User's portrait depicting method, device, medium and computing device
Technical field
Embodiments of the present invention are related to computer realm, more specifically, embodiments of the present invention are related to a kind of user
Portrait depicting method, a kind of family portrait characterization device, computer storage media and computing device.
Background technology
Background that this section is intended to provide an explanation of the embodiments of the present invention set forth in the claims or context.Herein
Description recognizes it is the prior art not because not being included in this part.
User portrait be virtual representations of the real user in internet, it be it is a kind of establish a series of truthful datas it
On user model.Usually, it can be realized based on user's portrait and the mesh for understanding user is removed by user's investigation and data mining
's.Specifically, can they be divided by difference according to the difference of the behavior of user, viewpoint and target of interest etc.
Type, characteristic feature is then extracted from each type, can also applied to systems such as recommendation-search-advertisement-push
Label characteristics are further taken out, applied to all kinds of marketing systems.
At present, occurred some users portrait method in the relevant technologies.
But during present inventive concept is realized, inventor has found that at least there are following defects in the relevant technologies:It is existing
Some user's portraits can only reflect feature of the user in specific area, it is impossible to reflect feature of the user in other field, for example, having
User's portrait can be only applied to the communications field, some then can be only applied to internet arena, and applied to internet arena
User portrait can only reflect this incidence relation of family's attribute of user.
For the above problem in the relevant technologies, effective solution is not proposed also at present.
Invention content
Since existing user portrait can only reflect feature of the user in specific area, it is impossible to reflect user in other field
Feature, cause user draw a portrait adaptation field it is very limited.
Therefore existing user's portrait cannot be applied in wider array of field, this is very bothersome.
Thus, it is also very desirable to a kind of improved user's portrait depicting method, so that it can be applied in wider array of field.
In the present context, embodiments of the present invention are intended to provide a kind of improved user's portrait depicting method and dress
It puts.
In the first aspect of embodiment of the present invention, a kind of user's portrait depicting method is provided, including:For target
User obtains to portray the initial data that the user of above-mentioned target user draws a portrait;Based on above-mentioned initial data, above-mentioned mesh is determined
The user characteristics that mark user shows in the multiple first default dimensions;And by above-mentioned target user above-mentioned multiple
The user characteristics showed in one default dimension portray user's portrait of above-mentioned target user.
In one embodiment of the invention, during the user's portrait for portraying above-mentioned target user:By above-mentioned mesh
The user characteristics that mark user shows at least one of above-mentioned multiple first default dimensions dimension carry out labeling, obtain
To corresponding user characteristics label;And by above-mentioned user characteristics label and remaining user characteristics, portray above-mentioned target user
User portrait, wherein, remaining above-mentioned user characteristics include:Above-mentioned target user is in above-mentioned multiple first default dimensions except upper
State the user characteristics showed in other dimensions outside at least one dimension.
In another embodiment of the invention, by above-mentioned target user in above-mentioned multiple first default dimensions at least one
The user characteristics showed in a dimension carry out labeling, obtain corresponding user characteristics label and include:For it is above-mentioned at least
Each dimension in one dimension determines the classification for the user characteristics that above-mentioned target user shows in the dimension;To upper
It states the user characteristics that target user shows in the dimension and carries out semantic analysis, obtain the user characteristics pair with above-mentioned classification
The semanteme answered;And be associated above-mentioned classification and semanteme corresponding with the user characteristics of above-mentioned classification, to obtain the dimension
User characteristics label on degree.
In yet another embodiment of the present invention, by above-mentioned target user in above-mentioned multiple first default dimensions at least
The user characteristics showed in one dimension carry out labeling, obtain corresponding user characteristics label and include:For it is above-mentioned extremely
Each dimension in a few dimension, the user characteristics showed in the dimension to above-mentioned target user are counted, are obtained
To corresponding statistical result;Obtain externally input additional information;And it based on above-mentioned statistical result and said additional information, gives
The user characteristics that above-mentioned target user shows in the dimension are tagged, to obtain the user characteristics mark in the dimension
Label.
In yet another embodiment of the present invention, based on above-mentioned initial data, determine above-mentioned target user multiple first
The user characteristics showed in default dimension include:It is relevant more that user's portrait is extracted and portrayed from above-mentioned initial data
Kind target data;And based on above-mentioned plurality of target data, determine above-mentioned target user in above-mentioned multiple first default dimensions
The user characteristics showed, wherein, a kind of target data corresponds to a first default dimension.
In yet another embodiment of the present invention, based on above-mentioned plurality of target data, determine above-mentioned target user above-mentioned
The user characteristics showed in multiple first default dimensions include:Each target data in above-mentioned plurality of target data is pressed
It is translated according to preset rules, obtains corresponding various structures data;To each structure in above-mentioned various structures data
Change data and carry out data analysis, obtain a variety of data objects, wherein, a kind of data object corresponds to a first default dimension;With
And based on above-mentioned a variety of data objects, determine the user that above-mentioned target user shows in above-mentioned multiple first default dimensions
Feature.
In yet another embodiment of the present invention, above-mentioned plurality of target data include at least two in data below:Base
Plinth data, original recorded data, factual data, statistical data and prediction data.
It is above-mentioned in yet another embodiment of the present invention in the case where above-mentioned target data includes above-mentioned basic data
Method further includes:Above-mentioned basic data is parsed, obtains one or more of following parsing data:Attribute data,
Character data, associated data, terminal data and log-on data, wherein, the first default dimension packet corresponding with above-mentioned basic data
It includes:Attribute dimensions, role's dimension, relevant dimension, terminal dimension and registration dimension;Include in the parsing data of above-mentioned basic data
In the case of above-mentioned attribute data, the attribute for being portrayed above-mentioned target user in above-mentioned attribute dimensions based on above-mentioned attribute data is drawn
Picture;In the case where the parsing data of above-mentioned basic data include above-mentioned character data, based on above-mentioned character data at above-mentioned angle
Role's portrait of above-mentioned target user is portrayed on color dimension;Include above-mentioned associated data in the parsing data of above-mentioned basic data
In the case of, the association for being portrayed above-mentioned target user on above-mentioned relevant dimension based on above-mentioned associated data is drawn a portrait;On above-mentioned basis
In the case that the parsing data of data include above-mentioned terminal data, portrayed in above-mentioned terminal dimension based on above-mentioned terminal data
State the terminal portrait of target user;And in the case where the parsing data of above-mentioned basic data include above-mentioned log-on data, base
The registration portrait of above-mentioned target user is portrayed in above-mentioned registration dimension in above-mentioned log-on data.
In yet another embodiment of the present invention in the case where above-mentioned target data includes above-mentioned original recorded data,
The above method further includes:Above-mentioned original recorded data is parsed, obtains one or more of following parsing data:It visits
Ask data, operation data, the first consumption data, the second consumption data and feedback data, wherein, with above-mentioned original recorded data pair
The the first default dimension answered includes:Access dimension, operation dimension, the first consumption dimension, the second consumption dimension and feedback dimension;
In the case that the parsing data of above-mentioned original recorded data include above-mentioned access data, based on above-mentioned access data in above-mentioned access
The access portrait of above-mentioned target user is portrayed in dimension;Include aforesaid operations data in the parsing data of above-mentioned original recorded data
In the case of, the operation for being portrayed above-mentioned target user in aforesaid operations dimension based on aforesaid operations data is drawn a portrait;In above-mentioned original
In the case that the parsing data for the record data that begin include above-mentioned first consumption data, based on above-mentioned first consumption data above-mentioned the
The first consumption portrait of above-mentioned target user is portrayed in one consumption dimension;Include in the parsing data of above-mentioned original recorded data upper
In the case of stating the second consumption data, above-mentioned target is portrayed in the above-mentioned second consumption dimension based on above-mentioned second consumption data and is used
The second consumption portrait at family;And in the case where the parsing data of above-mentioned original recorded data include above-mentioned feedback data, base
The feedback portrait of above-mentioned target user is portrayed in above-mentioned feedback dimension in above-mentioned feedback data.
It is above-mentioned in yet another embodiment of the present invention in the case where above-mentioned target data includes above-mentioned factual data
Method further includes:Above-mentioned factual data is parsed, obtains one or more of following parsing data:Disappear for M times earliest
Take data, last n times consumption data, preceding L consumption data, consumption data in preset time period, wherein, with above-mentioned factual data
Corresponding first default dimension includes:M consumption dimension, last n times consumption dimension, preceding L consumption dimension, preset time earliest
Consumption dimension in section;In the case where the parsing data of above-mentioned factual data include above-mentioned earliest M consumption data, based on above-mentioned
Earliest M consumption data portrays the earliest M consumption portrait of above-mentioned target user in above-mentioned earliest M consumption dimension;Upper
The parsing data of factual data are stated including in the case of above-mentioned last n times consumption data, existing based on above-mentioned last n times consumption data
The last n times consumption portrait of above-mentioned target user is portrayed in above-mentioned last n times consumption dimension;In the parsing number of above-mentioned factual data
In the case of including above-mentioned preceding L consumption data, carved in above-mentioned preceding L consumption dimension based on above-mentioned preceding L consumption data
Draw the preceding L consumption portrait for stating target user;And include above-mentioned preset time period in the parsing data of above-mentioned factual data
In the case of interior consumption data, consumed in above-mentioned preset time period based on consumption data in above-mentioned preset time period and carved in dimension
Draw consumption portrait in the preset time period for stating target user.
