CN104331429B - The method and device of multiple features dimension quantization is carried out to network object - Google Patents
The method and device of multiple features dimension quantization is carried out to network object Download PDFInfo
- Publication number
- CN104331429B CN104331429B CN201410564598.7A CN201410564598A CN104331429B CN 104331429 B CN104331429 B CN 104331429B CN 201410564598 A CN201410564598 A CN 201410564598A CN 104331429 B CN104331429 B CN 104331429B
- Authority
- CN
- China
- Prior art keywords
- data
- characteristic dimension
- network object
- feature
- weight
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000013139 quantization Methods 0.000 title claims abstract description 29
- 230000001419 dependent effect Effects 0.000 claims description 5
- 238000011002 quantification Methods 0.000 claims description 3
- 230000003247 decreasing effect Effects 0.000 description 5
- 230000010354 integration Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 238000012217 deletion Methods 0.000 description 3
- 230000037430 deletion Effects 0.000 description 3
- 241000408529 Libra Species 0.000 description 2
- 241000406668 Loxodonta cyclotis Species 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000005855 radiation Effects 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 241000475699 Euphaedusa digonoptyx comes Species 0.000 description 1
- 241000171154 Nematus <genus> Species 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000001035 drying Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Information Transfer Between Computers (AREA)
Abstract
The invention discloses a kind of method and device that multiple features dimension quantization is carried out to network object, this method includes：Determine a network object, obtain and describe data with the corresponding feature of the network object；Different characteristic according to pre-establishing describes the correspondence between data and different characteristic dimension, the feature corresponding with the network object is described data to be matched with characteristic dimension, and each characteristic dimension weight corresponding with the network object is obtained according to matching result；According to each characteristic dimension weight, the quantized value of each characteristic dimension of the network object is obtained.Method using the present invention, is handled by describing data to the feature of network object, forms the quantized value of each characteristic dimension of the network object, intuitively the feature of each dimension of reaction network object.
Description
Technical field
This application involves Internet technical field, more particularly to a kind of side that multiple features dimension quantization is carried out to network object
Method and device.
Background technology
With the rapid development of Internet technology, there is more and more network objects, such as video text on network
Part, audio file, text etc., go to browse or download for users.On the other hand, skill is searched for by web crawlers etc.
Art, for different network objects, can search on network and get a large amount of relevant information, such as these networks pair
Comment, recommendation or label data of elephant etc., these information are much all based on UGC (User Generated Content, user
Generation content) mode generate.
However, for the great deal of related information of network object, particularly UGC data messages, since different web sites are to difference
The subjective division of network object, and structural input limitation of the user on different web sites, causing these data can only
Isolated is present in different web sites, and is scattered in the different web pages of different web sites, can not produce between each other very big
Reference significance, although some websites may also can be integrated and be quantified to some of which information data, in view of different nets
The information standard and disunity stood to consolidated network object, therefore be generally also limited only to carry out the relevant information in our station
Integrate, and the quantized data after this integration is very single in structure and is difficult to extend, while also result in and measuring
Change and be very limited system on the exhibition method of data.
The content of the invention
In view of the above problems, it is proposed that the present invention overcomes the above problem in order to provide one kind or solves at least in part
That states problem carries out the method for multiple features dimension quantization and corresponding device to network object.
According to an aspect of the invention, there is provided a kind of method that multiple features dimension quantization is carried out to network object, bag
Include：Determine a network object, obtain and describe data with the corresponding feature of the network object；According to the different characteristic pre-established
The correspondence between data and different characteristic dimension is described, the feature corresponding with the network object is described into data and spy
Sign dimension is matched, and obtains each characteristic dimension weight corresponding with the network object according to matching result；According to described each
Characteristic dimension weight, obtains the quantized value of each characteristic dimension of the network object.
Wherein, the step of describing data with the corresponding feature of the network object is obtained, including：According to the network object
Type, predefine one or more web search addresses；Obtained by the web search address opposite with the network object
The feature answered describes data.
Wherein, the step of describing data with the corresponding feature of the network object is obtained, including：Heterogeneous networks will be come from
The feature of address describes data and carries out registration matching；Description data of the registration less than threshold value are deleted, generate final spy
Descriptor data set is levied to close.
Wherein, the step of describing data with the corresponding feature of the network object is obtained, including：Heterogeneous networks will be come from
The feature of address describes data and carries out registration matching；Data distribution is described accordingly for the feature according to the height of registration
Weight, the more high then weight of registration are higher；The feature repeated is described data to delete, generates the feature descriptor data set of Weight
Close.
Wherein, the step of describing data with the corresponding feature of the network object is obtained, is further comprised：According to true in advance
The fixed weight set for heterogeneous networks address, further describes the feature for coming from heterogeneous networks address the weight of data
It is adjusted.
Wherein, the characteristic dimension is determined according to the type of the network object.
Wherein, correspondence between data and different characteristic dimension is described according to the different characteristic pre-established, by institute
State feature corresponding with the network object and describe data and matched with characteristic dimension, and obtained and the network according to matching result
The step of object corresponding each characteristic dimension weight, including：According to the correspondence, by whole features describe data respectively with
Characteristic dimension is matched, and repeated matching is then carried out if the feature repeated describes data；Determined according to matching result to each
The ratio of the hit-count of characteristic dimension；According to the ratio of the hit-count, each feature corresponding with the network object is obtained
Dimension weight.
Wherein, correspondence between data and different characteristic dimension is described according to the different characteristic pre-established, by institute
State feature corresponding with the network object and describe data and matched with characteristic dimension, and obtained and the network according to matching result
The step of object corresponding each characteristic dimension weight, including：According to the correspondence, data difference is described into different features
Matched with characteristic dimension；The ratio of the hit-count to each characteristic dimension is determined according to matching result；According to different characteristic
The weight of data is described, the hit-count ratio of characteristic dimension will be adjusted；According to the ratio of the hit-count after adjustment,
Obtain each characteristic dimension weight corresponding with the network object.
Wherein, according to each characteristic dimension weight, the step of obtaining the quantized value of each characteristic dimension of the network object,
Including：Obtain score value corresponding with the network object and describe data；To be described with the corresponding score value of the network object data with
Each characteristic dimension weight is multiplied respectively, obtains the quantized value of each characteristic dimension of the network object.
Wherein, the step of score value corresponding with the network object describes data is obtained, including：Acquisition comes from heterogeneous networks
The score value corresponding with the network object of address describes data, and according to the predetermined power set for heterogeneous networks address
Weight, describes data by the score value for coming from heterogeneous networks address and is weighted summation, data are described as final score value.
Wherein, the description data include label data, comment data, and/or key data.
Wherein, further comprise：From described with choosing at least partly spy in the corresponding each characteristic dimension of the network object
Levy dimension；According to obtaining quantized value corresponding with the characteristic dimension of the selection, in a manner of multi-dimensional map by the quantized value into
Row displaying.
Wherein, further comprise：According to the quantized value of each characteristic dimension of the obtained network object, therefrom choose at least
Partial Feature dimension is shown.
According to another aspect of the present invention, there is provided a kind of device that multiple features dimension quantization is carried out to network object, is obtained
Module is obtained, for determining a network object, obtains and describes data with the corresponding feature of the network object；Matching module, is used for
Different characteristic according to pre-establishing describes the correspondence between data and different characteristic dimension, will the described and network object
Corresponding feature describes data and is matched with characteristic dimension, and obtains each spy corresponding with the network object according to matching result
Levy dimension weight；Quantization modules, for according to each characteristic dimension weight, obtaining the amount of each characteristic dimension of the network object
Change value.
Wherein, the acquisition module includes：Determining module, for the type according to the network object, predefines one
A or multiple web search addresses；Acquisition module, it is corresponding with the network object for being obtained by the web search address
Feature describes data.
Wherein, the acquisition module includes：First registration matching module, for the spy of heterogeneous networks address will to be come from
Sign description data carry out registration matching；First generation module, for description data of the registration less than threshold value to be deleted, generation
Final feature descriptor data set closes.
