CN106777132A - Data processing method and device - Google Patents
Data processing method and device Download PDFInfo
- Publication number
- CN106777132A CN106777132A CN201611173120.7A CN201611173120A CN106777132A CN 106777132 A CN106777132 A CN 106777132A CN 201611173120 A CN201611173120 A CN 201611173120A CN 106777132 A CN106777132 A CN 106777132A
- Authority
- CN
- China
- Prior art keywords
- data
- user
- dimension
- quality
- label
- 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
- G06F16/958—Organisation or management of web site content, e.g. publishing, maintaining pages or automatic linking
-
- 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 invention discloses a kind of data processing method and device.Wherein, the method includes：Data are obtained according at least one dimension, wherein, dimension is used to represent and the recommended related attribute of data；The data for getting are merged according to the corresponding dimension of each data；Data after merging are ranked up according to pre-defined rule, wherein, pre-defined rule can be edited, and the data after sequence are used to be recommended to user.The purpose of flexible recommendation is realized by the present invention, and then solves the problems, such as to carry out recommending the content service provider that cannot meet for causing to require in the prior art only according to the data of user's last visit.
Description
Technical field
The present invention relates to data processing field, in particular to a kind of data processing method and device.
Background technology
In the prior art, user can obtain the content for needing by browser or APP, access a content clothes
When business business, content and service provider it is generally desirable to the data that are wished to user recommended user, or, user is interested
Data, to increase the viscosity of user.
In the prior art, the general last data for accessing of concern user, data according to user's last visit to
Family is recommended, and this processing method is only to consider on one side, and does not consider to be entered in content service provider side
Capable treatment, so as to not reach the requirement of content service provider.
For above-mentioned problem, effective solution is not yet proposed at present.
The content of the invention
A kind of data processing scheme is the embodiment of the invention provides, solve at least in the prior art only according to user
The data of last visit recommend the problem that cannot meet content service provider's requirement for causing.
A kind of one side according to embodiments of the present invention, there is provided data processing method, including：According at least one dimension
Degree obtains data, wherein, the dimension is used to represent and the recommended related attribute of data；According to the corresponding dimension of each data
Data to getting are merged；Data after merging are ranked up according to pre-defined rule, wherein, the pre-defined rule energy
Enough to be edited, the data after sequence are used to be recommended to user.
Further, according to the corresponding dimension of described each data the data for getting are merged including：To correspondence
Duplicate removal is carried out in the data of different dimensions；By different dimensions correspondence in the data after duplicate removal.
Further, after the data that will be got are ranked up according to the pre-defined rule, methods described also includes：
Judge whether the data after sequence belong to the data of needs rejecting；Needs are picked from the data for getting according to judged result
The data removed are rejected；Preserve the data after rejecting.
Further, in the case of the quality of data is included at least one dimension, methods described also includes：According to
Quality model is estimated to the quality of data, wherein, the quality model is according to corresponding with the quality at least one
Parametric configuration；Preserve the corresponding quality of the data that assessment is obtained.
Further, it is that at least one parameter includes following in the case of being issued to multiple users in the data
At least one：The time of the data publication, the multiple user are to reply quantity, the multiple user of the data to institute
State the response rate of data, the multiple user to the collection number of the data, the multiple user to the collection rate of the data,
Ratio that quantity that the multiple user is praised the data, the multiple user are praised the data, the data
Multi-medium data that length, the packet contain, the rank of the user that the data are operated, when accessing total amount, unit
Between visit capacity, the visit capacity of isolated user.
Further, the response rate is the reply quantity and the access total amount or the visit capacity of the isolated user
Ratio, and/or；The collection rate is the collection number and the access total amount or the ratio of the visit capacity of the isolated user
Value, and/or；The ratio praised is the number of being praised with the access total amount or the ratio of the visit capacity of the isolated user
Value.
Further, the quality model is obtained using existing quality data as positive example sample training.
Further, also include：The information of user is obtained, wherein, the information of the user is used to identify user's logarithm
According to concern；Acquisition of information data according to the user simultaneously sort, and are obtained from the data after sequence and are pushed away to the user
The data recommended；The data recommendation that will recommend to the user gives the user.