It is above-mentioned in yet another embodiment of the present invention in the case where above-mentioned target data includes above-mentioned statistical data
Method further includes:Above-mentioned statistical data is parsed, obtains one or more of following parsing data:Preference data,
Geodata, regular data, wherein, the first default dimension corresponding with above-mentioned statistical data includes:Preference dimension, geography dimensionality
With regular dimension;In the case where the parsing data of above-mentioned statistical data include above-mentioned preference data, based on above-mentioned preference data
The preference portrait of above-mentioned target user is portrayed in above-mentioned preference dimension;Include in the parsing data of above-mentioned statistical data above-mentioned
In the case of managing data, the geographical portrait of above-mentioned target user is portrayed on above-mentioned geography dimensionality based on above-mentioned geodata;With
And in the case where the parsing data of above-mentioned statistical data include above-mentioned regular data, based on above-mentioned regular data in above-mentioned rule
The rule portrait of above-mentioned target user is portrayed in dimension.
It is above-mentioned in yet another embodiment of the present invention in the case where above-mentioned target data includes above-mentioned prediction data
Method further includes:Above-mentioned prediction data is parsed, obtains one or more of following parsing data:Probabilistic forecasting number
According to, risk profile data, tactful prediction data, wherein, the first default dimension corresponding with above-mentioned prediction data includes:Probability is pre-
Survey dimension, risk profile dimension and strategy prediction dimension;Include above-mentioned probabilistic forecasting number in the parsing data of above-mentioned prediction data
In the case of, the probabilistic forecasting of above-mentioned target user is portrayed in above-mentioned probabilistic forecasting dimension based on above-mentioned probabilistic forecasting data
Portrait;In the case where the parsing data of above-mentioned prediction data include above-mentioned risk profile data, based on above-mentioned risk profile number
It draws a portrait according to the risk profile that above-mentioned target user is portrayed in above-mentioned risk profile dimension;And the parsing in above-mentioned prediction data
In the case that data include above-mentioned tactful prediction data, portrayed in above-mentioned strategy prediction dimension based on above-mentioned tactful prediction data
The strategy prediction portrait of above-mentioned target user.
In the second aspect of embodiment of the present invention, a kind of computer readable storage medium is provided, is stored thereon with
Computer executable instructions are used to implement any one of above-described embodiment above-mentioned user when above-metioned instruction is performed by processing module
Portrait depicting method.
In the third aspect of embodiment of the present invention, a kind of user's portrait characterization device is provided, including:Obtain mould
Block for being directed to target user, obtains to portray the initial data that the user of above-mentioned target user draws a portrait;Determining module is used
In based on above-mentioned initial data, the user characteristics that above-mentioned target user shows in the multiple first default dimensions are determined;With
And first portray module, it is special for the user that is showed in above-mentioned multiple first default dimensions by above-mentioned target user
Sign portrays user's portrait of above-mentioned target user.
In one embodiment of the invention, above device further includes:Labeling module, for portraying above-mentioned target use
During user's portrait at family, by above-mentioned target user at least one of above-mentioned multiple first default dimensions dimension upper table
Reveal the user characteristics come and carry out labeling, obtain corresponding user characteristics label;And second portray module, for passing through
User characteristics label and remaining user characteristics are stated, portrays user's portrait of above-mentioned target user, wherein, remaining above-mentioned user characteristics
Including:Above-mentioned target user shows in other dimensions in above-mentioned multiple first default dimensions in addition to above-mentioned at least one dimension
User characteristics out.
In another embodiment of the invention, above-mentioned labeling module includes:First determination unit, for be directed to it is above-mentioned extremely
Each dimension in a few dimension determines the classification for the user characteristics that above-mentioned target user shows in the dimension;Point
Analyse unit, semantic analysis carried out for the user characteristics that are showed in the dimension to above-mentioned target user, obtain with it is above-mentioned
The corresponding semanteme of user characteristics of classification;And first generation unit, for by above-mentioned classification and user with above-mentioned classification
The corresponding semanteme of feature is associated, to obtain the user characteristics label in the dimension.
In yet another embodiment of the present invention, above-mentioned labeling module includes:Statistic unit, for be directed to it is above-mentioned at least
Each dimension in one dimension, the user characteristics showed in the dimension to above-mentioned target user are counted, are obtained
Corresponding statistical result;Acquiring unit, for obtaining externally input additional information;And second generation unit, for being based on
Above-mentioned statistical result and said additional information stamp mark to the user characteristics that above-mentioned target user shows in the dimension
Label, to obtain the user characteristics label in the dimension.
In yet another embodiment of the present invention, above-mentioned determining module includes:Extraction unit, for from above-mentioned initial data
In extract and portray user and draw a portrait relevant plurality of target data;And second determination unit, for being based on above-mentioned a variety of mesh
Data are marked, determine the user characteristics that above-mentioned target user shows in above-mentioned multiple first default dimensions, wherein, a kind of mesh
It marks data and corresponds to a first default dimension.
In yet another embodiment of the present invention, above-mentioned second determination unit includes:Subelement is translated, for will be above-mentioned more
Each target data in kind target data is translated according to preset rules, obtains corresponding various structures data;Analysis
Subelement for carrying out data analysis to each structural data in above-mentioned various structures data, obtains a variety of data pair
As, wherein, a kind of data object corresponds to a first default dimension;And determination subelement, for being based on above-mentioned a variety of data
Object determines the user characteristics that above-mentioned target user shows in above-mentioned multiple first default dimensions.
In yet another embodiment of the present invention, above-mentioned plurality of target data include at least two in data below:Base
Plinth data, original recorded data, factual data, statistical data and prediction data.
In yet another embodiment of the present invention, the second determination unit is additionally operable to include above-mentioned base in above-mentioned target data
In the case of plinth data, above-mentioned basic data is parsed, obtains one or more of following parsing data:Attribute number
According to, character data, associated data, terminal data and log-on data, wherein, the first default dimension corresponding with above-mentioned basic data
Including:Attribute dimensions, role's dimension, relevant dimension, terminal dimension and registration dimension;And above-mentioned first portray module and also use
In:In the case where the parsing data of above-mentioned basic data include above-mentioned attribute data, based on above-mentioned attribute data in above-mentioned category
Property dimension on portray above-mentioned target user attribute portrait;Include above-mentioned character data in the parsing data of above-mentioned basic data
In the case of, the role for being portrayed above-mentioned target user in above-mentioned role's dimension based on above-mentioned character data is drawn a portrait;On above-mentioned basis
In the case that the parsing data of data include above-mentioned associated data, portrayed on above-mentioned relevant dimension based on above-mentioned associated data
State the association portrait of target user;In the case where the parsing data of above-mentioned basic data include above-mentioned terminal data, based on upper
State the terminal portrait that terminal data portrays above-mentioned target user in above-mentioned terminal dimension;And the parsing in above-mentioned basic data
In the case that data include above-mentioned log-on data, above-mentioned target user is portrayed in above-mentioned registration dimension based on above-mentioned log-on data
Registration portrait.
In yet another embodiment of the present invention, the second determination unit is additionally operable to include above-mentioned original in above-mentioned target data
Begin in the case of record data, to parse above-mentioned original recorded data, obtain a kind of or more in following parsing data
Kind:Data, operation data, the first consumption data, the second consumption data and feedback data are accessed, wherein, with above-mentioned original record
The corresponding first default dimension of data includes:Access dimension, operation dimension, the first consumption dimension, the second consumption dimension and feedback
Dimension;And above-mentioned first portray module and be additionally operable to:Include above-mentioned access data in the parsing data of above-mentioned original recorded data
In the case of, the access for being portrayed above-mentioned target user in above-mentioned access dimension based on above-mentioned access data is drawn a portrait;In above-mentioned original
In the case that the parsing data for the record data that begin include aforesaid operations data, based on aforesaid operations data in aforesaid operations dimension
Portray the operation portrait of above-mentioned target user;Include above-mentioned first consumption data in the parsing data of above-mentioned original recorded data
In the case of, the first of above-mentioned target user is portrayed in the above-mentioned first consumption dimension consume picture based on above-mentioned first consumption data
Picture;In the case where the parsing data of above-mentioned original recorded data include above-mentioned second consumption data, based on the above-mentioned second consumption
Data portray the second consumption portrait of above-mentioned target user in the above-mentioned second consumption dimension;And in above-mentioned original recorded data
Parsing data include above-mentioned feedback data in the case of, above-mentioned mesh is portrayed in above-mentioned feedback dimension based on above-mentioned feedback data
Mark the feedback portrait of user.