Wherein, the acquisition module includes：Second registration matching module, for the spy of heterogeneous networks address will to be come from
Sign description data carry out registration matching；Weight distribution module, for describing data according to the height of registration for the feature
Corresponding weight is distributed, the more high then weight of registration is higher；Second generation module, is deleted for the feature repeated to be described data
Remove, the feature descriptor data set for generating Weight closes.
Wherein, the acquisition module further comprises：Weight adjusts module, for being directed to different nets according to predetermined
The weight that network address is set, the weight that data are further described to the feature for coming from heterogeneous networks address are adjusted.
Wherein, the characteristic dimension is determined according to the type of the network object.
Wherein, the matching module includes：First matching module, for according to the correspondence, whole features to be retouched
State data to be matched with characteristic dimension respectively, repeated matching is then carried out if the feature repeated describes data；First ratio
Determining module, for determining the ratio of the hit-count to each characteristic dimension according to matching result；First weight obtains module, uses
In the ratio according to the hit-count, each characteristic dimension weight corresponding with the network object is obtained.
Wherein, the matching module includes：Second matching module, for according to the correspondence, by different features
Description data are matched with characteristic dimension respectively；Second ratio-dependent module, for being determined according to matching result to each feature
The ratio of the hit-count of dimension；Ratio adjusts module, will be to characteristic dimension for describing the weight of data according to different characteristic
Hit-count ratio be adjusted；Second weight obtains module, for the ratio according to the hit-count after adjustment, obtain with
The corresponding each characteristic dimension weight of the network object.
Wherein, the quantization modules, including：Score value acquisition module, is retouched for obtaining score value corresponding with the network object
State data；Quantification treatment module, is weighed for that will describe data with the corresponding score value of the network object with each characteristic dimension
It is multiplied respectively again, obtains the quantized value of each characteristic dimension of the network object.
Wherein, the score value acquisition module, is further used for：Obtain come from heterogeneous networks address with the network object
Corresponding score value describes data, and according to the predetermined weight set for heterogeneous networks address, will come from different nets
The score value of network address describes data and is weighted summation, and data are described as final score value.
Wherein, the description data include label data, comment data, and/or key data.
Wherein, further comprise：Module is chosen, for being selected from corresponding each characteristic dimension with the network object
Take at least part characteristic dimension；First display module, the quantized value corresponding with the characteristic dimension of the selection obtained for basis,
The quantized value is shown in a manner of multi-dimensional map.
Wherein, the second display module, for the quantized value of each characteristic dimension according to the obtained network object, Cong Zhongxuan
At least part characteristic dimension is taken to be shown.
Technique according to the invention scheme, there are following technique effect：
The present invention describes the correspondence between data and different characteristic dimension according to the different characteristic pre-established, by net
The corresponding feature of network object describes data and is matched with characteristic dimension, obtains each characteristic dimension power corresponding with the network object
Weight, and then the quantized value of each characteristic dimension of network object is obtained according to the weight of each characteristic dimension, it can intuitively react net
The feature of each characteristic dimension of network object, simultaneously as can according to different network object types and to characteristic dimension into
Row increase or decrease, and and then to description data and characteristic dimension between correspondence be adjusted at any time, therefore, the present invention
It is as rich as Croesus that the technical solution of offer, which is directed in the structure of quantized data of the relevant information of heterogeneous networks object after integration,
Elasticity simultaneously easily extends, this also provides possibility for the more diversified exhibition method of offer.
Described above is only the general introduction of technical solution of the present invention, in order to better understand the technological means of the present invention,
And can be practiced according to the content of specification, and in order to allow above and other objects of the present invention, feature and advantage can
Become apparent, below especially exemplified by the embodiment of the present invention.
Brief description of the drawings
Attached drawing described herein is used for providing further understanding of the present application, forms the part of the application, this Shen
Schematic description and description please is used to explain the application, does not form the improper restriction to the application.In the accompanying drawings：
Fig. 1 is the main-process stream of the method according to an embodiment of the invention that multiple features dimension quantization is carried out to network object
Figure；
Fig. 2 is the step of acquisition according to an embodiment of the invention describes data with the corresponding feature of the network object
Flow chart；
Fig. 3 is that the different characteristic that basis according to an embodiment of the invention pre-establishes describes data and different characteristic dimension
Correspondence between degree, describes data by the feature corresponding with the network object and is matched with characteristic dimension, and root
The flow chart for the step of obtaining each characteristic dimension weight corresponding with the network object according to matching result；
Fig. 4 is according to an embodiment of the invention according to each characteristic dimension weight, obtains each of the network object
The flow chart of the step of quantized value of characteristic dimension；
Fig. 5 is the step of acquisition according to another embodiment of the present invention describes data with the corresponding feature of the network object
Flow chart；
Fig. 6 is that the different characteristic that basis according to another embodiment of the present invention pre-establishes describes data and different characteristic dimension
Correspondence between degree, describes data by the feature corresponding with the network object and is matched with characteristic dimension, and root
The flow chart for the step of obtaining each characteristic dimension weight corresponding with the network object according to matching result；
Fig. 7 is the structural frames of the device according to an embodiment of the invention that multiple features dimension quantization is carried out to network object
Figure；And
Fig. 8 is the quantized value of each characteristic dimension according to an embodiment of the invention to a film with multi-dimensional map
Displaying examples of interfaces figure when mode is shown.
Embodiment
The exemplary embodiment of the disclosure is more fully described below with reference to accompanying drawings.Although the disclosure is shown in attached drawing
Exemplary embodiment, it being understood, however, that may be realized in various forms the disclosure without should be by embodiments set forth here
Limited.On the contrary, these embodiments are provided to facilitate a more thoroughly understanding of the present invention, and can be by the scope of the present disclosure
Completely it is communicated to those skilled in the art.
As shown in Figure 1, Fig. 1 is the side according to an embodiment of the invention that multiple features dimension quantization is carried out to network object
The general flow chart of method.
At step S110, a network object is determined, obtain and describe data with the corresponding feature of the network object.
Specifically, network object can be determined according to the search of user, e.g., determines to use according to searching request input by user
The network object of family search；Or network object can be determined to the access of preset link according to user, e.g., user is to preset chain
Click connect etc..
Wherein, network object can include：Resource on internet, can include video file, audio file, text text
Part etc., e.g., film, TV, music, article, books etc..Describing data with the corresponding feature of the network object can include：
The label data of the network object, comment data, and/or key data etc..Wherein, label data for example can be that user gives birth to
Into for label that the network object is recommended, introduces, commented on, for example, substantial amounts of electricity can be searched on network
Shadow label；Comment data can be the comment that user carries out the network object；Key data can be to network object
Recommendation, comment data carry out the pass for the network object that specifically processing (e.g., carrying out word segmentation processing to comment word) is got
Key word.It will be appreciated by those skilled in the art that the embodiment of the present invention not limited to this.
According to one embodiment of present invention, can be obtained using following sub-step corresponding with the network object
Feature describes data.
First, according to the type of the network object, one or more web search address is predefined.For example, interconnection
The special sub- column for having many special websites either number of site on the net issues or searches for certain types of net for user
The feature of network object describes data, e.g., website or the net of information is issued or searched for for offers such as film, TV, books and periodicals, magazines
Stand sub- column etc., therefore, the website of data can be described according to the type of network object, predetermined search characteristics, that is, determine
One or more web search addresses.
Then, obtained by the web search address and describe data with the corresponding feature of the network object.That is, by true
The feature that the fixed one or more web search address obtains the network object describes data.It is pre- according to the type of network object
First determine web search address, search can be made to have more directionality, avoid scanning for from incoherent website, improve search effect
Rate.Also, the feature that network object is obtained from different network address describes data, can be by from different web sites and network
The corresponding feature of object describes data and is compared comprehensive, unified integration, and the number between data is described according to different characteristic
The correspondence that magnitude relation and feature are described between data and characteristic dimension quantifies the feature of network object.