Further, the information of the user includes at least two class labels, wherein, first kind label is used to identify the use
Data that family accessed and/or the data for operating, Equations of The Second Kind label are used to identify the state of the user.
Further, the first kind label includes at least one label, and at least one label is according to the user
Access and/or the time of peration data determine the weight of the label, the weight as to user's recommending data according to
According to.
Other side according to embodiments of the present invention, additionally provides a kind of data processing equipment, including：Acquisition module,
For obtaining data according at least one dimension, wherein, the dimension is used to represent and the recommended related attribute of data；Merge
Module, for being merged to the data for getting according to the corresponding dimension of each data；Order module, for by after merging
Data are ranked up according to pre-defined rule, wherein, the pre-defined rule can be edited, and the data after sequence are used to enter to user
Row is recommended.
In embodiments of the present invention, data are processed by content service provider side, then the number after these treatment
According to the recommendation that will be used for data, it is achieved thereby that flexibly recommend purpose, and then solve in the prior art only according to
The data of family last visit recommend the problem that cannot meet content service provider's requirement for causing.
Brief description of the drawings
Accompanying drawing described herein is used for providing a further understanding of the present invention, constitutes the part of the application, this hair
Bright schematic description and description does not constitute inappropriate limitation of the present invention for explaining the present invention.In the accompanying drawings：
Fig. 1 is the schematic diagram of a kind of optional data processing method according to embodiments of the present invention；
Fig. 2 is the schematic diagram of a kind of optional data processing equipment according to embodiments of the present invention；
Fig. 3 is a kind of schematic diagram in the content recommendation source of alternative embodiment according to embodiments of the present invention；
Fig. 4 is the schematic diagram of commending system according to embodiments of the present invention.
Specific embodiment
In order that those skilled in the art more fully understand the present invention program, below in conjunction with the embodiment of the present invention
Accompanying drawing, is clearly and completely described to the technical scheme in the embodiment of the present invention, it is clear that described embodiment is only
The embodiment of a part of the invention, rather than whole embodiments.Based on the embodiment in the present invention, ordinary skill people
The every other embodiment that member is obtained under the premise of creative work is not made, should all belong to the model of present invention protection
Enclose.
It should be noted that term " first ", " in description and claims of this specification and above-mentioned accompanying drawing
Two " it is etc. for distinguishing similar object, without for describing specific order or precedence.It should be appreciated that so using
Data can exchange in the appropriate case, so as to embodiments of the invention described herein can with except illustrating herein or
Order beyond those of description is implemented.Additionally, term " comprising " and " having " and their any deformation, it is intended that cover
Lid is non-exclusive to be included, for example, the process, method, system, product or the equipment that contain series of steps or unit are not necessarily limited to
Those steps or unit clearly listed, but may include not list clearly or for these processes, method, product
Or other intrinsic steps of equipment or unit.
According to embodiments of the present invention, there is provided a kind of embodiment of data processing method, it is necessary to explanation, in accompanying drawing
The step of flow is illustrated can perform in the such as one group computer system of computer executable instructions, and, although
Logical order is shown in flow chart, but in some cases, can be performing shown different from order herein or retouch
The step of stating.
Fig. 1 is the flow chart of data processing method according to embodiments of the present invention, as shown in figure 1, the method is including as follows
Step：
Step S102, data are obtained according at least one dimension, wherein, dimension is used to represent recommended to data related
Attribute；
The data for getting are merged by step S104 according to the corresponding dimension of each data；
Step S106, the data after merging are ranked up according to pre-defined rule, wherein, pre-defined rule can be edited,
Data after sequence are used to be recommended to user.
In above-mentioned steps, data can be processed in server side, the data after treatment can be used for
Family is recommended, and by above-mentioned steps, can get the data of at least one dimension, then data is processed, phase
Than in prior art, to processing advantageously in the recommendation to user for data, it is achieved thereby that the flexibly purpose of recommendation, and then
To solve carried out only according to the data of user's last visit in the prior art and recommend that causes cannot meet content service provider
It is required that problem.