In yet another embodiment of the present invention, the second determination unit is additionally operable to include above-mentioned thing in above-mentioned target data
In the case of real data, above-mentioned factual data is parsed, obtains one or more of following parsing data:Earliest M
Secondary consumption data, last n times consumption data, preceding L consumption data, consumption data in preset time period, wherein, with the above-mentioned fact
The corresponding first default dimension of data includes:M consumption dimension, last n times are consumed dimension, preceding L consumption dimension, are preset earliest
Consumption dimension in period;And above-mentioned first portray module and be additionally operable to:Include in the parsing data of above-mentioned factual data above-mentioned
In the case of earliest M consumption data, portrayed in above-mentioned earliest M consumption dimension based on above-mentioned earliest M consumption data
State the earliest M consumption portrait of target user;Include above-mentioned last n times consumption data in the parsing data of above-mentioned factual data
In the case of, the last N that above-mentioned target user is portrayed in dimension is consumed in above-mentioned last n times based on above-mentioned last n times consumption data
Secondary consumption portrait;In the case where the parsing data of above-mentioned factual data include above-mentioned preceding L consumption data, based on above-mentioned preceding L
Secondary consumption data portrays the preceding L consumption portrait of above-mentioned target user in above-mentioned preceding L consumption dimension;And in the above-mentioned fact
The parsing data of data are included in above-mentioned preset time period in the case of consumption data, based on consuming number in above-mentioned preset time period
Portrait is consumed according to being consumed in above-mentioned preset time period in the preset time period that above-mentioned target user is portrayed in dimension.
In yet another embodiment of the present invention, the second determination unit is additionally operable to include above-mentioned system in above-mentioned target data
In the case of counting, above-mentioned statistical data is parsed, obtains one or more of following parsing data:Preference number
According to, geodata, regular data, wherein, the first default dimension corresponding with above-mentioned statistical data includes:Preference dimension, geography
Dimension and regular dimension;And above-mentioned first portray module and be additionally operable to:Include in the parsing data of above-mentioned statistical data above-mentioned inclined
In the case of good data, the preference portrait of above-mentioned target user is portrayed in above-mentioned preference dimension based on above-mentioned preference data;
In the case that the parsing data of above-mentioned statistical data include above-mentioned geodata, based on above-mentioned geodata in above-mentioned geography dimensionality
On portray the geographical portrait of above-mentioned target user;And include the feelings of above-mentioned regular data in the parsing data of above-mentioned statistical data
Under condition, the rule portrait of above-mentioned target user is portrayed in above-mentioned regular dimension based on above-mentioned regular data.
In yet another embodiment of the present invention, the second determination unit is additionally operable to include in above-mentioned target data above-mentioned pre-
In the case of measured data, above-mentioned prediction data is parsed, obtains one or more of following parsing data:Probability is pre-
Measured data, risk profile data, tactful prediction data, wherein, the first default dimension corresponding with above-mentioned prediction data includes:Generally
Rate prediction dimension, risk profile dimension and strategy prediction dimension;And above-mentioned first portray module and be additionally operable to:In above-mentioned prediction number
According to parsing data include above-mentioned probabilistic forecasting data in the case of, tieed up based on above-mentioned probabilistic forecasting data in above-mentioned probabilistic forecasting
The probabilistic forecasting portrait of above-mentioned target user is portrayed on degree;Include above-mentioned risk profile number in the parsing data of above-mentioned prediction data
In the case of, the risk profile of above-mentioned target user is portrayed in above-mentioned risk profile dimension based on above-mentioned risk profile data
Portrait;It is pre- based on above-mentioned strategy and in the case where the parsing data of above-mentioned prediction data include above-mentioned tactful prediction data
Measured data portrays the strategy prediction portrait of above-mentioned target user in above-mentioned strategy prediction dimension.
In the fourth aspect of embodiment of the present invention, a kind of computing device is provided, including:Processing module;And it deposits
Module is stored up, is stored thereon with executable instruction, above-metioned instruction is used to implement above-described embodiment when being performed by above-mentioned processing module
Any one of described in user draw a portrait depicting method.
According to the user of embodiment of the present invention portrait depicting method and device, due to that can reflect user using multiple
Different characteristic latitude on portray any one user portrait mode, it is thus possible to realize that multi-angle stage construction portrays use
The purpose of family portrait, it is possible to overcome at least partly in the relevant technologies due to only being closed from communication level or family's attribute
The defects of application field that connection relationship level portrays user's portrait and user is caused to draw a portrait is limited, so as to significantly extend user
The adaptation field of portrait.
Description of the drawings
Detailed description below, above-mentioned and other mesh of exemplary embodiment of the invention are read by reference to attached drawing
, feature and advantage will become prone to understand.In the accompanying drawings, if showing the present invention's by way of example rather than limitation
Dry embodiment, wherein:
Fig. 1, which is schematically shown, according to embodiments of the present invention to be suitable for user and draws a portrait the exemplary of depicting method and device
Application scenarios;
Fig. 2 schematically shows the flow charts of user's portrait depicting method according to embodiments of the present invention;
Fig. 3 A schematically show the schematic diagram for portraying user's portrait according to embodiments of the present invention;
Fig. 3 B schematically show the flow chart of labeling user characteristics according to embodiments of the present invention;
Fig. 3 C schematically show the flow chart of labeling user characteristics according to another embodiment of the present invention;
Fig. 3 D schematically show the flow chart of determining user characteristics according to embodiments of the present invention;
Fig. 3 E schematically show the flow chart of determining user characteristics according to another embodiment of the present invention;
Fig. 3 F schematically show the schematic diagram of the composition of target data according to embodiments of the present invention;
Fig. 3 G schematically show the schematic diagram of the composition of basic data according to embodiments of the present invention;
Fig. 3 H schematically show the schematic diagram of the composition of original recorded data according to embodiments of the present invention;
Fig. 3 I schematically show the block diagram of the composition of the fact that according to embodiments of the present invention data;And
Fig. 3 J schematically show the schematic diagram of the composition of statistical data according to embodiments of the present invention;
Fig. 3 K schematically show the schematic diagram of the composition of prediction data according to embodiments of the present invention;
The block diagram for characterization device of drawing a portrait Fig. 4 schematically shows user according to embodiments of the present invention;
Fig. 5 schematically shows the schematic diagram of computer readable storage medium product according to embodiments of the present invention;With
And
Fig. 6 schematically shows the block diagram of computing device according to embodiments of the present invention.
In the accompanying drawings, identical or corresponding label represents identical or corresponding part.
Specific embodiment
The principle and spirit of the invention are described below with reference to several illustrative embodiments.It should be appreciated that provide this
A little embodiments are not with any just for the sake of better understood when those skilled in the art and then realize the present invention
Mode limits the scope of the invention.On the contrary, these embodiments are provided so that the disclosure is more thorough and complete, and energy
It is enough that the scope of the present disclosure is completely communicated to those skilled in the art.
One skilled in the art will appreciate that embodiments of the present invention can be implemented as a kind of system, device, equipment, method
Or computer program product.Therefore, the disclosure can be with specific implementation is as follows, i.e.,:Complete hardware, complete software
The form that (including firmware, resident software, microcode etc.) or hardware and software combine.
According to the embodiment of the present invention, it is proposed that a kind of user's portrait depicting method, medium, device and computing device.
Herein, it is to be understood that any number of elements in attached drawing is used to example and unrestricted and any
Name is only used for distinguishing, without any restrictions meaning.
Below with reference to several representative embodiments of the present invention, the principle and spirit of the invention are illustrated in detail.
Summary of the invention
During the embodiment of the present invention is realized, inventor has found:Existing user's portrait can only reflect user in spy
Determine the feature in field, it is impossible to reflect feature of the user in other field, for example, some users portrait can be only applied to communication neck
Domain, some then can be only applied to internet arena, and user's portrait applied to internet arena can only reflect the family of user
This incidence relation of front yard attribute.
Embodiments of the present invention provide a kind of user's portrait depicting method and device, this method include:For target
User obtains to portray the initial data that the user of target user draws a portrait;Based on initial data, determine target user multiple
The user characteristics showed in first default dimension;And it is showed in the multiple first default dimensions by target user
User characteristics, portray target user user portrait.Since the present invention is using in multiple different characteristics that can reflect user
Latitude on portray any one user portrait mode, it is thus possible to realize multi-angle stage construction portray user portrait mesh
, it is possible to overcome at least partly in the relevant technologies due to only from communication level or family's Attribute Association relationship level
The defects of application field portrayed user's portrait and user is caused to draw a portrait is limited, so as to significantly extend the adaptation of user's portrait
Field.
