommenderInteraction) flow process, it comprises commending system and user two class.
2, the mixing commending system based on man-machine interaction that has incremental learning function is established.
The system starting stage, when there is not any man-machine intersection record, system will be recommended by random algorithm; In transition period, when namely human-machine interaction data is also very sparse, the mode that system is then mixed mutually by random algorithm and KNN algorithm is recommended; In the stabilization sub stage, after the data namely collected exceed given threshold, then KNN algorithm is mainly relied on to recommend.
3, use recall rate and diversity as evaluation index, be used for assessing the commending system quality under man-machine interaction scene.
Recall rate: the ratio being relevant documentation numbers all in the relevant documentation number and document library retrieved, measurement be the recall ratio of commending system.
Diversity: describe recommendation results in commending system and can cover the different interest worlds of user.General by recommending row
in tablearticle between any two dissimilarity calculate, and between article, more dissimilar then diversity is better.
Each recommendation taken turns is in fact the constraint condition as subsequent recommendation, the selection of user or behavior under system log (SYSLOG), then applies to during next round recommends to calculate, thus content recommendation is more and more close to the users interest.
The meaning of the present embodiment mainly contains following three aspects:
(1) define closer to the man-machine interaction behavior under actual application environment, there is stronger Practical significance.
(2) based on the accuracy of error criterion, as root-mean-square error (RMSE) and mean absolute error (MAE), and be not suitable for evaluating interactive commending system.Therefore, in practical application scene, be more suitable for using recall rate and diversity index to evaluate the mixing commending system based on man-machine interaction described by the present embodiment.
(3) core that hybrid algorithm is interactive commending system is proposed.
In famous MovieLens data set experiment (http://www.movielens.org/), the experimental result performance of the hybrid algorithm of the present embodiment is better than wherein any one single algorithm.
Although specifically show in conjunction with preferred embodiment and describe the present invention; but those skilled in the art should be understood that; not departing from the spirit and scope of the present invention that appended claims limits; can make a variety of changes the present invention in the form and details, be protection scope of the present invention.
Claims (5)
1., for recommending the mixing commending system based on man-machine interaction on tourism and leisure trip ground, it is characterized in that: comprising:
Subscriber interface module, for presenting information content, obtains user instruction and recording user behavior,
User model module, for according to user history information and behavior, carries out cunalysis and calculation to user behavior,
Recommended engine module, adopts in conjunction with man-machine interaction reuse algorithm and mixing proposed algorithm according to user model, carries out recommendation results calculating,
Recommended model module, to classify to recommended according to user interest dimension in conjunction with recommendation results, and provides content recommendation for recommended.
2. according to claim 1 for recommending the mixing commending system based on man-machine interaction on tourism and leisure trip ground, it is characterized in that: user behavior comprises three steps:
Step 1: obtain N number of recommendation items,
Step 2: recommendation items and personal interest are compared,
Step 3: therefrom select a recommendation items meeting personal interest most to browse.
3. according to claim 1 for recommending the mixing commending system based on man-machine interaction on tourism and leisure trip ground, it is characterized in that: the recommendation results of recommended engine module calculates, and specifically comprises the following steps:
Step 1: accept user's request,
Step 2: according to special algorithm by qualified N number of proposed recommendations to user,
Step 3: recording user is selected to be used for further recommending.
4. according to claim 1 for recommending the mixing commending system based on man-machine interaction on tourism and leisure trip ground, it is characterized in that: described special algorithm comprises mixing proposed algorithm, based on the data characteristic browsed in the matrix of project, be divided into three phases: the starting stage, transition period and stabilization sub stage.
5. according to claim 1 for recommending the mixing commending system based on man-machine interaction on tourism and leisure trip ground, it is characterized in that: man-machine interaction reuse algorithm.Man-machine interaction algorithm, based on man-machine interaction scene, is divided into user and system two angles.