In the present embodiment, a kind of data processing equipment is additionally provided, as shown in Fig. 2 the device includes：
Acquisition module 22, for obtaining data according at least one dimension, wherein, dimension is used to represent recommended with data
Related attribute；
Merging module 24, for being merged to the data for getting according to the corresponding dimension of each data；
Order module 26, for the data after merging to be ranked up according to pre-defined rule, wherein, pre-defined rule can be by
Editor, the data after sequence are used to be recommended to user.
In the foregoing description, at least one dimension can be configured according to the actual needs, for example, COLLECTIDN can
As a dimension, model quality can be as dimension etc. more than threshold value.
In the case of certain is specific, same data (for example, article) are likely to occur in the middle of multiple dimensions, for example,
The data can be COLLECTIDN, while being again that model quality exceedes threshold value.In order to prevent same data from occurring, can be with
Duplicate removal is carried out when merging, however, it is desirable to multiple dimensions of the data are identified on the data.The duplicate removal can be according to
Following steps are carried out：Duplicate removal is carried out to the data corresponding to different dimensions first, then can be by different dimensions correspondence in duplicate removal
In data afterwards.Such treatment both reduces data volume, the different dimensions corresponding to a data is remained again, so as to be
Follow-up recommendation is prepared.
Sometimes, consider for certain purpose, some data can not be recommended, for example, violating state's laws or disobeying
The data of antisocial social morality.These data can now be deleted by way of blacklist, if necessary to the data ratio deleted
It is more, it is also possible to be carried out by way of white list.Alternatively, the geographical position that the embodiment of the present invention can also be according to where user
Put the recommended data of selection.In the alternative embodiment, the data that will be got are ranked up according to pre-defined rule, can sentence
Whether the data after disconnected sequence belong to the data of needs rejecting；Rejecting will be needed from the data for getting according to judged result
Data rejected；Preserve the data after rejecting.By the optional embodiment, some data can be shielded,
So as to meet the requirements.
Rejecting to data, the data that only will have harm are shielded.Certainly, it is also to need according to the matter of data
Measure to be processed.In this alternative embodiment, in the case where at least one dimension includes the quality of data, can also basis
Quality model is estimated to the quality of data, and the quality model is according at least one parametric configuration corresponding with quality；
Preserve the corresponding quality of the data that assessment is obtained.The Quality advance of the recommendation of data can be made by Evaluation Model on Quality.Matter
Amount model can rule of thumb be built, and some quality models are also to be instructed using existing quality data as positive example sample
Get.This processing mode can flexibly be adjusted to the assessment of quality.
The quality of data has a variety of embodiment modes, for example, being in the case of being issued to multiple users at least one in data
Individual parameter includes at least one of：The time of data publication, multiple users are to reply quantity, multiple users of data to data
Response rate, multiple user to collection number, multiple users of data to collection rate, multiple user's logarithm of data it is said that the number praised
Amount, multiple user's logarithm it is said that the ratio praised, the length of data, packet contain multi-medium data, data are operated
The rank of user, access total amount, unit interval visit capacity, the visit capacity of isolated user.
Used as an optional implementation method, above-mentioned parameter can also include at least one following：Response rate is reply number
Amount and access total amount or the ratio of the visit capacity of isolated user, and/or；Collection rate is collection number and access total amount or isolated user
Visit capacity ratio, and/or；The ratio praised is the ratio for being praised number and the visit capacity for accessing total amount or isolated user.
After above-mentioned data are obtained, can also be recommended according to above-mentioned data.In an optional implementation method,
It is also conceivable to reference to the information of user.In the optional implementation method, the information of user can be obtained, wherein, user's
Information is used to identify concern of the user to data；Then acquisition of information data according to user simultaneously sort, from after sequence
Obtained in data to user recommend data and by data recommendation to user.By the optional embodiment, user can be combined
Information recommended.
In an optional implementation method, the information of tag representation user can be used, for example, the information of user is included extremely
Few two class labels, wherein, first kind label is used to identify the data that user accessed and/or the data for operating, the second category
Sign the state for identifying user.When label is more, weight is may be incorporated into.For example, first kind label is included extremely
A few label, at least one label determines the weight of the label, weight according to the time of user's access and/or peration data
As the foundation to user's recommending data.