In a specific embodiment, above-mentioned sub-step can be realized by web crawlers technology, i.e. by program or
Script obtains feature corresponding with the network object and describes data automatically.
With reference to figure 2, Fig. 2 is that acquisition according to an embodiment of the invention is described with the corresponding feature of the network object
The flow chart of the step of data.
As shown in Fig. 2, obtaining the step of describing data with the corresponding feature of the network object, following step can be included
Suddenly：
Step S210, describes data by the feature for coming from heterogeneous networks address and carries out registration matching.
The feature for coming from heterogeneous networks address describes in data that there may be (identical) data of many repetitions, i.e.
Same feature, which describes data, to be occurred repeatedly, wherein, similar data (the same or like data of meaning) can also regard repetition as
(identical) data.Specifically, a certain feature describe Data duplication appearance number it is more, or this feature describes data
And described with its similar feature data appearance number (overlapping number) it is more, then it is believed that this feature describes the weights of data
It is right higher.For example, being directed to a certain portion's film, there is the substantial amounts of label (tag) from each different web sites, in these labels
It may include many identical labels or similar label, then the label of the film got can be subjected to registration
Match somebody with somebody, the number (coincidence number) that a certain label repeats is more, or time that the label and the label similar to its occur
Number is more, then its registration is higher.
Step S220, describes data less than the feature of threshold value by registration and deletes, generate final feature descriptor data set
Close.
Specifically, feature of the registration less than the threshold value can be described by data according to pre-set registration threshold value
Delete, generate final feature descriptor data set and close.For example, can be according to pre-set coincidence frequency threshold value.Delete and be somebody's turn to do
The corresponding feature of network object describes to overlap the data that number is less than the coincidence frequency threshold value in data, generates final feature and retouches
State data acquisition system.Data are described less than the feature of threshold value by deleting registration, subsequent process number to be treated can be reduced
Data bulk, makes the data processing in subsequent process easier.
It should be noted that during the final feature descriptor data set of generation closes, deletion registration can be included and be less than the threshold value
Feature describe whole features of the remaining network object after data and describe data, in other words, including repeating or
Similar feature describes data.
Fig. 1 is returned to, after obtaining and describing data with the corresponding feature of the network object, next, in step
At S120, the correspondence between data and different characteristic dimension is described according to the different characteristic pre-established, will be described with being somebody's turn to do
The corresponding feature of network object describes data and is matched with characteristic dimension, and is obtained and the network object pair according to matching result
Each characteristic dimension weight answered.
Specifically, the corresponding characteristic dimension of the network object first can be determined according to the type of the network object, then
The correspondence that different characteristic according to pre-establishing is described between data and different characteristic dimension is corresponding by the network object
Feature describes data and is matched with characteristic dimension.If for example, the network object is film, the relevant spy of film can be determined
Sign dimension is the corresponding characteristic dimension of the network object, for example, the corresponding characteristic dimension of film can include：Audiovisual, plot, think of
Think, the characteristic dimension such as atmosphere, or the characteristic dimension of other attributes can also be included, e.g., the dimension such as spectators' gender, age, constellation
Degree.Also, can also be according to the type of different network objects, the demand of user etc., feature corresponding to network object
Dimension is increased or decreased, great flexibility.
Wherein, the correspondence that different characteristic is described between data and different characteristic dimension can be advance by following steps
Establish：
First, multiple features are obtained and describe data, and determine multiple characteristic dimensions.Wherein, the multiple characteristic dimension can
To include each multiple characteristic dimension included in one or more features dimension classification respectively.In other words, it is described
Multiple characteristic dimensions include one or more features dimension classification, and each of which characteristic dimension classification includes multiple feature dimensions
Degree.
Then, it is determined that the multiple feature describes the correspondence of data and the multiple characteristic dimension, for example, can root
Determine that feature describes whether data and characteristic dimension have correspondence according to the result that a large amount of historical datas are carried out with statistical analysis.
Also, the correspondence that can also be described different characteristic between data and different characteristic dimension is configured to a mapping table.
By taking film label as an example, configure corresponding between each film label (feature describes data) and each characteristic dimension
Relation table can be as shown in table 1：
Table 1
Wherein, for each label (feature describes data), pair it is marked with its characteristic dimension with correspondence,
That is, marking the label and this feature dimension has correspondence, for example, by the use of " 1 " as the mark with correspondence in table 1,
For example, label " comedy " has corresponding relation with characteristic dimension " plot ", then " plot " dimension of " comedy " is marked.
It should be noted that the correspondence that Partial Feature describes data and Partial Feature dimension is merely illustrated in table 1.
It should be noted that different characteristic is pre-established in the present invention describes correspondence between data and different characteristic dimension
Mode not limited to this.Also, since the correspondence that the different characteristic describes between data and different characteristic dimension is advance
Configured, therefore, in actual application, dimension extension can be carried out according to different demands, i.e. increase new feature
Dimension, also, during progress dimension extension, it is not necessary to collect relevant feature for newly-increased dimension and describe data, only need to be according to
Some features describe data, and the correspondence that complementary features describe data and newly-increased feature dimension is adjusted, the extension of dimension
Quite flexible.
With reference to figure 3, Fig. 3 be according to one embodiment of present invention according to the different characteristic that pre-establishes describe data with not
With the correspondence between characteristic dimension, data and characteristic dimension progress are described into the feature corresponding with the network object
The flow of the step of matching somebody with somebody, and each characteristic dimension weight corresponding with the network object is obtained according to matching result (step S120)
Figure.
As shown in figure 3, step S120 may further include following steps：
Whole features according to the correspondence, are described data and are matched respectively with characteristic dimension, such as by step S310
The feature that fruit has repetition describes data and then carries out repeated matching.
Specifically, the different characteristic that can be pre-established according to this describes corresponding between data and different characteristic dimension
Acquisition whole features corresponding with the network object are described data and are matched respectively with characteristic dimension by relation, that is,
Each feature during the feature descriptor data set of the network object is closed describes data and is matched respectively with characteristic dimension.Wherein,
If the feature that whole features describes to there is repetition in data describes data (that is, the same or similar feature describes data),
Then carry out repeated matching, that is to say, that no matter describing data in the presence of how much the same or similar features, all respectively with characteristic dimension
Matched so that the number of repetition that different characteristic describes data has an impact final quantized value result.
More specifically, it may be determined that whole features of the network object describe to retouch in the feature pre-established in data
State feature present in the correspondence of data and different characteristic dimension time and describe data, and there will be feature data are described
Matched respectively with each characteristic dimension in the correspondence, if a feature describes data in the feature description pre-established
Data then determine to describe data tool with this feature with the correspondence of different characteristic dimension time existing according to the correspondence
There is the one or more features dimension of correspondence, and record the network object and the one or more features dimension is hit respectively
Once, if the feature repeated describes data, then the feature repeated is described into data and carries out repeated matching with each characteristic dimension.
Step S320, the ratio of the hit-count to each characteristic dimension is determined according to matching result.
Specifically, hit that can be according to each characteristic dimension in the corresponding whole features of the network object describe data
Number, determines the hit-count ratio of each characteristic dimension, that is to say, that the whole features for accumulating the network object describe data pair
The hit-count of each characteristic dimension, and determine the network object to each characteristic dimension according to the hit-count of each characteristic dimension
Hit-count ratio.
Step S330, according to the ratio of the hit-count, obtains each characteristic dimension weight corresponding with the network object.
That is, according to the hit-count ratio of each characteristic dimension, the corresponding each characteristic dimension of the network object is determined
Weight.Specifically, how corresponding the weight of hit-count characteristic dimension more be bigger, thus can according to hit-count ratio from
Small order is arrived greatly, and the weight of each characteristic dimension is determined in a manner of weight is successively decreased successively, or can be directly by each feature dimensions
Weight of the hit-count ratio of degree as each characteristic dimension.And wherein, if the plurality of characteristic dimension includes multiple features
Dimension classification, then can determine weight of each characteristic dimension in each affiliated classification respectively according to characteristic dimension classification,
Wherein, the sum of weight of each characteristic dimension is equal to 1 in any one characteristic dimension classification.