After the basic principle for describing the present invention, lower mask body introduces the various nonrestrictive embodiment party of the present invention
Formula.
Application scenarios overview
User's portrait depicting method of the embodiment of the present invention and the exemplary application field of device are elaborated referring initially to Fig. 1
Scape.
After internet gradually steps into the big data epoch, the inevitably behavior to product and user brings a series of change
Become, all behaviors of user should be " visualization " in face of product.With the further investigation and application of big data technology, enterprise
The absorbed point of industry increasingly focus on how using big data come for precision marketing service, and then deeply excavate potential business valency
Value.
User's portrait is the basic means of production currently without product under product or line on opinion line, by user's portrait to
Family carry out it is comprehensive portrays, can help enterprise more comprehensively, gain more insight into the feature of user, so as to adjust marketing plan
More matched personal marketing scheme is provided slightly or automatically using the processing capacity of server, increase conclusion of the business probability so as to reach
Purpose.
The purposes of user's portrait is very extensive, as shown in Figure 1, user's portrait 110 can be used for ad system 120, push
System 130, commending system 140 and search system 150.
Specifically, user's portrait primarily can be used for:(1) precision marketing:User group is cut into thinner granularity, it is auxiliary
With push note and mail and the means such as advertising campaign are done, different users is shown loving care for, retrieve, is encouraged;(2) number
According to application:Such as commending system, ad system and search system, can draw a portrait according to user data custom strategies or by user
Characteristic carries out machine learning and promotes rate of product transformation or user experience;(3) product analysis:The data warehouse of service layer,
All kinds of labels are the natural element of multidimensional analysis, thus can inquire related data therein by Data Query Platform.
To sum up, million people raises you, not as good as a people understands you, veritably understand user, can just obtain user, can just occupy
The market of bigger.As it can be seen that the importance of user's portrait is self-evident.User's portrait is the requisite data storehouse of a good product
Important model in library and AI systems.
It should be understood that the application scenarios of the present embodiment are only schematical, the model of the disclosure can not be limited or reduced
It encloses.
Illustrative methods
It is described below with reference to Fig. 2 according to the user of exemplary embodiment of the invention portrait depicting method.It should be noted that
, above application scene, which is for only for ease of, to be understood spirit and principles of the present invention and shows, embodiments of the present invention exist
It is unrestricted in this respect.On the contrary, embodiments of the present invention can be applied to applicable any scene.
An embodiment of the present invention provides a kind of user portrait depicting methods.
Fig. 2 schematically shows the flow charts for depicting method of being drawn a portrait according to the user of embodiment of the present invention.Such as Fig. 2 institutes
Show, this method can include following operation:
S210 is operated, for target user, obtains to portray the initial data that the user of target user draws a portrait;
S220 is operated, based on initial data, determines the user that target user shows in the multiple first default dimensions
Feature;And
S230 is operated, the user characteristics showed in the multiple first default dimensions by target user portray target
User's portrait of user.
It should be noted that target user of the present invention can be arbitrary user.Further, for any user and
Speech, initial data can be all data related with the user, the data including online and offline.In addition, obtain original number
According to mode can include a variety of, for example crawled from network, by various interfaces from corresponding database, memory (as schemed
Can be file memory 301 and data storage 302 shown in 3A) in read, etc., do not limit herein.
For target user, after its initial data is obtained, the initial data of acquisition can be parsed, and to parsing
As a result classify, different classes of initial data is made to correspond to different default dimensions, and then to right in each default dimension
The initial data of type is answered to be analyzed, the user characteristics showed with determining this kind of data.
After user characteristics that Various types of data is showed are determined, it is possible to portray user in each default dimension
User draw a portrait.
For example, as shown in Figure 3A, the user that portrays for portraying user's portrait draws a portrait system 305, can be stored from file
The initial data of user is obtained in device 301 and data storage 302, and the initial data based on acquisition performs corresponding user and draws
As portraying flow, after user's portrait is portrayed and finished, user can be drawn a portrait for marketing system 303 and other online services
In 304.
Through the embodiment of the present invention, due to arbitrary using being portrayed on the latitude of multiple different characteristics that can reflect user
The mode of the portrait of one user, it is thus possible to portray the purpose of user's portrait with realizing multi-angle stage construction, it is possible to extremely
Partially overcome in the relevant technologies due to only portraying user's portrait from communication level or family's Attribute Association relationship level
And the defects of causing the application field of user's portrait limited, so as to significantly extend the adaptation field of user's portrait.
Depicting method of drawing a portrait below in conjunction with Fig. 3 B~Fig. 3 K to user shown in Fig. 2 is described in detail.
Generally, due to the user characteristics that user shows in each first default dimension are usually relatively abstract, more superfluous
It is long, complicated, in order to overcome these defects, the user characteristics that user shows in each default dimension can be carried out simplifying place
Reason, simplify to be abstracted, tediously long, complicated user characteristics embody,.For example, label can be made, make its labeling.
Wherein it is possible to which the user characteristics in each default dimension are carried out labeling, (user in some default dimensions is special
Sign can not also carry out labeling and directly use), in labeling flow, the technologies such as cluster or artificial mark side can be used
Formula concludes the user in certain default dimensions with similar users feature, is easy to use.In addition, the mark of user characteristics
Labelization mode can at least include mode 1 and mode 2 as described below.
Mode 1, as shown in Figure 3B, the user characteristics that can be showed to every class initial data cluster, and tie
Neural LISP program LISP NLP is closed, each type is formed and is described.
Specifically, as a kind of optional embodiment, above-mentioned user draws a portrait depicting method in the user for portraying target user
It can also include during portrait:The target user is showed at least one of the multiple first default dimension dimension
User characteristics out carry out labeling, obtain corresponding user characteristics label;And pass through the user characteristics label and remaining
User characteristics portray user's portrait of the target user, wherein, which includes:The target user is multiple
The user characteristics showed in other dimensions in first default dimension in addition at least one dimension.
It should be noted that during user's portrait is portrayed, labeling user characteristics are a preferred embodiments, for
It for multiple first default dimensions, can select its whole labeling, can also select a portion labeling, specifically
How to select, can determine, not limit herein according to actual needs.
Further, as a kind of optional embodiment, by the target user in the multiple first default dimension extremely
The user characteristics showed in a few dimension carry out labeling, obtain corresponding user characteristics label and include:For this extremely
Each dimension in a few dimension determines the classification for the user characteristics that the target user shows in the dimension;To this
The user characteristics that target user shows in the dimension carry out semantic analysis, obtain corresponding with the user characteristics of the category
It is semantic;And be associated the category and semanteme corresponding with the user characteristics of the category, to obtain the use in the dimension
Family feature tag.
Specifically, for the user characteristics in any default dimension, user spy can be obtained by technological means such as clusters
The classification of sign, then the semanteme by such user characteristics of the technical limit spacings such as natural language processing, so as to classify and the reason of the mankind
De-association gets up, this is preset user characteristics in dimension and practical significance associates, forms label.
Mode 2 as shown in Figure 3 C, can carry out sorting-out in statistics to the user characteristics that every class initial data is showed,
And plus artificial acceptable environment, each type is formed and is described.
Specifically, as a kind of optional embodiment, by the target user in the multiple first default dimension at least
The user characteristics showed in one dimension carry out labeling, and obtaining corresponding user characteristics label can include:For this
Each dimension at least one dimension, the user characteristics showed in the dimension to the target user are counted, are obtained
To corresponding statistical result;Obtain externally input additional information;And based on the statistical result and the additional information, give the mesh
The user characteristics that mark user shows in the dimension are tagged, to obtain the user characteristics label in the dimension.
It, can be by (i.e. structuring) data of labeling and non-labeling after by corresponding user characteristics labeling
User's (such as online service, marketing system) is transported to, saves as corresponding form, for its use.Wherein, these data supply line
Upper service is mainly used for automated process flow when using, as carried out machine learning, into push on line, recommend, search, fixed
Valency etc., these data for marketing system using when be mainly used for such as crm system, push systems etc..
As shown in Figure 3D, as a kind of optional embodiment, operation S220 is based on the initial data, determines the target user
The user characteristics showed in the multiple first default dimensions can include:
S221 is operated, extracts and portray the relevant plurality of target data of user's portrait from the initial data;And
S222 is operated, based on the plurality of target data, determines that the target user shows in the multiple first default dimension
User characteristics out, wherein, a kind of target data corresponds to a first default dimension.
Specifically, the above-mentioned flow of the embodiment of the present invention can be refined as the progress of following three step:Parsing:For from original
The relevant data of beginning extracting data translate into the data language of structuring, and for example data flow is become by unserializing can
The key-value key-value pairs of identification, to facilitate subsequent processing;Processing:To the user data that upper step parses, carry out further
Analyzing and processing, such as statistics are excavated, prediction;Structuring:By upper step, treated that data are assembled into centainly by the content of specific data
The data shape of form can only blame, in order to subsequently make in order to be inserted into database or write corresponding file
It is convenient with (being checked on line or under line).