Illustrated with reference to an alternative embodiment.
In this alternative embodiment, it is possible to achieve following effect：
1. dynamic to recommend to be personalized, what each user saw is the content for being adapted to oneself and high-quality.That is " thousand people
Thousand faces ".
In order to accomplish this point, can read each user sees note record, search record, and different type is seen according to user
The number (being classified with model label herein) of content recommends her may content interested guessing the interest of user.Can be with
According to the information of user, for example, user breeds the information of child, now can be according to baby's user age, the information such as expected date of childbirth
To recommend to breed user helpful content.In addition part topic can also be recommended using collaborative filtering.
In the present embodiment, whether model is a high-quality, and the model for being worth pushing out mainly is come by quality model
Judge.The model calculated by indexs such as topic reply volume, topic length, concerned degree model whether high-quality.
2. practiced every conceivable frugality the human cost of operation, and major part work should be automatically performed by algorithm.
There is algorithm help, it is model classification (labelling) that the daily maximum work of staff is exactly, due to daily
The new post for having magnanimity is produced, in order to save manpower, in the present embodiment with machine learning automatically for model labels.Staff
Only need to a small number of typical case's model classification as training set.
3. can the different business of flexible access, including user produce content (UGC), professional production content (PGC), advertisement,
Operation activity.
A set of content ordering rule is devised in the present embodiment to meet flexibly intervention.This set ordering rule can meet
Advertiser and the demand of all departments' operation.
4. appropriate random push, it is ensured that the open content ecology of user one is presented to, without limitation in the emerging of oneself
Interesting circle the inside.
Above-mentioned several aspects are illustrated below.
Commending system content frame
There are number of different types in content recommendation source, and Fig. 3 is a kind of recommendation of alternative embodiment according to embodiments of the present invention
The schematic diagram of content sources, in figure 3, has been related to according to quality-ordered, randomly ordered, according to weight sequencing, according to delivering
Time-sequencing etc..These different sequences can be regarded as different dimensions.
For essence, the commending system content sources include four classes：
The first kind：Pure intelligent recommendation-include that, by tag computation mass note higher, collaborative filtering note is matched by state
Content, this partial content accounts for major part.
Equations of The Second Kind：Semi-intelligent is semi-artificial-on the basis of being matched by interest, can manually define some topics more has preferentially
Level.
3rd class：It is pure artificial operation that pure artificial recommendation-small volume is pushed away by force；Small volume thinks model that can be fiery, can be set to " potentiality
Note " can increase its exposure.
4th class：The content that user's self-sizing-people of user's concern delivers appears in dynamic.
Label framework
In order to machine can more accurately be recommended, new label framework includes three dimensions:
The main object of theme-content description：Theme can be identified in terms of content, can be also included in user interest.For example
User often pays close attention to the model of " cuisines " theme, in the present embodiment, also will be considered that user likes " cuisines " this theme, we
" cuisines " related topic can be recommended to user priority.
Theme is a tree structure.
Motivation, demand that type-content is delivered.Type includes " study course/knowledge ", " discussion ", " blueprint ", " question and answer " etc.
On the opportunity that, type can influence to recommend.The topic of such as " cuisines " theme has " advertisement " type and " study course " type, " cuisines "
Related " study course " would not be recommended in inappropriate time to user.
It refers to applicable which type of user to be applicable state-content, for example, what stage is the content for breeding correlation be applied to
Mother or ready-to-be mother.The state of being applicable includes breeding the stage to baby 6 years old is all of from standby pregnant, pregnancy.It is ensured that user is every
One stage can be seen suitable oneself breeds knowledge.
In the present embodiment, each content all labels comprising above three dimension, the accuracy recommended is ensured with this.
The technological frame of commending system is shown in Fig. 4, in fig. 4, is related to content sources, Quality estimation model etc.,
It is explained below.
1. many kinds source content mix
Dynamic information includes that similar microblogging (content that the people that I pays close attention to delivers), today's tops (match according to interest
Content), magazine (small daily to compile selected content), various shapes such as pregnancy period companion (and I breeds the closely bound up content of state)
Formula.