The corresponding weight of corresponding with the network object each characteristic dimension can be obtained by above-mentioned steps S310~S330,
After each characteristic dimension weight corresponding with the network object is obtained, Fig. 1 is next returned, carries out step S130.
At step S130, according to each characteristic dimension weight, the quantization of each characteristic dimension of the network object is obtained
Value.
In order to more clearly explain the step, a kind of optional specific embodiment of the step is described below with reference to Fig. 4.
Fig. 4 is foundation each characteristic dimension weight according to one embodiment of the invention, obtains each characteristic dimension of the network object
Quantized value the step of flow chart.
Step S410, obtains score value corresponding with the network object and describes data.
Specifically, it can obtain and come from the score value corresponding with the network object of heterogeneous networks address and describe data, and
According to it is predetermined for heterogeneous networks address set weight, by the score value for coming from heterogeneous networks address describe data into
Row weighted sum, data are described as final score value.That is, data are described into the score value from heterogeneous networks address
The weighted average score value final as the network object describes data.For example, for a certain portion's film, it can obtain and come from
User's scoring score value of the film of heterogeneous networks address, and will be come from according to the weight set in advance for heterogeneous networks address
User's scoring score value of heterogeneous networks address is weighted summation, using point that obtained weighted average is final as the film
Value.
The above-mentioned weight according to network address describes data to the score value from heterogeneous networks address and is weighted, and obtains
Final score value data are described will be more objective and accurate.Wherein, the weight set for heterogeneous networks address, can be with
It is configured in advance, for example, the weight of the website is set according to user to the fancy grade or degree of concern of website, that is,
Say, the weight of the network address of the website is set, for the network address that user preferences degree is higher or degree of concern is higher, its
Corresponding weight can set higher.
Step S420, data will be described with the corresponding score value of the network object and will be multiplied respectively with each characteristic dimension weight,
Obtain the quantized value of each characteristic dimension of the network object.
That is, based on the score value of the network object describes data, the corresponding score value of the network object is retouched
State each characteristic dimension multiplied by weight of the data respectively with the network object, quantization of the obtained result respectively as each characteristic dimension
Value, i.e. the corresponding score value of each characteristic dimension of the network object.
Detailed above in association with Fig. 1~Fig. 4 according to embodiments of the present invention to network object progress multiple features dimension quantization
Method idiographic flow.
Multiple features dimension is carried out to network object to according to another embodiment of the present invention with reference to Fig. 1, Fig. 5 and Fig. 6
The method of quantization is described in detail.
As shown in Figure 1, at step S110, a network object is determined, acquisition is retouched with the corresponding feature of the network object
State data.
Specifically, network object can be determined according to the search of user, e.g., according in search query input by user
Keyword determine network object；Or network object can be determined to the access of preset link according to user, e.g., user is to pre-
Click for the access buttons put etc..After determining network object, obtain and describe data, example with the corresponding feature of the network object
Such as, it can automatically be obtained using crawlers or script and retouched with the corresponding feature of the network object by web crawlers technology
State data.
Fig. 5 is the step of describing data with the corresponding feature of the network object according to the acquisition of another embodiment of the present invention
Flow chart.As shown in figure 5, in the present embodiment, obtaining the step of describing data with the corresponding feature of the network object can
To further comprise the steps：
Step S510, describes data by the feature for coming from heterogeneous networks address and carries out registration matching.
Specifically, the feature for coming from heterogeneous networks address that can be according to corresponding to the network object is described in data
The number that each Data duplication occurs, determines that each feature describes the registration height of data, i.e. comes from the network object is corresponding
Being described in the feature of heterogeneous networks in data, identical feature describes the number of Data duplication appearance, wherein, similar data
(the same or like data of meaning) can also regard (identical) data of repetition as.A certain feature describes Data duplication appearance
Number is more, or this feature describes data and the feature similar with its describes the number (overlapping number) of data appearance more
It is more, then it is believed that this feature describe data registration it is higher.
Step S520, describes data for the feature according to the height of registration and distributes corresponding weight, registration is higher
Then weight is higher.
That is, describing the registration height of data according to feature, the corresponding power of data distribution is described for each feature
Weight, the higher feature of registration describe data, then the weight for its distribution is higher.In a specific embodiment, can basis
Each feature that feature corresponding with the network object is described in data describes the registrations of data ranking from high to low, with weight
The mode successively decreased successively describes data for each feature and distributes corresponding weight, wherein, identical or similar feature describes number
According to same data are considered as, identical weight is distributed for it.
Step S530, describes data by the feature repeated and deletes, the feature descriptor data set for generating Weight closes.
Specifically, the corresponding feature of the network object can be described in data repeat feature describe data (including
Identical or similar feature describes data) one of them is only stayed, other data repeated are deleted, the feature for generating Weight is retouched
State data acquisition system, that is to say, that the corresponding feature of the network object is described into the same or similar feature in data and describes data
One of them is only taken, remaining deletion, i.e. there is no the same or similar feature in the feature descriptor data set of generation to describe number
According to.In this way, different features just to be described to the number of Data duplication appearance, the weight that different characteristic describes data is converted into, letter
Follow-up data handling procedure is changed.
Alternatively, step S540 can also be included after step S530, according to predetermined for heterogeneous networks
The weight that location is set, the weight that data are further described to the feature for coming from heterogeneous networks address are adjusted.
Specifically, weight can be set for heterogeneous networks address in advance, for example, hobby that can be according to user to website
Degree or degree of concern set the weight of different websites, that is to say, that the weight of different network address are set, for user
The network address that fancy grade is higher or degree of concern is higher, the weight corresponding to it can set higher.According to different nets
The weight that the weight of network address describes each feature data is adjusted, and has considered fancy grade or pass of the user to website
Influence of the note degree to quantized result, enables to the quantized value of each characteristic dimension that finally obtains more accurate.
, can be further to coming from heterogeneous networks according to the predetermined weight set for heterogeneous networks address
The weight that the feature of location describes data is adjusted, in other words, according to each feature describe data from different networks
Weight corresponding to address, the weight that data are described to each feature are adjusted, and specifically, can be retouched according to each feature
Ratio/ratio relation between the weight corresponding to the heterogeneous networks address of data is stated, adjusts the power that each feature describes data
Weight.
At step S120, the corresponding pass between data and different characteristic dimension is described according to the different characteristic pre-established
System, describes data by the feature corresponding with the network object and is matched with characteristic dimension, and obtained according to matching result
Each characteristic dimension weight corresponding with the network object.
Based on the above embodiments, step S120 may further include following step：
As shown in fig. 6, the different characteristic to be pre-established according to the basis of another embodiment of the present invention describe data with
Correspondence between different characteristic dimension, describes data by the feature corresponding with the network object and is carried out with characteristic dimension
The flow of the step of matching, and each characteristic dimension weight corresponding with the network object is obtained according to matching result (step S120)
Figure.
Step S610, according to the correspondence, describes data by different features and is matched respectively with characteristic dimension.
Specifically, the different characteristic that can be pre-established according to this describes corresponding between data and different characteristic dimension
Relation, describes data by acquisition different feature corresponding from the network object and is matched respectively with characteristic dimension.More
For body, it may be determined that the feature descriptor data set of the network object close in the feature pre-established describe data from it is different
Feature present in the correspondence of characteristic dimension time describes data, and there will be feature to describe data corresponding with this respectively
Each characteristic dimension in relation is matched, if a feature describes data describes data and different spies in the feature pre-established
Levy and exist in the correspondence of dimension time, then determine that describing data with this feature has correspondence according to the correspondence
One or more features dimension, and record the network object and the one or more features dimension is hit once respectively, so that really
The fixed corresponding different feature of the network object describes hit-count of the data to each characteristic dimension.
Step S620, the ratio of the hit-count to each characteristic dimension is determined according to matching result.
That is, hit of the data to each characteristic dimension time is described according to the corresponding different feature of the network object
Number, determines hit-count ratio of the network object to each characteristic dimension.