From user perspective, man-machine interaction scene is as follows:
User, to system input request, obtains N number of content of system recommendation;
When user determines to browse content recommendation, whether the content no matter user browses is based on her interest inventory, and recommended flowsheet all can continue; If user does not like the project of system recommendation, then can exit;
From commending system angle, man-machine interaction scene is as follows;
Commending system accepts the request of user and utilizes hybrid algorithm to recommend N number of item to this user;
If user has browsed wherein one, this selection can be recorded to and browse in a matrix by commending system, otherwise Flow ends;
All users use commending system not have particular order.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510449722.XA CN105138574A (en) | 2015-07-28 | 2015-07-28 | Man-machine interaction based hybrid recommendation system used for recommending travel and leisure place |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510449722.XA CN105138574A (en) | 2015-07-28 | 2015-07-28 | Man-machine interaction based hybrid recommendation system used for recommending travel and leisure place |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105138574A true CN105138574A (en) | 2015-12-09 |
Family
ID=54723923
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510449722.XA Pending CN105138574A (en) | 2015-07-28 | 2015-07-28 | Man-machine interaction based hybrid recommendation system used for recommending travel and leisure place |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105138574A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106202252A (en) * | 2016-06-29 | 2016-12-07 | 厦门趣处网络科技有限公司 | Method, system are recommended in a kind of trip analyzed based on user emotion |
| CN107358471A (en) * | 2017-07-17 | 2017-11-17 | 桂林电子科技大学 | A kind of tourist resources based on visit behavior recommends method and system |
| CN115018474A (en) * | 2022-08-03 | 2022-09-06 | 山东美丽乡村云计算有限公司 | Text and travel consumption heat degree analysis method based on big data |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050193002A1 (en) * | 2004-02-26 | 2005-09-01 | Yahoo! Inc. | Method and system for generating recommendations |
| CN101894351A (en) * | 2010-08-09 | 2010-11-24 | 北京邮电大学 | Multi-agent based tour multimedia information personalized service system |
| CN102750318A (en) * | 2005-11-08 | 2012-10-24 | 真实网络公司 | Method of providing content items |
| CN103309866A (en) * | 2012-03-09 | 2013-09-18 | 华为技术有限公司 | Method and device for generating recommendation result |
| CN103678672A (en) * | 2013-12-25 | 2014-03-26 | 北京中兴通软件科技股份有限公司 | Method for recommending information |
| CN103793537A (en) * | 2014-03-04 | 2014-05-14 | 南京大学 | System for recommending individual music based on multi-dimensional time series analysis and achieving method of system |
-
2015
- 2015-07-28 CN CN201510449722.XA patent/CN105138574A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050193002A1 (en) * | 2004-02-26 | 2005-09-01 | Yahoo! Inc. | Method and system for generating recommendations |
| CN102750318A (en) * | 2005-11-08 | 2012-10-24 | 真实网络公司 | Method of providing content items |
| CN101894351A (en) * | 2010-08-09 | 2010-11-24 | 北京邮电大学 | Multi-agent based tour multimedia information personalized service system |
| CN103309866A (en) * | 2012-03-09 | 2013-09-18 | 华为技术有限公司 | Method and device for generating recommendation result |
| CN103678672A (en) * | 2013-12-25 | 2014-03-26 | 北京中兴通软件科技股份有限公司 | Method for recommending information |
| CN103793537A (en) * | 2014-03-04 | 2014-05-14 | 南京大学 | System for recommending individual music based on multi-dimensional time series analysis and achieving method of system |
Non-Patent Citations (2)
| Title |
|---|
| HENG-RU ZHANG等: "Interactive hybrid recommendation with granule selection", 《IEEE INTERNATIONAL CONFERENCE ON GRANULAR COMPUTING》 * |
| 九天科技: "《精通微信应用、技巧与营销》", 30 June 2014 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106202252A (en) * | 2016-06-29 | 2016-12-07 | 厦门趣处网络科技有限公司 | Method, system are recommended in a kind of trip analyzed based on user emotion |
| CN107358471A (en) * | 2017-07-17 | 2017-11-17 | 桂林电子科技大学 | A kind of tourist resources based on visit behavior recommends method and system |
| CN107358471B (en) * | 2017-07-17 | 2020-05-22 | 桂林电子科技大学 | Tour resource recommendation method and system based on tour behaviors |
| CN115018474A (en) * | 2022-08-03 | 2022-09-06 | 山东美丽乡村云计算有限公司 | Text and travel consumption heat degree analysis method based on big data |
| CN115018474B (en) * | 2022-08-03 | 2022-11-08 | 山东美丽乡村云计算有限公司 | Text and travel consumption heat degree analysis method based on big data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Rafailidis et al. | Modeling users preference dynamics and side information in recommender systems | |
| Neidhardt et al. | Eliciting the users' unknown preferences | |
| CN104680453B (en) | Course based on student's attribute recommends method and system | |
| Xie et al. | Breaking out of the box of recommendations: from items to packages | |
| Stella et al. | Influence of augmented humans in online interactions during voting events | |
| Feng et al. | Recommendation via user's personality and social contextual | |
| CN111061946A (en) | Scenario content recommendation method and device, electronic equipment and storage medium | |
| CN113688167A (en) | Deep interest capture model construction method and device based on deep interest network | |
| CN110321291A (en) | Test cases intelligent extraction system and method | |
| CN105210061A (en) | Tagged search result maintenance | |
| CN107292648A (en) | A kind of user behavior analysis method and device | |
| Parvin et al. | Using best-worst scaling method to examine consumers’ value preferences: A multidimensional perspective | |
| Li et al. | Warehouse vis: A visual analytics approach to facilitating warehouse location selection for business districts | |
| Koed Madsen | Web-visions as controversy-lenses | |
| El-Korany | Integrated expert recommendation model for online communities | |
| CN104008204A (en) | Dynamic multi-dimensional context awareness film recommending system and achieving method thereof | |
| WO2017203672A1 (en) | Item recommendation method, item recommendation program, and item recommendation apparatus | |
| CN105138574A (en) | Man-machine interaction based hybrid recommendation system used for recommending travel and leisure place | |
| US10409830B2 (en) | System for facet expansion | |
| CN116501979A (en) | Information recommendation method, information recommendation device, computer equipment and computer readable storage medium | |
| CN116049549A (en) | Activity recommendation method based on multi-granularity feature fusion | |
| CN114861079A (en) | Collaborative filtering recommendation method and system fusing commodity features | |
| Soltani et al. | Identity matching in social media platforms | |
| Zhang et al. | A collaborative filtering recommendation system by unifying user similarity and item similarity | |
| Gu et al. | Developing a scholar classification scheme from publication patterns in academic science: A cluster analysis approach |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20151209 |