2. model Quality estimation model
The computational methods first assume that following characteristics have certain relevance with topic quality：
It is m- during PV/ final updatings to post the time
Reply number/PV
Collection number/UV
Praised number/UV
Model content-length
Whether picture is had
The average rank posted with money order receipt to be signed and returned to the sender user
Remarks：PV is page view, i.e. page access amount；UV is unique visitor independence guest access numbers
Assuming that after these features meet certain " condition ", the topic can be changed into " recommending note ".The elite that staff selects
Note/recommendation note calculates each characteristic value weight as sample, after each characteristic value linear weighted function, result of calculation is brought into
Sigmoid functions, its output valve, for example, being considered to recommend note if value is more than 0.5, is recognized as model quality less than 0.5
For be not recommend note.
Remarks：Logistic regression involved by Sigmoid functions, is a model for standard.
3. User Status label, interest tags are distinguished
To user's content recommendation be according to user with label determine, and label includes two classes, and one is interest tags, example
Such as Eight Diagrams, pet, cosmetic；One class is state tag (label related to the state that breeds), such as baby's nursing, early education, the moon
Son etc..For the mode that user plays two class labels is different.
User interest label
User interest label sees that note behavior is produced by user's, and user tag has been with the theme of the topic read
Total correlation
The interest tags of user are chosen from theme, and each label of each user has weight, the bigger table of weight
Show interested in the theme
Weight can decay with natural time, be currently the circular function curve decay of simulation, and formula is：
Δ days refers to update the day that date of this label and user stamp the date of certain label and differ for the first time last time
Number.If in x days, user does not have behavior to certain label, and the weight of the label will decay to 0.
User can update label when note behavior is seen, renewal process is the label first decayed with it, then will be current
See that the label and weight of note are added to user.
Example：Assuming that it is 45 to take x, certain user has tag1, tag2, tag3 totally 3 labels, respectively in 2015.12.13,
2016.1.14,2015.11.11 is stamped, and weight is respectively 12,10,5.In 2016.2.2, user sees note behavior, the model seen
There are 3 labels, tag2, tag3, tag4, the quality of the model is 0.6.
By updating, due to being separated by 51 days with current date, more than 45 days, weight decayed to 0 to tag1, and newly see
Note does not have this label of tag1, so the weight of tag1 is 0.Tag2 is separated by 19 days with current date, then the weight of tag2 should
ForTag3 is separated by 83 days with current date, and more than 45 days, weight decayed to 0,
Again because the note newly seen includes tag3, therefore the weight of tag3 is 0+0.6=0.6.Tag4 is newly-increased label, and weight is model
Quality 0.6.
Part tag is strong correlation, and that for example pays close attention to " mother-in-law and daughter-in-law's relation " can be considered same tag with concern " man and wife's emotion "
Tag merge principle be:Multiple tag that concern crowd highly overlaps and content is close or machine cannot be distinguished by can be closed
And.
4. User Status label
The related theme of each state, there is " applicable state " attribute, and such as " diatery supplement " is applied to the treasured of 4 months to 1 years old
It is precious.We can in proportion for she recommends the related content of the state, because every class user is to breeding the different (example of the interest of topic
The user that is such as pregnant is substantially stronger than the user's request for having baby to such topic), ratio also can be with adjustment.
5. it is simple and convenient that ordering rule is changed.
Ordering rule is controlled by way of configuring, thus want increase delete data source, change it is each come source sequence, all may be used
Realized with by way of changing configuration file, it is not necessary to change code, it is simple and convenient.
6. real-time recommendation system
Behavior can be clicked on according to active user, the point of interest of user be calculated in real time, and update the interest tags of user.Therefore
The interests change of user can be in real time caught, and is responded at once, be real-time recommendation system on a undelayed line.
Different scenes can be finely adjusted with reference to new label system, because new label system has " theme " and " type "
Two dimensions, therefore easily trickle scene can be optimized.For example equally it is use interested in " mother-in-law and daughter-in-law's Eight Diagrams "
Family, the user to liking money order receipt to be signed and returned to the sender recommends " discussion " type topic, and to liking seeing that the user that note is not returned recommends " story " type topic, it is real
The existing optimization of resources alocation.