Step S630, the weight of data is described according to different characteristic, and the hit-count ratio of characteristic dimension will be adjusted
It is whole.
In step S530 or step S530~S540 above, the feature repeated data are described into and has deleted generation band
The feature descriptor data set of weight closes, that is to say, that there is no the number repeated during the feature descriptor data set of the network object closes
According to, and each feature describes data and has corresponding weight, therefore, the corresponding power of data can be described according to different characteristic
Weight, will be adjusted the ratio of the hit-count of characteristic dimension.Specifically, data can be described according to each different characteristic
Weight between proportionate relationship/ratio relation, adjust the hit-count ratio to each characteristic dimension.
Step S640, according to the ratio of the hit-count after adjustment, obtains each characteristic dimension corresponding with the network object
Weight.
That is, according to the hit-count ratio to each characteristic dimension after adjustment, determine that the network object is corresponding
The weight of each characteristic dimension.Specifically, can according to after adjustment to each characteristic dimension hit-count ratio from big to small
Sequentially, the weight of each characteristic dimension is determined in a manner of weight is successively decreased successively, or can be directly by each characteristic dimension after adjustment
Weight of the hit-count ratio as each characteristic dimension.And wherein, if the corresponding characteristic dimension of the network object includes
Multiple characteristic dimension classifications, then can determine each characteristic dimension in each affiliated classification respectively according to characteristic dimension classification
Weight.
At step S130, according to each characteristic dimension weight, the quantization of each characteristic dimension of the network object is obtained
Value.
Specifically, it can obtain and describe data with the corresponding score value of the network object, and will be opposite with the network object
The score value answered describes data and is multiplied respectively with each characteristic dimension weight, obtains the quantized value of each characteristic dimension of the network object.
That is, based on the score value of the network object describes data, data point are described into the network object corresponding score value
Not with each characteristic dimension multiplied by weight of the network object, quantized value of the obtained result respectively as each characteristic dimension.
By above-mentioned steps S110~S130, the quantized value of each characteristic dimension of network object can be obtained, so that more straight
The feature of some dimension of the reaction network object of sight.Meanwhile can be according to different network object types and to characteristic dimension
Increased or decreased, and and then the correspondence between description data and characteristic dimension is adjusted at any time.Therefore, for
High resilience and easily extended very much in the structure of quantized data of the relevant information of heterogeneous networks object after integration, be
More diversified exhibition method is provided and provides possibility.
According to one embodiment of present invention, which can also be into one
Step includes the quantized value of each characteristic dimension according to the obtained network object, therefrom chooses at least part characteristic dimension and is opened up
The step of showing.
Specifically, can according to the quantized value of each characteristic dimension of the obtained network object, therefrom selected part compared with
Excellent characteristic dimension (for example, quantized value higher characteristic dimension) is shown, for example, to Mr. Yu portion film, its " plot " dimension
The score value of degree is higher, then " plot " dimension can be carried out protruding displaying, for example, illustrating with word, image etc..
According to another embodiment of the invention, which can also be into
One step comprises the following steps：
First, from described with choosing at least part characteristic dimension in the corresponding each characteristic dimension of the network object.For example,
If the corresponding each characteristic dimension of the network object includes one or more features dimension classification, can therefrom choose at least one
Characteristic dimension classification, i.e. choose each characteristic dimension that at least one characteristic dimension classification is included.
Then, according to quantized value corresponding with the characteristic dimension of the selection is obtained, by the amount in a manner of multi-dimensional map
Change value is shown.
According to one embodiment of present invention, can according to the quantity of the characteristic dimension of selection, generate respective numbers from
The ray that same origin is sent but direction is different, further according to the corresponding quantized value of each characteristic dimension of selection, determines each
Position of the quantized value on corresponding ray, and the corresponding position of quantized value between adjacent ray is connected, generation shows the net
The multi-dimensional map of the quantized value of each characteristic dimension of network object.Wherein, the angle between adjacent ray is equal, so that generation is more
Dimension figure more intuitively reflects the magnitude relationship of each quantized value.
Specifically, the quantized value of each characteristic dimension according to unit length and the correspondence of quantized value, can be determined
The corresponding position on corresponding ray, reconnects the corresponding position of quantized value between adjacent ray, and the net is shown with generation
The multi-dimensional map of the quantized value of each characteristic dimension of network object.Wherein, the correspondence of unit length and quantized value can be set in advance
Put, if for example, quantized value is score value, can be with the corresponding score value of setting unit length.
More specifically, the quantization of each characteristic dimension can be determined according to unit length and the correspondence of quantized value
It is worth the quantity of corresponding unit length, i.e. determine that the quantized value of each characteristic dimension corresponds to how many a unit lengths respectively, so
Afterwards, according to the quantity of the corresponding unit length of quantized value of definite each characteristic dimension, determine each quantized value corresponding
Position on ray.Alternatively, multiple points can also be set on every ray, wherein, on every ray, each two is adjacent
Point between distance it is equal, in other words, the line segment length all same on every ray between two neighboring point, and each two phase
The length of line segment between adjacent point is equal to the unit length.It is corresponding according to the quantized value of definite each characteristic dimension
The quantity of unit length, you can determine the quantized value of each characteristic dimension point corresponding on its corresponding ray, the correspondence
Point be position of the quantized value of this feature dimension on corresponding ray.Determine each quantized value on corresponding ray
Behind position, the position corresponding to the quantized value between adjacent ray is connected, you can generation shows each feature dimensions of the network object
The multi-dimensional map of the quantized value of degree.Wherein, one can be encircled into after connecting the position corresponding to the quantized value between adjacent ray
Polygon, the quantity of characteristic dimension of the quantity on the side of polygon with choosing is identical, if for example, the characteristic dimension quantity chosen is
N, the then amount of radiation generated are N, if connecting and can enclosing the position corresponding to quantized value on every two adjacent rays
Form a N sides shape.The polygon being encircled into by this, is capable of the feature of each characteristic dimension of reaction network object.
It is alternatively possible to the region for connecting and being enclosed is set to be shown with foreground, that is to say, that by N bars
The N sides shape that position line on ray corresponding to quantized value is enclosed is shown with foreground.Pass through this display side
Formula, can be distinct represent the magnitude relationship of each characteristic dimension quantized value, and reflect some dimension of the network object
Feature.Fig. 8 is the quantized value of each characteristic dimension according to an embodiment of the invention to a film with the side of multi-dimensional map
Displaying examples of interfaces figure when formula is shown, as can be seen from Figure 8, the plot of the film, which is deduced, more to be protruded.
According to one embodiment of present invention, M unit length can also be chosen in every ray of generation, will be adjacent
The corresponding position of same units length between ray connects, and is shown with background colour.
Specifically, if the amount of radiation of generation is N, M unit length is chosen respectively on every ray of N bar rays
Degree, from the common origin of the ray of generation, every a unit length, by same units length pair on every two adjacent rays
The position answered connects, formed area gradually increases from the inside to the outside M N while shape (while quantity and the characteristic dimension of selection
Quantity it is identical), the M N sides shape, formed a net structure figure, the figure of formation can be shown with background colour,
Enable a user to more intuitively judge that each characteristic dimension corresponds to the magnitude relationship of quantized value, so as to know the network pair
The more prominent characteristic dimension of elephant.
According to one embodiment of present invention, can also be opened up in the side of the ray of generation (including upper and lower, left and right side)
Multiple figures are shown with, wherein each figure corresponds to a characteristic dimension not characterized by the ray.That is, penetrated unused
The characteristic dimension of line displaying, is shown with other figures, wherein each figure corresponds to a characteristic dimension.So that user is also
It will appreciate that the network object others characteristic dimension.