As shown in FIGURE 3 E, as a kind of optional embodiment, operation S222 is based on the plurality of target data, determines the target
The user characteristics that user shows in the multiple first default dimension can include:
S2221 is operated, each target data in the plurality of target data is translated according to preset rules, is obtained pair
The various structures data answered;
S2222 is operated, data analysis is carried out to each structural data in the various structures data, obtains a variety of numbers
According to object, wherein, a kind of data object corresponds to a first default dimension;And
S2223 is operated, based on a variety of data objects, determines that the target user shows in the multiple first default dimension
User characteristics out.
The depicting method it should be noted that user proposed by the present invention draws a portrait, by above-mentioned flow by original and user
Associated data configuration is drawn a portrait into corresponding user, and so-called user's portrait herein refers to the data of structuring, to represent user
The user characteristics showed in various default dimensions form close copy file, can be loaded into marketing system, advertisement system
System, supplying system, commending system and search system can be used for strategy use, and use method provided by the invention
The user's portrait come is depicted with all readable characteristic of machine and people, can also be used as user model use.
In addition, in embodiments of the present invention, the data source of user's portrait generally includes data warehouse, can generally distinguish
For data file or the database of structuring, the former such as file, hdfs, document electronic document, the latter such as mysql, ddb,
The databases such as hive, redis, tair, ncr.During embodiment, first initial data is taken out from data warehouse, then from initial data
Middle extraction related data is used as target data in order to further build user's portrait.Wherein, initial data substantially can be with
Divide into online basic data, behavioral data in product, user content preference data, customer transaction data this four classes.
Specifically, these files can be imported into computing unit, calculates or extract each pre- according to specific rule
If the user characteristics of dimension, specific rules will introduce below.
Through the embodiment of the present invention, user's portrait systematically can be robustly portrayed, so as to provide abundant user
Data facilitate automated system or automanual operation system on line to use, personalized recommendation and crowd are facilitated to divide, Neng Gouda
To the purpose of precision marketing.In addition, the information that the user's portrait depicted by scheme provided by the invention includes is more comprehensive,
Subjective conjecture can be reduced to the greatest extent in the running of product, walk close to user, understand their real demand, so as to know how more
It is different types of user service well.
Further, as a kind of optional embodiment, which includes at least two in data below:Base
Plinth data, original recorded data, factual data, statistical data and prediction data.As illustrated in Figure 3 F, for the sake of comprehensively, herein
Target data can include basic data, original recorded data, factual data, statistical data and prediction data simultaneously.
Specifically, degree is enriched according to specific product orientation and product data, adjustment target can be taken the circumstances into consideration in specific implementation
Content that data are included and the sequence for obtaining these contents.Wherein:Basic data, also known as basic information, it is to obtain
With processing user information in itself, here in the information of the user person of including the use of itself and platform virtual portrait role information;
Original recorded data, it is acquisition and handles the data record of user most original in interiors of products and Related product, including
But behavior record is not limited to, such as purchasing history;Factual data, also known as Fact Model, it is acquisition and processing with the fact
User information based on dimension (state) obtains the user characteristics under the description by a kind of significant description;Statistics
Data, also known as statistical model, it is acquisition and handles user's portrait based on statistical information, is excavated by statistical means latent
User's feature, for example taste, preference, often go to place etc.;Prediction data, also known as prediction model, it is acquisition and place
Manage some related datas, and pass through model, typically machine learning model predicts some features of the user, as risk estimates,
Taste is estimated etc., also including the user information lacked in basic data, original recorded data, factual data, statistical data, such as
Occupation, age, gender prediction etc..
Further, as a kind of optional embodiment, in the case where the target data includes the basic data, this method
It can also include:The basic data is parsed, obtains one or more of following parsing data:Attribute data, angle
(as shown in Figure 3 G, basic data can include attribute data, angle simultaneously for chromatic number evidence, associated data, terminal data and log-on data
Chromatic number evidence, associated data, terminal data and log-on data), wherein, the first default dimension corresponding with the basic data includes:
Attribute dimensions, role's dimension, relevant dimension, terminal dimension and registration dimension;Include the category in the parsing data of the basic data
Property data in the case of, portrayed in the attribute dimensions based on the attribute data target user attribute portrait;On the basis
In the case that the parsing data of data include the character data, target use is portrayed in role's dimension based on the character data
Role's portrait at family;In the case where the parsing data of the basic data include the associated data, based on the associated data at this
The association portrait of the target user is portrayed on relevant dimension;Include the situation of the terminal data in the parsing data of the basic data
Under, the terminal for being portrayed the target user in the terminal dimension based on the terminal data is drawn a portrait;And the solution in the basic data
Data are analysed including in the case of the log-on data, portraying the registration of the target user in the registration dimension based on the log-on data
Portrait.
It should be noted that different types of data in above-mentioned basic data, for generating the different piece of user's portrait,
In the different piece for using different types of basic data generation user's portrait, the present invention does not limit its genesis sequence.
Specifically, attribute data includes but not limited to one or more of characteristic herein:Identifier (such as account,
User id, equipment id, cookieid etc.), age of user, user's gender, user's native place, common cell-phone number, primary account number, birth
Days, constellation, user's educational background, user's personality, user nationality, the growth age of user's religion, user in this product etc.;Angle
Chromatic number is according to including but not limited to one or more of characteristic:Role in standing, stand outer role, occupational role, family role,
Community role etc.;Associated data includes but not limited to one or more of characteristic:User whether have baby (if any,
Further include record child's age, weight, gender), user for Whom buys on behalf, user holds vehicle, user's real estate, user company
(i.e. corporate site, company's industry), user's income level;Terminal data includes but not limited to one or more of characteristic
According to:The terminal parameter used, containing cpu, memory, screen, system, brand, Time To Market etc.;Log-on data include but be not limited to
A kind of lower or various features data:Information when registration such as register channel, logon mode, contact method, the time, read it is emerging
Interest, life interest etc., registration portrait here further include user's information for actively filling in and change in the product after the completion of registration.
It further,, should in the case where the target data includes the original recorded data as a kind of optional embodiment
Method further includes:The original recorded data is parsed, obtains one or more of following parsing data:Access number
According to, (as shown in figure 3h, original recorded data can wrap for operation data, the first consumption data, the second consumption data and feedback data
Include while access data, operation data, the first consumption data, the second consumption data and feedback data), wherein, with the original note
The corresponding first default dimension of record data includes:Access dimension, operation dimension, the first consumption dimension, the second consumption dimension and anti-
Present dimension;In the case where the parsing data of the original recorded data include the access data, based on the access data in the visit
Ask the access portrait that the target user is portrayed in dimension;Include the feelings of the operation data in the parsing data of the original recorded data
Under condition, the operation portrait of the target user is portrayed in the operation dimension based on the operation data;In the original recorded data
Data are parsed including in the case of first consumption data, this is portrayed in the first consumption dimension based on first consumption data
The first consumption portrait of target user;In the case where the parsing data of the original recorded data include second consumption data,
The second consumption portrait of the target user is portrayed in the second consumption dimension based on second consumption data;It is and original at this
The parsing data of data are recorded including in the case of the feedback data, the mesh is portrayed in the feedback dimension based on the feedback data
Mark the feedback portrait of user.
It should be noted that different types of data in above-mentioned original recorded data, for generating the difference of user's portrait
Part, in the different piece for using different types of original recorded data generation user's portrait, the present invention is to its genesis sequence
It does not limit.
Specifically, original recorded data is obtained and handles herein to include but not limited to from access, operation, consumption, feedback, note
5 aspects of record obtain data, to complete portraying for certain customers portrait.Here it accesses, operate, consumption, feeding back, pair of record
Element is liked, element includes one or more of " commodity, article, classification, brand, album, activity " etc., is described in detail such as
Under.
Wherein, data are accessed to include being not limited to following one or more features data:Element in the product accesses trace,
The displaying trace of user stops trace;Operation data includes but not limited to following one or more features data:In the product
Element is browsed, is thumbed up plus purchases, collects, buying, sharing, forming a team, search behavior trace;First consumption data include be not limited to
Next or multiple characteristics:The information of real consumption the reason of including consumption, channel, source, financial details, preferential makes
Use situation;Second consumption data, for generating feedback portrait, including being not limited to following one or more features data:Customer service
Feedback content, comment content, scoring content, negative-feedback content, content of loseing interest in, complaint content;Feedback data, for generating
Record portrait, including being not limited to following one or more features data:Current consumption total value, currently in way commodity, current etc.
Current state value to be paid.