Generally speaking, " theme " is used to judge " whether user is interested " that " type " to be used to judge " to be pushed away under what scene
Recommend (or not recommending) ".
The embodiments of the present invention are for illustration only, and the quality of embodiment is not represented.
In the above embodiment of the present invention, the description to each embodiment all emphasizes particularly on different fields, and does not have in certain embodiment
The part of detailed description, may refer to the associated description of other embodiment.
In several embodiments provided herein, it should be understood that disclosed technology contents, can be by other
Mode is realized.Wherein, device embodiment described above is only schematical, such as division of described unit, Ke Yiwei
A kind of division of logic function, can there is other dividing mode when actually realizing, such as multiple units or component can combine or
Person is desirably integrated into another system, or some features can be ignored, or does not perform.Another, shown or discussed is mutual
Between coupling or direct-coupling or communication connection can be the INDIRECT COUPLING or communication link of unit or module by some interfaces
Connect, can be electrical or other forms.
The unit that is illustrated as separating component can be or may not be it is physically separate, it is aobvious as unit
The part for showing can be or may not be physical location, you can with positioned at a place, or can also be distributed to multiple
On unit.Some or all of unit therein can be according to the actual needs selected to realize the purpose of this embodiment scheme.
In addition, during each functional unit in each embodiment of the invention can be integrated in a processing unit, it is also possible to
It is that unit is individually physically present, it is also possible to which two or more units are integrated in a unit.Above-mentioned integrated list
Unit can both be realized in the form of hardware, it would however also be possible to employ the form of SFU software functional unit is realized.
If the integrated unit is to realize in the form of SFU software functional unit and as independent production marketing or use
When, can store in a computer read/write memory medium.Based on such understanding, technical scheme is substantially
The part for being contributed to prior art in other words or all or part of the technical scheme can be in the form of software products
Embody, the computer software product is stored in a storage medium, including some instructions are used to so that a computer
Equipment (can be personal computer, server or network equipment etc.) perform each embodiment methods described of the invention whole or
Part steps.And foregoing storage medium includes：USB flash disk, read-only storage (ROM, Read-Only Memory), arbitrary access are deposited
Reservoir (RAM, Random Access Memory), mobile hard disk, magnetic disc or CD etc. are various can be with store program codes
Medium.
The above is only the preferred embodiment of the present invention, it is noted that for the ordinary skill people of the art
For member, under the premise without departing from the principles of the invention, some improvements and modifications can also be made, these improvements and modifications also should
It is considered as protection scope of the present invention.
Claims (11)
1. a kind of data processing method, it is characterised in that including：
Data are obtained according at least one dimension, wherein, the dimension is used to represent and the recommended related attribute of data；
The data for getting are merged according to the corresponding dimension of each data；
Data after merging are ranked up according to pre-defined rule, wherein, the pre-defined rule can be edited, the number after sequence
According to for being recommended to user.
2. method according to claim 1, it is characterised in that according to the corresponding dimension of described each data to getting
Data merge including：
Duplicate removal is carried out to the data corresponding to different dimensions；
By different dimensions correspondence in the data after duplicate removal.
3. method according to claim 1, it is characterised in that carried out according to the pre-defined rule in the data that will be got
After sequence, methods described also includes：
Judge whether the data after sequence belong to the data of needs rejecting；
The data for needing to reject are rejected from the data for getting according to judged result；
Preserve the data after rejecting.
4. method according to claim 1, it is characterised in that include the feelings of the quality of data at least one dimension
Under condition, methods described also includes：
The quality of data is estimated according to quality model, wherein, the quality model is according to corresponding with the quality
At least one parametric configuration；
Preserve the corresponding quality of the data that assessment is obtained.
5. method according to claim 4, it is characterised in that be in the case of being issued to multiple users in the data,
At least one parameter includes at least one of：
The time of the data publication, the multiple user are to reply quantity, the multiple user of the data to the number
According to response rate, the multiple user to the collection number of the data, the multiple user to the collection rate of the data, described
Ratio that quantity that multiple users are praised the data, the multiple user are praised the data, the length of the data,
Multi-medium data that the packet contains, the rank of the user operated to the data, access total amount, the visit of unit interval
The amount of asking, the visit capacity of isolated user.