Alternatively, the corresponding characteristic dimension of each figure can be more excellent in the characteristic dimension classification belonging to this feature dimension
Characteristic dimension, i.e. each figure corresponds in one characteristic dimension classification of displaying preferably dimension, for example, can be such as Fig. 8 institutes
Show, in the figure shown below the ray of generation, the figure of " Libra favorite " represents be used as characteristic dimension using constellation
Dimension classification in, the user of Libra may be the user for most liking the film, and the figure of " male audience is in the majority " represents to exist
In dimension classification using gender as characteristic dimension, liking may be in the majority with male audience in the user of the film.This mode
More prominent dimension of the network object in other each different dimensions classifications can be embodied, so as to from a variety of differences
Angle, provides the decision information of selection network object to the user.
, can also be according to the control instruction received, by the multiple figure along orientation at displaying interface based on this
It is middle to roll displaying.That is, according to the control instruction received, by multiple other characteristic dimensions of displaying in interface is shown
Figure is scrolled along the direction (such as direction or from left to right direction from top to bottom) of arrangement, wherein, the control received
System instruction, can be instruction input by user, or pre-set control instruction, for example, pre-setting when user selects
When selecting the multiple features dimension present graphical for checking network object, that is, scrolled, alternatively, carrying out once at predetermined time intervals
Scrolling display.
The quantized value of each characteristic dimension of network object is shown by the exhibition method of above-mentioned multi-dimensional map, Neng Gouzhi
The feature of each characteristic dimension of ground reaction network object is seen, so as to obtain the decision information of network object as user.
Present invention also offers a kind of device that multiple features dimension quantization is carried out to network object.As shown in fig. 7, Fig. 7 is
The structure diagram of the device 700 according to an embodiment of the invention that multiple features dimension quantization is carried out to network object.
Device 700 includes：Obtain module 710, matching module 720 and quantization modules 730.
Obtain module 710 and be determined for a network object, obtain and describe number with the corresponding feature of the network object
According to.Matching module 720 can be used for being described according to the different characteristic pre-established corresponding between data and different characteristic dimension
Relation, describes data by the feature corresponding with the network object and is matched with characteristic dimension, and is obtained according to matching result
Obtain each characteristic dimension weight corresponding with the network object.Quantization modules 730 can be used for according to each characteristic dimension weight,
Obtain the quantized value of each characteristic dimension of the network object.
In the apparatus according to the invention 700, the acquisition module 710 can include：Determining module and acquisition module.
Wherein it is determined that module can be used for the type according to the network object, predefine one or more network and search
Rope address.Acquisition module can be used for obtaining by the web search address describes number with the corresponding feature of the network object
According to.
According to another embodiment of the invention, the acquisition module 710 can include：First registration matching module and
First generation module.
Wherein, the first registration matching module can be used for the feature for coming from heterogeneous networks address describing data progress
Registration matches.First generation module can be used for deleting description data of the registration less than threshold value, generate final feature
Descriptor data set closes.
According to one embodiment of present invention, the matching module 720 can include：First matching module, the first ratio
Determining module and the first weight obtain module.
Specifically, the first matching module can be used for according to the correspondence, by whole features describe data respectively with
Characteristic dimension is matched, and repeated matching is then carried out if the feature repeated describes data.First ratio-dependent module can be with
For determining the ratio of the hit-count to each characteristic dimension according to matching result.First weight obtains module and can be used for basis
The ratio of the hit-count, obtains each characteristic dimension weight corresponding with the network object.
According to another embodiment of the invention, the acquisition module 710 can include：Second registration matching module is weighed
Reallocation module and the second generation module.
Wherein, the second registration matching module can be used for the feature for coming from heterogeneous networks address describing data progress
Registration matches.Weight distribution module can be used for describing the corresponding power of data distribution according to the height of registration for the feature
Weight, the more high then weight of registration are higher.Second generation module can be used for the feature repeated describing data deletion, generate cum rights
The feature descriptor data set of weight closes.
Further, in the apparatus according to the invention 700, the acquisition module 710 can further include：Weight
Adjust module, the module can be used for according to it is predetermined for heterogeneous networks address set weight, further to from
The weight that feature in heterogeneous networks address describes data is adjusted.
According to another embodiment of the invention, the matching module can include：Second matching module, the second ratio are true
Cover half block, ratio adjustment module and the second weight obtain module.
Wherein, the second matching module can be used for according to the correspondence, by different features describe data respectively with
Characteristic dimension is matched.Second ratio-dependent module can be used for determining the hit time to each characteristic dimension according to matching result
Several ratios.Ratio adjustment module can be used for the weight that data are described according to different characteristic, by the hit time to characteristic dimension
Number ratio is adjusted.Second weight obtains module, for the ratio according to the hit-count after adjustment, obtains and the network pair
As corresponding each characteristic dimension weight.
In the apparatus according to the invention 700, the characteristic dimension is determined according to the type of the network object.
In the apparatus according to the invention 700, the quantization modules 730 can include：At score value acquisition module and quantization
Manage module.
Wherein, score value acquisition module, which can be used for obtaining corresponding with network object score value, describes data；Quantification treatment
Module can be used for that data will be described with the corresponding score value of the network object being multiplied respectively with each characteristic dimension weight, obtain
To the quantized value of each characteristic dimension of the network object.
Score value acquisition module, which can be further used for obtaining, comes from the corresponding with the network object of heterogeneous networks address
Score value describes data, and according to the predetermined weight set for heterogeneous networks address, will come from heterogeneous networks address
Score value describe data and be weighted summation, data are described as final score value.
In the apparatus according to the invention 700, the feature describe data can include label data, comment data, and/
Or key data.
In the apparatus according to the invention 700, it can further include：Choose module and the first display module.
Specifically, module is chosen to can be used for from described with choosing at least in the corresponding each characteristic dimension of the network object
Partial Feature dimension.First display module can be used for according to obtained quantized value corresponding with the characteristic dimension of the selection, with
The quantized value is shown by the mode of multi-dimensional map.
In the apparatus according to the invention 700, it can further include：Second display module, obtains for basis
The quantized value of each characteristic dimension of the network object, therefrom chooses at least part characteristic dimension and is shown.
In the apparatus according to the invention 700, the second display module can be further used for：Setting connects enclosing shape
Into region be shown with foreground.
In the apparatus according to the invention 700, the second display module may further include：Choose submodule be connected and
Show submodule.Wherein, submodule is chosen to can be used for choosing M unit length in every ray of generation.Connection and exhibition
Show that module can be used for connecting the corresponding position of same units length between adjacent ray, and opened up with background colour
Show.
In the apparatus according to the invention 700, the angle between the adjacent ray is identical.
In the apparatus according to the invention 700, the second display module can further include：Figure shows submodule,
For the side of the ray in generation, displaying has multiple figures, wherein what each figure correspondence one was not characterized by the ray
Characteristic dimension.
According to one embodiment of present invention, the figure displaying submodule can also be further used for：According to receiving
Control instruction, the multiple figure rolls displaying along orientation in interface is shown.
The device described above that multiple features dimension quantization is carried out to network object with before describe to network object into
The processing for the method that row multiple features dimension quantifies is corresponding, accordingly, with respect to more detailed ins and outs, is retouched before may refer to
The method stated.
Algorithm and display be not inherently related to any certain computer, virtual system or miscellaneous equipment provided herein.
Various general-purpose systems can also be used together with teaching based on this.As described above, required by constructing this kind of system
Structure be obvious.In addition, the present invention is not also directed to any certain programmed language.It should be understood that it can utilize various
Programming language realizes the content of invention described herein, and the description done above to language-specific is to disclose this hair
Bright preferred forms.
In the specification that this place provides, numerous specific details are set forth.It is to be appreciated, however, that the implementation of the present invention
Example can be put into practice in the case of these no details.In some instances, known method, structure is not been shown in detail
And technology, so as not to obscure the understanding of this description.
Similarly, it will be appreciated that in order to simplify the disclosure and help to understand one or more of each inventive aspect,
Above in the description to the exemplary embodiment of the present invention, each feature of the invention is grouped together into single implementation sometimes
In example, figure or descriptions thereof.However, the method for the disclosure should be construed to reflect following intention：I.e. required guarantor
The application claims of shield features more more than the feature being expressly recited in each claim.It is more precisely, such as following
Claims reflect as, inventive aspect is all features less than single embodiment disclosed above.Therefore,
Thus the claims for following embodiment are expressly incorporated in the embodiment, wherein each claim is in itself
Separate embodiments all as the present invention.