Further, as a kind of optional embodiment, in the case where the target data includes the factual data, this method
It further includes:The factual data is parsed, obtains one or more of following parsing data:Earliest M consumption data,
Last n times consumption data, preceding L consumption data, (as shown in fig. 31, factual data can be same for consumption data in preset time period
When include earliest M consumption data, last n times consumption data, preceding L consumption data, consumption data in preset time period),
In, the first default dimension corresponding with the factual data includes:M consumption dimension, last n times consume dimension, disappear for first L times earliest
Take consumption dimension in dimension, preset time period;Include the situation of the earliest M consumption data in the parsing data of the factual data
Under, the earliest M consumption for being portrayed the target user in the earliest M consumption dimension based on the earliest M consumption data is drawn a portrait;
In the case where the parsing data of the factual data include the last n times consumption data, based on the last n times consumption data at this
The last n times consumption portrait of the target user is portrayed in last n times consumption dimension;Including in the parsing data of the factual data should
In the case of preceding L consumption data, portray the target user's in the preceding L consumption dimension based on the preceding L consumption data
Preceding L consumption portrait;And include in the preset time period in the case of consumption data in the parsing data of the factual data, base
The preset time period that the target user is portrayed in dimension is consumed in the preset time period in consumption data in the preset time period
Interior consumption portrait.
It should be noted that different types of data in above-mentioned factual data, for generating the different piece of user's portrait,
In the different piece for using different types of original recorded data generation user's portrait, the present invention does not limit its genesis sequence
It is fixed.
Specifically, it obtains and is handled in factual data flow herein, need to including at least FirstN/LastN/TopN/
Any one in tetra- models of TimeN or multiple models carry out user characteristics extraction, including one or more in each model
A project, specifically this four to establish model step (in no particular order) as follows.
Generate FirstN models:Record and processing user after the equipment registration time of preceding n times purchase, the amount of money, classification,
Preferential dynamics, number of packages, each classification order distribution, each classification amount of money distribution, each classification commodity number, preceding n times using payment accounts,
Identity information, consignee's information;Content, the tendency of preceding n times comment;The article that preceding n times are collected, thumbed up, paying close attention to.
Generate LastN models:Record and processing user after the equipment registration in current time, nearest n times purchase
Time, the amount of money, classification, preferential dynamics, number of packages, the distribution of each classification order, the distribution of each classification amount of money, each classification commodity number, most
Nearly n times use payment accounts, identity information, consignee's information;Content, the tendency of nearest n times comment;Nearest n times collection, point
The article praise, paid close attention to.
Generate TopN models:Record and processing user after the equipment registration in current time, the order numbers of TopN,
TopN Brand Buyings goods amount, TopN countries purchase order number, TopN countries purchase goods amount, TopN frequently pay a bill by branch
Number, TopN consignee's information (containing address mark, name mark, administrative region), N is parameter, represents frequency, be typically chosen 10,
20、50、100。
Generate TimeN models:TimeN is the statistical data recorded in a time window, and the time, (Time) was generally comprised
10 minutes, 60 minutes, 4 hours, 12 hours, 24 hours, the same day, last week, last month, this month, this month, current season, 180 days, this year
Deng description, within the period of specific time N, record and one or more following user behavior project of processing:Purchase is real to pay gold
Volume, purchase number of packages, order numbers, consumption number of times, consumption range, click volume, amount of collection plus purchase amount, purchase volume, user are inclined to classification
Good, user is to Brang Preference etc..
Further, as a kind of optional embodiment, in the case where the target data includes the statistical data, this method
It further includes:The statistical data is parsed, obtains one or more of following parsing data:Preference data, geographical number
According to, regular data (as shown in figure 3j, statistical data can include preference data, geodata, regular data simultaneously), wherein,
The first default dimension corresponding with the statistical data includes:Preference dimension, geography dimensionality and regular dimension;In the statistical data
Data are parsed including in the case of the preference data, the inclined of the target user is portrayed in the preference dimension based on the preference data
Good portrait;In the case where the parsing data of the statistical data include the geodata, based on the geodata in the geographical dimension
The geographical portrait of the target user is portrayed on degree;And include the situation of the regular data in the parsing data of the statistical data
Under, the rule for being portrayed the target user in the rule dimension based on the regular data is drawn a portrait.
It should be noted that different types of data in above-mentioned statistical data, for generating the different piece of user's portrait,
In the different piece for using different types of original recorded data generation user's portrait, the present invention does not limit its genesis sequence
It is fixed.
Specifically, it obtains and is handled in statistical model flow herein, including processing preference pattern, geographic model and regular mould
One or more in type, further, each model can include time parameter, can obtain including real-time model in this way
With the model data of the different times window such as long-term model, these models are specifically:
Handle user preferences modeling:Including processing user's classification preference, Brang Preference, price preference, color preference, style
Preference, equity preference, service preferences, time preference are obtained by each model according to the weight calculation clicked and purchase distribution is different
It arrives.
Handle user's geographic model:Geographic properties including calculating user, including GPS coordinate, IP network address, shopping position
It is at county level not etc. to put custom, metropolitan district.
User policy model is handled, including handling one or more following project:A user price sensitivity models,
For example the average price that nearest 60 days each item are purchased first is calculated, it is every then to calculate user's purchase according to average price
The average discount of a classification;B user's purchasing power model goes out user couple according to the item price rating calculation marked in object model
Purchasing power model of the average price grade of classification as user;C discount coupon degrees of dependence, i.e. user to discount coupon and promotion according to
Bad degree;D subscriber lifecycle, for example be sunk into sleep and (have within such as nearly 90 days purchase, nearly 60 days without purchase), for example actively (be such as divided into
High frequency, intermediate frequency, low frequency three classes cluster to obtain according to user's purchase number of days of nearly 60 days), for example be lost in (such as nearly 90 days without purchase,
Once bought, do not counted temporarily), for example register and do not buy and (if only registration is not bought, do not count temporarily), such as new user
(such as just registration, having first single purchase or without purchase);The search listing of E-search rope model, i.e. user temporally dimension;FRFM moulds
Type, such as Recency, a such as recent the last time in middle of the month consumption time, such as Frequency, such as a recent middle of the month
Consuming frequency, such as Monetary, the spending amount of such as recent one month, etc..
Further, as a kind of optional embodiment, in the case where the target data includes the prediction data, this method
It further includes:The prediction data is parsed, obtains one or more of following parsing data:Probabilistic forecasting data, wind
Dangerous prediction data, (as shown in Fig. 3 K, prediction data can include probabilistic forecasting data, risk profile to tactful prediction data simultaneously
Data, tactful prediction data), wherein, the first default dimension corresponding with the prediction data includes:Probabilistic forecasting dimension, risk
Predict dimension and strategy prediction dimension;In the case where the parsing data of the prediction data include the probabilistic forecasting data, it is based on
The probabilistic forecasting data portray the probabilistic forecasting portrait of the target user in the probabilistic forecasting dimension;In the solution of the prediction data
Data are analysed including in the case of the risk profile data, the mesh is portrayed in the risk profile dimension based on the risk profile data
Mark the risk profile portrait of user;And in the case where the parsing data of the prediction data include the strategy prediction data, base
The strategy prediction portrait of the target user is portrayed in strategy prediction dimension in the strategy prediction data.
It should be noted that different types of data in above-mentioned prediction data, for generating the different piece of user's portrait,
In the different piece for using different types of original recorded data generation user's portrait, the present invention does not limit its genesis sequence
It is fixed.
Specifically, it obtains and is handled in prediction model flow herein, including processing probabilistic model, Policy model and risk mould
One or more in type, specifically:
Probabilistic model is handled, it is sensitive including processing user active degree in the product, loyalty, shopping type, promotion
One or more feature in degree, purchase attribute bias.
Risk model is handled, including processing user credit risk in the product, potential loss risk, risk is satisfied with, abandons single wind
One or more feature in danger, risk of fraud, ox/reptile identity authentication.
Policy model is handled, includes the one or more features of following items in processing user.
Shopping stage forecast, arbitrarily strolls, focusing searching, gets a good buy by shopping around, waits for price reduction, comes with preparation.
Missing information strategy, including such as gender prediction, work prediction, not consumer purpose consuming capacity prediction, etc..
Migration efficiency is predicted, including user index such as to be developed, user index to be retrieved, user index to be activated, is treated
Safeguard user index, etc..
Customer value is predicted, including the value that prediction user brings to product, such as monetary value, prestige value, influence
Power value, etc..
Domestic demand prediction, current demand and potential demand including predicting user, etc..
Detest degree predicts, including prediction user to brand, classification, commodity detest degree, etc..