6. method according to claim 5, it is characterised in that
The response rate is the ratio of the reply quantity and the visit capacity for accessing total amount or the isolated user, and/or；
The collection rate is the ratio of the collection number and the visit capacity for accessing total amount or the isolated user, and/or；
The ratio praised is the number of being praised with the access total amount or the ratio of the visit capacity of the isolated user.
7. method according to claim 4, it is characterised in that the quality model is to be made using existing quality data
For positive example sample training is obtained.
8. method according to any one of claim 1 to 7, it is characterised in that also include：
The information of user is obtained, wherein, the information of the user is used to identify concern of the user to data；
Acquisition of information data according to the user are simultaneously sorted, and the number recommended to the user is obtained from the data after sequence
According to；
The data recommendation that will recommend to the user gives the user.
9. method according to claim 8, it is characterised in that the information of the user includes at least two class labels, wherein,
First kind label is used to identify the data that the user accessed and/or the data for operating, and Equations of The Second Kind label is used to identify institute
State the state of user.
10. method according to claim 9, it is characterised in that the first kind label includes at least one label, described
At least one label is accessed according to the user and/or the time of peration data determines the weight of the label, and the weight makees
It is the foundation to user's recommending data.
A kind of 11. data processing equipments, it is characterised in that including：
Acquisition module, for obtaining data according at least one dimension, wherein, the dimension is used to represent be recommended phase with data
The attribute of pass；
Merging module, for being merged to the data for getting according to the corresponding dimension of each data；
Order module, for the data after merging to be ranked up according to pre-defined rule, wherein, the pre-defined rule can be compiled
Volume, the data after sequence are used to be recommended to user.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611173120.7A CN106777132A (en) | 2016-12-18 | 2016-12-18 | Data processing method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611173120.7A CN106777132A (en) | 2016-12-18 | 2016-12-18 | Data processing method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106777132A true CN106777132A (en) | 2017-05-31 |
Family
ID=58889642
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611173120.7A Pending CN106777132A (en) | 2016-12-18 | 2016-12-18 | Data processing method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106777132A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107562912A (en) * | 2017-09-12 | 2018-01-09 | 电子科技大学 | Sina weibo event recommendation method |
| CN107688959A (en) * | 2017-07-24 | 2018-02-13 | 平安科技（深圳）有限公司 | Processing method, storage medium and the server of breakpoint list |
| CN108777785A (en) * | 2018-04-26 | 2018-11-09 | 广州坚和网络科技有限公司 | A kind of method and system carrying out automatic scoring to media quality |
| CN109657130A (en) * | 2018-12-10 | 2019-04-19 | 陆少杰 | Querying method, device and the electronic equipment of automobile information |
| CN110069617A (en) * | 2017-11-09 | 2019-07-30 | 北京嘀嘀无限科技发展有限公司 | Management method, device, server and the computer readable storage medium of forum's note |
| CN112506975A (en) * | 2020-11-30 | 2021-03-16 | 橙脑教育科技(上海)有限公司 | Information recommendation method, device and system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102611785A (en) * | 2011-01-20 | 2012-07-25 | 北京邮电大学 | Personalized active news recommending service system and method for mobile phone user |
| CN104615680A (en) * | 2015-01-21 | 2015-05-13 | 广州神马移动信息科技有限公司 | Method and device for establishing web page quality model |
| US20150371347A1 (en) * | 2014-06-20 | 2015-12-24 | William E. Hayward | Estimating impact of property on individual health - property health advice |
| US20160180269A1 (en) * | 2013-08-02 | 2016-06-23 | Toshiba Mitsubishi-Electric Industrial Systems Corporation | Energy-saving-operation recommending system |
-
2016