Those skilled in the art, which are appreciated that, to carry out adaptively the module in the equipment in embodiment
Change and they are arranged in one or more equipment different from the embodiment.Can be the module or list in embodiment
Member or component be combined into a module or unit or component, and can be divided into addition multiple submodule or subelement or
Sub-component.In addition at least some in such feature and/or process or unit exclude each other, it can use any
Combination is disclosed to all features disclosed in this specification (including adjoint claim, summary and attached drawing) and so to appoint
Where all processes or unit of method or equipment are combined.Unless expressly stated otherwise, this specification (including adjoint power
Profit requires, summary and attached drawing) disclosed in each feature can be by providing the alternative features of identical, equivalent or similar purpose come generation
Replace.
In addition, it will be appreciated by those of skill in the art that although some embodiments described herein include other embodiments
In included some features rather than further feature, but the combination of the feature of different embodiments means in of the invention
Within the scope of and form different embodiments.For example, in the following claims, embodiment claimed is appointed
One of meaning mode can use in any combination.
The all parts embodiment of the present invention can be with hardware realization, or to be run on one or more processor
Software module realize, or realized with combinations thereof.It will be understood by those of skill in the art that it can use in practice
Microprocessor or digital signal processor (DSP) are according to embodiments of the present invention to network object progress multiple features dimension to realize
The some or all functions of some or all components in the device of metrization.The present invention is also implemented as being used to perform
The some or all equipment or program of device of method as described herein are (for example, computer program and computer journey
Sequence product).It is such realize the present invention program can store on a computer-readable medium, either can have one or
The form of multiple signals.Such signal can be downloaded from internet website and obtained, either provide on carrier signal or
There is provided in the form of any other.
It should be noted that the present invention will be described rather than limits the invention for above-described embodiment, and ability
Field technique personnel can design alternative embodiment without departing from the scope of the appended claims.In the claims,
Any reference symbol between bracket should not be configured to limitations on claims.Word "comprising" does not exclude the presence of not
Element or step listed in the claims.Word "a" or "an" before element does not exclude the presence of multiple such
Element.The present invention can be by means of including the hardware of some different elements and being come by means of properly programmed computer real
It is existing.In if the unit claim of equipment for drying is listed, several in these devices can be by same hardware branch
To embody.The use of word first, second, and third does not indicate that any order.These words can be explained and run after fame
Claim.
Claims (26)
1. a kind of method that multiple features dimension quantization is carried out to network object, including：
Determine a network object, obtain and describe data with the corresponding feature of the network object；
Different characteristic according to pre-establishing describes the correspondence between data and different characteristic dimension, will the described network pair
Matched as corresponding feature describes data characteristic dimension corresponding with the described network object, and obtained according to matching result
Each characteristic dimension weight corresponding with the network object；
According to each characteristic dimension weight, the quantized value of each characteristic dimension of the network object is obtained.
2. the method as described in claim 1, it is characterised in that obtain and describe data with the corresponding feature of the network object
Step, including：
According to the type of the network object, one or more web search address is predefined；
Obtained by the web search address and describe data with the corresponding feature of the network object.
3. the method as described in claim 1, it is characterised in that obtain and describe data with the corresponding feature of the network object
Step, including：
The feature for coming from heterogeneous networks address is described into data and carries out registration matching；
Description data of the registration less than threshold value are deleted, final feature descriptor data set is generated and closes.
4. the method as described in claim 1, it is characterised in that obtain and describe data with the corresponding feature of the network object
Step, including：
The feature for coming from heterogeneous networks address is described into data and carries out registration matching；
Data are described for the feature according to the height of registration and distribute corresponding weight, the more high then weight of registration is higher；
The feature repeated is described data to delete, the feature descriptor data set for generating Weight closes.
5. method as claimed in claim 4, it is characterised in that obtain and describe data with the corresponding feature of the network object
Step, further comprises：
According to the predetermined weight set for heterogeneous networks address, further to coming from the feature of heterogeneous networks address
The weight of description data is adjusted.
6. the method as described in claim 1, it is characterised in that the characteristic dimension is true according to the type of the network object
Fixed.
7. the method as described in claim 1 or 3, it is characterised in that
Different characteristic according to pre-establishing describes the correspondence between data and different characteristic dimension, will the described network pair
Matched as corresponding feature describes data characteristic dimension corresponding with the described network object, and obtained according to matching result
The step of each characteristic dimension weight corresponding with the network object, including：
According to the correspondence, whole features are described into data and are matched respectively with characteristic dimension, if the spy repeated
Sign description data then carry out repeated matching；
The ratio of the hit-count to each characteristic dimension is determined according to matching result；
According to the ratio of the hit-count, each characteristic dimension weight corresponding with the network object is obtained.
8. method as claimed in claim 4, it is characterised in that
Different characteristic according to pre-establishing describes the correspondence between data and different characteristic dimension, will the described network pair
Matched as corresponding feature describes data characteristic dimension corresponding with the described network object, and obtained according to matching result
The step of each characteristic dimension weight corresponding with the network object, including：
According to the correspondence, different features is described into data and is matched respectively with characteristic dimension；
The ratio of the hit-count to each characteristic dimension is determined according to matching result；
The weight of data is described according to different characteristic, the hit-count ratio of characteristic dimension will be adjusted；
According to the ratio of the hit-count after adjustment, each characteristic dimension weight corresponding with the network object is obtained.
9. the method as described in claim 1, it is characterised in that according to each characteristic dimension weight, obtain the network object
Each characteristic dimension quantized value the step of, including：
Obtain score value corresponding with the network object and describe data；
Data will be described with the corresponding score value of the network object to be multiplied respectively with each characteristic dimension weight, obtain the network
The quantized value of each characteristic dimension of object.
10. method as claimed in claim 9, it is characterised in that obtain score value corresponding with the network object and describe data
Step, including：
The score value corresponding with the network object that acquisition comes from heterogeneous networks address describes data, and according to predetermined pin
The weight set to heterogeneous networks address, describes data by the score value for coming from heterogeneous networks address and is weighted summation, as
Final score value describes data.
11. the method as described in claim 1, it is characterised in that it is described description data include label data, comment data, and/
Or key data.
12. the method as described in claim 1, it is characterised in that further comprise：
From described with choosing at least part characteristic dimension in the corresponding each characteristic dimension of the network object；
According to quantized value corresponding with the characteristic dimension of the selection is obtained, the quantized value is opened up in a manner of multi-dimensional map
Show.
13. the method as described in claim 1, it is characterised in that further comprise：
According to the quantized value of each characteristic dimension of the obtained network object, therefrom choose at least part characteristic dimension and opened up
Show.
14. a kind of device that multiple features dimension quantization is carried out to network object, including：
Module is obtained, for determining a network object, obtains and describes data with the corresponding feature of the network object；
Matching module, for describing the correspondence between data and different characteristic dimension according to the different characteristic pre-established,
The corresponding feature of the described network object is described data characteristic dimension corresponding with the described network object to be matched, and root
Each characteristic dimension weight corresponding with the network object is obtained according to matching result；
Quantization modules, for according to each characteristic dimension weight, obtaining the quantized value of each characteristic dimension of the network object.
15. device as claimed in claim 14, it is characterised in that the acquisition module includes：
Determining module, for the type according to the network object, predefines one or more web search address；
Acquisition module, data are described for being obtained by the web search address with the corresponding feature of the network object.
16. device as claimed in claim 14, it is characterised in that the acquisition module includes：
First registration matching module, registration matching is carried out for the feature for coming from heterogeneous networks address to be described data；
First generation module, for description data of the registration less than threshold value to be deleted, generates final feature descriptor data set
Close.