Through the embodiment of the present invention, as a result of the side that user's portrait is portrayed initial data in multiple default dimensions
Formula, thus can generate and draw a portrait than more comprehensive user, in this way, in order to supervise the Team Member of relevant enterprise in product design
Personal like can be cast aside in the process, focus is subjected to product design in the motivation of target user and behavior, therefore, used
Family portrait can be such that the service object of product more focuses on, and more be absorbed in.And due to doing product design for specific personage
Thing that will be far superior to fabricate in brain design, thus when all people for participating in product designs be all based on consistent user into
Row discusses and decision, and being just easy to constrain each side can be maintained in same general orientation, improve the efficiency of decision.
Also, the present invention is using a variety of composition datas, that is, a variety of data samples of initial data as the sample of machine learning,
The convenient abundant and training that sample data is carried out to certain machine learning example, so as to the business effect of automation services on increase line
Fruit, and then user's portrait generation scheme of regulation and standardization is formed, convenient for Rapid Implementation.
It should be noted that the user's portrait arrived involved in the whole flow process of the embodiment of the present invention and each flow
It can include a kind of general parameter, such as time parameter in acquisition methods and Partial Feature.
Exemplary means
After the method for exemplary embodiment of the invention is described, next, with reference to figure 4 to the exemplary reality of the present invention
Apply mode, be used to implement user draw a portrait depicting method device be described in detail.
An embodiment of the present invention provides a kind of user portrait characterization devices.
The block diagram for characterization device of drawing a portrait Fig. 4 schematically shows user according to embodiments of the present invention.As shown in figure 4,
User portrait characterization device 400 can include:Acquisition module 410 for being directed to target user, obtains to portray the target
The initial data of user's portrait of user;Determining module 420 for being based on the initial data, determines the target user multiple
The user characteristics showed in first default dimension;And first portray module 430, it is more at this for passing through the target user
The user characteristics showed in a first default dimension portray user's portrait of the target user.
Through the embodiment of the present invention, due to arbitrary using being portrayed on the latitude of multiple different characteristics that can reflect user
The mode of the portrait of one user, it is thus possible to portray the purpose of user's portrait with realizing multi-angle stage construction, it is possible to extremely
Partially overcome in the relevant technologies due to only portraying user's portrait from communication level or family's Attribute Association relationship level
And the defects of causing the application field of user's portrait limited, so as to significantly extend the adaptation field of user's portrait.
As a kind of optional embodiment, described device further includes:Labeling module, for portraying the target user
User portrait during, the target user is showed at least one of the multiple first default dimension dimension
User characteristics out carry out labeling, obtain corresponding user characteristics label;And second portray module, it is described for passing through
User characteristics label and remaining user characteristics portray user's portrait of the target user, wherein, remaining described user characteristics packet
It includes:The target user shows in other dimensions in the multiple first default dimension in addition at least one dimension
The user characteristics come.
As a kind of optional embodiment, the labeling module includes:First determination unit, for be directed to it is described at least
Each dimension in one dimension determines the classification for the user characteristics that the target user shows in the dimension;Analysis
Unit, the user characteristics for being showed in the dimension to the target user carry out semantic analysis, obtain and the class
The corresponding semanteme of other user characteristics;And first generation unit, for by the classification and with the user of the classification it is special
It levies corresponding semanteme to be associated, to obtain the user characteristics label in the dimension.
As a kind of optional embodiment, the labeling module includes:Statistic unit, it is described at least one for being directed to
Each dimension in dimension, the user characteristics showed in the dimension to the target user are counted, are corresponded to
Statistical result;Acquiring unit, for obtaining externally input additional information;And second generation unit, it is described for being based on
Statistical result and the additional information, the user characteristics showed in the dimension to the target user are tagged, with
Obtain the user characteristics label in the dimension.
As a kind of optional embodiment, the determining module includes:Extraction unit, for being carried from the initial data
Take out and portray the relevant plurality of target data of user's portrait;And second determination unit, for being based on the plurality of target number
According to, determine the user characteristics that the target user shows in the multiple first default dimension, wherein, a kind of number of targets
According to a corresponding first default dimension.
As a kind of optional embodiment, second determination unit includes:Subelement is translated, for by a variety of mesh
Each target data in mark data is translated according to preset rules, obtains corresponding various structures data;Analysis is single
Member for carrying out data analysis to each structural data in the various structures data, obtains a variety of data objects,
In, a kind of data object corresponds to a first default dimension;And determination subelement, for being based on a variety of data objects,
Determine the user characteristics that the target user shows in the multiple first default dimension.
As a kind of optional embodiment, the plurality of target data include at least two in data below:Basic number
According to, original recorded data, factual data, statistical data and prediction data.
As a kind of optional embodiment, the second determination unit is additionally operable to include the basic number in the target data
In the case of, the basic data is parsed, obtains one or more of following parsing data:Attribute data,
Character data, associated data, terminal data and log-on data, wherein, the first default dimension packet corresponding with the basic data
It includes:Attribute dimensions, role's dimension, relevant dimension, terminal dimension and registration dimension;And described first portray module and be additionally operable to:
In the case where the parsing data of the basic data include the attribute data, based on the attribute data in the attribute dimension
The attribute portrait of the target user is portrayed on degree;Include the situation of the character data in the parsing data of the basic data
Under, the role for being portrayed the target user in role's dimension based on the character data is drawn a portrait;In the basic data
Parsing data include the associated data in the case of, the mesh is portrayed on the relevant dimension based on the associated data
Mark the association portrait of user;In the case where the parsing data of the basic data include the terminal data, based on the end
End data portrays the terminal portrait of the target user in the terminal dimension;And the parsing data in the basic data
In the case of including the log-on data, the note of the target user is portrayed in the registration dimension based on the log-on data
Volume portrait.
As a kind of optional embodiment, the second determination unit is additionally operable to include the original note in the target data
In the case of recording data, the original recorded data is parsed, obtains one or more of following parsing data:It visits
Ask data, operation data, the first consumption data, the second consumption data and feedback data, wherein, with the original recorded data pair
The the first default dimension answered includes:Access dimension, operation dimension, the first consumption dimension, the second consumption dimension and feedback dimension;With
And described first portray module and be additionally operable to:Include the situation for accessing data in the parsing data of the original recorded data
Under, the access for being portrayed the target user in the access dimension based on the access data is drawn a portrait;In the original record
In the case that the parsing data of data include the operation data, institute is portrayed in the operation dimension based on the operation data
State the operation portrait of target user;Include the situation of first consumption data in the parsing data of the original recorded data
Under, the first consumption for being portrayed the target user in the described first consumption dimension based on first consumption data is drawn a portrait;
In the case that the parsing data of the original recorded data include second consumption data, existed based on second consumption data
The second consumption portrait of the target user is portrayed in the second consumption dimension;And the parsing in the original recorded data
In the case that data include the feedback data, the target user is portrayed in the feedback dimension based on the feedback data
Feedback portrait.
As a kind of optional embodiment, the second determination unit is additionally operable to include the true number in the target data
In the case of, the factual data is parsed, obtains one or more of following parsing data:Disappear for M times earliest
Take data, last n times consumption data, preceding L consumption data, consumption data in preset time period, wherein, with the factual data
Corresponding first default dimension includes:M consumption dimension, last n times consumption dimension, preceding L consumption dimension, preset time earliest
Consumption dimension in section;And described first portray module and be additionally operable to:Include the earliest M in the parsing data of the factual data
In the case of secondary consumption data, the target is portrayed in the earliest M consumption dimension based on the earliest M consumption data
The earliest M consumption portrait of user;Include the situation of the last n times consumption data in the parsing data of the factual data
Under, the last n times for being portrayed the target user in the last n times consumption dimension based on the last n times consumption data are disappeared
Take portrait;In the case where the parsing data of the factual data include the preceding L consumption data, disappeared based on described first L times
Expense data portray the preceding L consumption portrait of the target user in the preceding L consumption dimension;And in the factual data
Parsing data include in the preset time period in the case of consumption data, based on consumption data in the preset time period existing
Consumption portrait in the preset time period of the target user is portrayed in the preset time period in consumption dimension.
As a kind of optional embodiment, the second determination unit is additionally operable to include the statistical number in the target data
In the case of, the statistical data is parsed, obtains one or more of following parsing data:Preference data,
Geodata, regular data, wherein, the first default dimension corresponding with the statistical data includes:Preference dimension, geography dimensionality
With regular dimension;And described first portray module and be additionally operable to:Include the preference number in the parsing data of the statistical data
In the case of, the preference portrait of the target user is portrayed in the preference dimension based on the preference data;Described
In the case that the parsing data of statistical data include the geodata, carved on the geography dimensionality based on the geodata
Draw the geographical portrait of the target user;And include the situation of the regular data in the parsing data of the statistical data
Under, the rule for being portrayed the target user in the regular dimension based on the regular data is drawn a portrait.