- 2016-12-18 CN CN201611173120.7A patent/CN106777132A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102611785A (en) * | 2011-01-20 | 2012-07-25 | 北京邮电大学 | Personalized active news recommending service system and method for mobile phone user |
| US20160180269A1 (en) * | 2013-08-02 | 2016-06-23 | Toshiba Mitsubishi-Electric Industrial Systems Corporation | Energy-saving-operation recommending system |
| US20150371347A1 (en) * | 2014-06-20 | 2015-12-24 | William E. Hayward | Estimating impact of property on individual health - property health advice |
| CN104615680A (en) * | 2015-01-21 | 2015-05-13 | 广州神马移动信息科技有限公司 | Method and device for establishing web page quality model |
Non-Patent Citations (1)
| Title |
|---|
| 于敬: ""达观数据个性化推荐系统实践"", 《INFOQ—HTTPS://WWW.INFOQ.CN/ARTICLE/PERSONALIZED-RECOMMENDATION-SYSTEM-PRACTISE》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107688959A (en) * | 2017-07-24 | 2018-02-13 | 平安科技（深圳）有限公司 | Processing method, storage medium and the server of breakpoint list |
| CN107688959B (en) * | 2017-07-24 | 2020-12-29 | 平安科技（深圳）有限公司 | Breakpoint list processing method, storage medium and server |
| CN107562912A (en) * | 2017-09-12 | 2018-01-09 | 电子科技大学 | Sina weibo event recommendation method |
| CN110069617A (en) * | 2017-11-09 | 2019-07-30 | 北京嘀嘀无限科技发展有限公司 | Management method, device, server and the computer readable storage medium of forum's note |
| CN108777785A (en) * | 2018-04-26 | 2018-11-09 | 广州坚和网络科技有限公司 | A kind of method and system carrying out automatic scoring to media quality |
| CN109657130A (en) * | 2018-12-10 | 2019-04-19 | 陆少杰 | Querying method, device and the electronic equipment of automobile information |
| CN112506975A (en) * | 2020-11-30 | 2021-03-16 | 橙脑教育科技(上海)有限公司 | Information recommendation method, device and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106777132A (en) | Data processing method and device | |
| CN111008278B (en) | Content recommendation method and device | |
| CN106484801B (en) | A kind of dialogue method and its knowledge base management system of intelligent customer service robot | |
| CN103914550B (en) | Show the method and apparatus of content recommendation | |
| CN110046952A (en) | A kind of training method and device, a kind of recommended method and device of recommended models | |
| US7610255B2 (en) | Method and system for computerized searching and matching multimedia objects using emotional preference | |
| CN106803190A (en) | A kind of ad personalization supplying system and method | |
| Weber et al. | Coding the News: The role of computer code in filtering and distributing news | |
| CN109816421A (en) | Advertisement machine launches contents controlling method, device, computer equipment and storage medium | |
| CN108256119A (en) | A kind of construction method of resource recommendation model and the resource recommendation method based on the model | |
| CN106897892A (en) | Advertisement placement method and device | |
| CN107798027A (en) | A kind of heatrate Forecasting Methodology, information recommendation method and device | |
| CN109903127A (en) | A kind of group recommending method, device, storage medium and server | |
| CN107577655A (en) | Name acquiring method and apparatus | |
| CN109308332B (en) | Target user acquisition method and device and server | |
| CN111814034A (en) | Information processing method, information processing apparatus, storage medium, and electronic device | |
| US20070266048A1 (en) | System and Method for Determining Affinity Profiles for Research, Marketing, and Recommendation Systems | |
| CN108153754A (en) | A kind of data processing method and its device | |
| CN108694211A (en) | Using distribution method and device | |
| Martin-Biggers et al. | Cover lines using positive, urgent, unique language entice moms to read health communications | |
| Ntalianis et al. | Non-Gatekeeping on Social Media: A Reputation Monitoring Approach and its Application in Tourism Services. | |
| CN106844743B (en) | Emotion classification method and device for Uygur language text | |
| CN111225009B (en) | Method and device for generating information | |
| CN104281599A (en) | Method and device for recommending information to user in social network | |
| Kuş | The influence of cultural background on content marketing practices: the approach of Turkish and Lithuanian youth to the main components of content marketing |
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: 20170531 |
|
| RJ01 | Rejection of invention patent application after publication |