As a kind of optional embodiment, the second determination unit is additionally operable to include the prediction number in the target data
In the case of, the prediction data is parsed, obtains one or more of following parsing data:Probabilistic forecasting number
According to, risk profile data, tactful prediction data, wherein, the first default dimension corresponding with the prediction data includes:Probability is pre-
Survey dimension, risk profile dimension and strategy prediction dimension;And described first portray module and be additionally operable to:In the prediction data
In the case that parsing data include the probabilistic forecasting data, based on the probabilistic forecasting data in the probabilistic forecasting dimension
Portray the probabilistic forecasting portrait of the target user;Include the risk profile data in the parsing data of the prediction data
In the case of, the risk profile for being portrayed the target user in the risk profile dimension based on the risk profile data is drawn
Picture;And in the case where the parsing data of the prediction data include the tactful prediction data, based on the strategy prediction
Data portray the strategy prediction portrait of the target user in the strategy prediction dimension.
It should be noted that present apparatus section Example and the embodiment of aforementioned method part are same or similar,
The technical issues of solution, the particular technique means used and the technique effect reached are also all same or similar, elaborate
It please join the description of aforementioned method part, details are not described herein.
Exemplary media
An embodiment of the present invention provides a kind of computer readable storage mediums, are stored thereon with executable instruction, the instruction
User's portrait depicting method of any one of this method embodiment is used to implement when being performed by the processing module.
In some possible embodiments, various aspects of the invention are also implemented as a kind of shape of program product
Formula, including program code, when described program product is run on the terminal device, said program code is used to make the terminal
Equipment performs the use according to the various illustrative embodiments of the present invention described in above-mentioned " illustrative methods " part of this specification
Step in the portrait depicting method of family, for example, the terminal device can perform operation S210 as shown in Figure 2, for mesh
User is marked, obtains to portray the initial data that the user of target user draws a portrait;S220 is operated, based on initial data, determines mesh
The user characteristics that mark user shows in the multiple first default dimensions;And operation S230, by target user multiple
The user characteristics showed in first default dimension portray user's portrait of target user.
The arbitrary combination of one or more readable mediums may be used in described program product.Readable medium can be readable letter
Number medium or readable storage medium storing program for executing.Readable storage medium storing program for executing for example may be-but not limited to-electricity, magnetic, optical, electromagnetic, red
The system of outside line or semiconductor, device or device or arbitrary above combination.The more specific example of readable storage medium storing program for executing
(non exhaustive list) includes:Electrical connection, portable disc with one or more conducting wires, hard disk, random access memory
(RAM), read-only memory (ROM), erasable programmable read only memory (EPROM or flash memory), optical fiber, portable compact disc
Read memory (CD-ROM), light storage device, magnetic memory device or above-mentioned any appropriate combination.
As shown in figure 5, the program product 50 portrayed for user's portrait according to the embodiment of the present invention is described,
Portable compact disc read only memory (CD-ROM) may be used and including program code, and can be such as a in terminal device
It is run on people's computer.However, the program product of the present invention is without being limited thereto, in this document, readable storage medium storing program for executing can be any
Comprising or storage program tangible medium, the program can be commanded execution system, device either device using or tied with it
It closes and uses.
Readable signal medium can be included in a base band or as the data-signal that a carrier wave part is propagated, wherein carrying
Readable program code.Diversified forms may be used in the data-signal of this propagation, believe including --- but being not limited to --- electromagnetism
Number, optical signal or above-mentioned any appropriate combination.Readable signal medium can also be other than readable storage medium storing program for executing it is any can
Read medium, which can send, propagate either transmit for being used by instruction execution system, device or device or
Program in connection.
The program code included on readable medium can be transmitted with any appropriate medium, including --- but being not limited to ---
Wirelessly, wired, optical cable, RF etc. or above-mentioned any appropriate combination.
It can combine to write to perform the program that the present invention operates with the arbitrary of one or more programming languages
Code, described program design language is including object oriented program language --- such as Java, C++ etc. further include routine
Procedural programming language --- such as " C ", language or similar programming language.Program code can fully exist
It performs in user calculating equipment, partly perform on a user device, the software package independent as one performs, partly in user
Part performs or is performed in remote computing device or server completely on a remote computing on computing device.It is relating to
And in the situation of remote computing device, remote computing device can pass through the network of any kind --- including LAN (LAN)
Or wide area network (WAN) one be connected to user calculating equipment or, it may be connected to external computing device (such as utilizes internet
Service provider passes through Internet connection).
Exemplary computer device
After method, medium and the system for describing exemplary embodiment of the invention, next, introducing according to this hair
The computing device portrayed for user's portrait of a bright illustrative embodiments.
The embodiment of the present invention additionally provides a kind of computing device.The computing device includes:Processing module;And storage is single
Member is stored with computer executable instructions, which, which is used to implement when being performed by the processing module in this method embodiment, appoints
User's portrait depicting method described in one.
Person of ordinary skill in the field it is understood that various aspects of the invention can be implemented as device, method or
Program product.Therefore, various aspects of the invention can be with specific implementation is as follows, i.e.,:It is complete hardware embodiment, complete
The embodiment combined in terms of full Software Implementation (including firmware, microcode etc.) or hardware and software, can unite here
Referred to as " circuit ", " module " or " device ".
In some possible embodiments, computing device according to the present invention can include at least at least one processing mould
Block and at least one storage unit.Wherein, the storage unit has program stored therein code, when said program code is described
Processing module perform when so that the processing module perform described in above-mentioned " illustrative methods " part of this specification according to this
Step in the user's portrait depicting method for inventing various illustrative embodiments.For example, the processing module can perform such as
Operation S210 shown in Fig. 2 for target user, obtains to portray the initial data that the user of target user draws a portrait;Behaviour
Make S220, based on initial data, determine the user characteristics that target user shows in the multiple first default dimensions;And behaviour
Make S230, the user characteristics showed in the multiple first default dimensions by target user portray the user of target user
Portrait.
The computing device portrayed for user's portrait of this embodiment according to the present invention is described referring to Fig. 6
60.Computing device 60 as shown in Figure 6 is only an example, and the function and use scope of the embodiment of the present invention should not be brought
Any restrictions.
As shown in fig. 6, computing device 60 is showed in the form of universal computing device.The component of computing device 60 can include
But it is not limited to:Above-mentioned at least one processing module 601, above-mentioned at least one storage unit 602, connection different system component (packet
Include storage unit 602 and processing module 601) bus 603.
Bus 603 represents one or more in a few class bus structures, including memory bus or Memory Controller,
Peripheral bus, graphics acceleration port, processor or the local bus using the arbitrary bus structures in a variety of bus structures.
Storage unit 602 can include the readable medium of form of volatile memory, such as random access memory (RAM)
6021 and/or cache memory 6022, it can further include read-only memory (ROM) 6023.
Storage unit 602 can also include program/utility with one group of (at least one) program module 6024
6025, such program module 6024 includes but not limited to:Operating system, one or more application program, other program moulds
Block and program data may include the realization of network environment in each or certain combination in these examples.
Computing device 60 can also be with one or more external equipments 604 (such as keyboard, sensing equipment, bluetooth equipment etc.)
Communication can also enable a user to the equipment interacted with computing device 60 communication with one or more and/or be set with to calculate
The standby 60 any equipment (such as router, modem etc.) that can be communicated with one or more of the other computing device are logical
Letter.This communication can be carried out by input/output (I/O) interface 605.Also, computing device 60 can also be fitted by network
Orchestration 606 and one or more network (such as LAN (LAN), wide area network (WAN) and/or public network, such as because of spy
Net) communication.As shown in the figure, network adapter 606 is communicated by bus 603 with other modules of computing device 60.It should be understood that
Although not shown in the drawings, can combine computing device 60 uses other hardware and/or software module, including but not limited to:Micro- generation
Code, device driver, redundancy processing module, external disk drive array, RAID system, tape drive and data backup are deposited
Storage system etc..
It should be noted that although be referred in above-detailed user draw a portrait characterization device several units/modules or
Subelement/module, but it is this division be only exemplary it is not enforceable.In fact, embodiment party according to the present invention
Formula, the feature and function of two or more above-described units/modules can embody in a units/modules.Conversely,
The feature and function of an above-described units/modules can be further divided into being embodied by multiple units/modules.
In addition, although the operation of the method for the present invention is described with particular order in the accompanying drawings, this do not require that or
The operation that these operations must be performed or have to carry out shown in whole according to the particular order by implying could be realized desired
As a result.Additionally or alternatively, it is convenient to omit multiple steps are merged into a step and performed and/or by one by certain steps
Step is decomposed into execution of multiple steps.
Although describe spirit and principles of the present invention by reference to several specific embodiments, it should be appreciated that, this
Invention is not limited to disclosed specific embodiment, does not also mean that the feature in these aspects cannot to the division of various aspects
Combination is this to divide merely to the convenience of statement to be benefited.The present invention is directed to cover appended claims spirit and
In the range of included various modifications and equivalent arrangements.