17. device as claimed in claim 14, it is characterised in that the acquisition module includes：
Second registration matching module, registration matching is carried out for the feature for coming from heterogeneous networks address to be described data；
Weight distribution module, corresponding weight, registration are distributed for describing data according to the height of registration for the feature
More high then weight is higher；
Second generation module, is deleted for the feature repeated to be described data, and the feature descriptor data set for generating Weight closes.
18. device as claimed in claim 17, it is characterised in that the acquisition module further comprises：
Weight adjust module, for according to it is predetermined for heterogeneous networks address set weight, further to coming from
The weight that the feature of heterogeneous networks address describes data is adjusted.
19. device as claimed in claim 14, it is characterised in that the characteristic dimension is the type according to the network object
Definite.
20. the device as described in claim 14 or 16, it is characterised in that the matching module includes：
Whole features for according to the correspondence, are described data and are carried out respectively with characteristic dimension by the first matching module
Match somebody with somebody, repeated matching is then carried out if the feature repeated describes data；
First ratio-dependent module, for determining the ratio of the hit-count to each characteristic dimension according to matching result；
First weight obtains module, for the ratio according to the hit-count, obtains each feature corresponding with the network object
Dimension weight.
21. device as claimed in claim 17, it is characterised in that the matching module includes：
Second matching module, for according to the correspondence, different features being described data and is carried out respectively with characteristic dimension
Matching；
Second ratio-dependent module, for determining the ratio of the hit-count to each characteristic dimension according to matching result；
Ratio adjust module, for describing the weight of data according to different characteristic, by the hit-count ratio of characteristic dimension into
Row adjustment；
Second weight obtains module, for the ratio according to the hit-count after adjustment, obtains corresponding with the network object each
Characteristic dimension weight.
22. device as claimed in claim 14, it is characterised in that the quantization modules, including：
Score value acquisition module, data are described for obtaining score value corresponding with the network object；
Quantification treatment module, for data and each characteristic dimension weight point will to be described with the corresponding score value of the network object
It is not multiplied, obtains the quantized value of each characteristic dimension of the network object.
23. device as claimed in claim 22, it is characterised in that the score value acquisition module, is further used for：
The score value corresponding with the network object that acquisition comes from heterogeneous networks address describes data, and according to predetermined pin
The weight set to heterogeneous networks address, describes data by the score value for coming from heterogeneous networks address and is weighted summation, as
Final score value describes data.
24. device as claimed in claim 14, it is characterised in that the description data include label data, comment data,
And/or key data.
25. device as claimed in claim 14, it is characterised in that further comprise：
Module is chosen, for choosing at least part characteristic dimension from the described and corresponding each characteristic dimension of the network object；
First display module, the quantized value corresponding with the characteristic dimension of the selection obtained for basis, in a manner of multi-dimensional map
The quantized value is shown.
26. device as claimed in claim 14, it is characterised in that further comprise：
Second display module, for the quantized value of each characteristic dimension according to the obtained network object, therefrom chooses at least portion
Point characteristic dimension is shown.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410564598.7A CN104331429B (en) | 2014-10-21 | 2014-10-21 | The method and device of multiple features dimension quantization is carried out to network object |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410564598.7A CN104331429B (en) | 2014-10-21 | 2014-10-21 | The method and device of multiple features dimension quantization is carried out to network object |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104331429A CN104331429A (en) | 2015-02-04 |
| CN104331429B true CN104331429B (en) | 2018-04-27 |
Family
ID=52406156
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410564598.7A Active CN104331429B (en) | 2014-10-21 | 2014-10-21 | The method and device of multiple features dimension quantization is carried out to network object |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104331429B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106033457B (en) * | 2015-03-18 | 2019-10-18 | 阿里巴巴集团控股有限公司 | A kind of method and apparatus of the attribute information of determining fruit objective attribute target attribute |
| CN106897208A (en) * | 2015-12-18 | 2017-06-27 | 阿里巴巴集团控股有限公司 | A kind of quantization method and device of application software Consumer's Experience |
| CN107749006A (en) * | 2017-11-01 | 2018-03-02 | 广州爱九游信息技术有限公司 | Game appraisal procedure, device and equipment |
| CN111177515B (en) * | 2019-12-25 | 2023-07-21 | 中国平安财产保险股份有限公司 | Product label matching method, device, computer equipment and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101211339A (en) * | 2006-12-29 | 2008-07-02 | 上海芯盛电子科技有限公司 | Intelligent web page classifier based on user behaviors |
| WO2010044780A1 (en) * | 2008-10-14 | 2010-04-22 | Hewlett-Packard Development Company, L.P. | Dynamic content sorting using tags |
| CN102902790A (en) * | 2012-09-29 | 2013-01-30 | 北京奇虎科技有限公司 | Web page classification system and method |
| CN102955813A (en) * | 2011-08-29 | 2013-03-06 | 中国移动通信集团四川有限公司 | Information searching method and information searching system |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101551806B (en) * | 2008-04-03 | 2012-04-18 | 北京搜狗科技发展有限公司 | Personalized website navigation method and system |
| CN103886090B (en) * | 2014-03-31 | 2018-01-02 | 北京搜狗科技发展有限公司 | Content recommendation method and device based on user preferences |
-
2014
- 2014-10-21 CN CN201410564598.7A patent/CN104331429B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101211339A (en) * | 2006-12-29 | 2008-07-02 | 上海芯盛电子科技有限公司 | Intelligent web page classifier based on user behaviors |
| WO2010044780A1 (en) * | 2008-10-14 | 2010-04-22 | Hewlett-Packard Development Company, L.P. | Dynamic content sorting using tags |
| CN102955813A (en) * | 2011-08-29 | 2013-03-06 | 中国移动通信集团四川有限公司 | Information searching method and information searching system |
| CN102902790A (en) * | 2012-09-29 | 2013-01-30 | 北京奇虎科技有限公司 | Web page classification system and method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104331429A (en) | 2015-02-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110020094A (en) | A kind of methods of exhibiting and relevant apparatus of search result | |
| CN102306171B (en) | A kind of for providing network to access suggestion and the method and apparatus of web search suggestion | |
| CN107423355A (en) | Information recommendation method and device, electronic equipment | |
| CN102968413B (en) | A kind of method and apparatus for being used to provide search result | |
| CN107229666B (en) | A kind of interest heuristic approach and device based on recommender system | |
| CN104331429B (en) | The method and device of multiple features dimension quantization is carried out to network object | |
| CN106688215A (en) | Automated click type selection for content performance optimization | |
| CN102354315A (en) | Generation method of site navigation page and device thereof | |
| CN102831117B (en) | Select font, the determination of font, recommendation, generation method and equipment thereof | |
| CN105874449A (en) | Systems and methods for extracting and generating images for display content | |
| CN106484766A (en) | Searching method based on artificial intelligence and device | |
| CN110267097A (en) | Video pushing method, device and electronic equipment based on characteristic of division | |
| CN103514282A (en) | Method and device for displaying search results of videos | |
| CN105224657B (en) | A kind of information recommendation method and electronic equipment based on search engine | |
| CN110110201A (en) | A kind of content recommendation method and system | |
| CN103473282B (en) | A kind of apparatus and method generating the Hot Contents page | |
| CN106354855A (en) | Recommendation method and system | |
| CN104699844B (en) | The method and device of video tab is determined for advertisement | |
| CN104199865B (en) | Searching method, client and the system of the customization result of content providers are provided | |
| CN102073670A (en) | Method, equipment and system for debugging online webpage template | |
| CN110278447A (en) | Video pushing method, device and electronic equipment based on continuous feature | |
| CN105139233A (en) | Advertisement putting method, device, and system | |
| CN109075987A (en) | Optimize digital assembly analysis system | |
| CN110688578A (en) | Screen locking wallpaper recommendation method and device and electronic equipment | |
| CN109509051A (en) | A kind of item recommendation method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20220803 Address after: Room 801, 8th floor, No. 104, floors 1-19, building 2, yard 6, Jiuxianqiao Road, Chaoyang District, Beijing 100015 Patentee after: BEIJING QIHOO TECHNOLOGY Co.,Ltd. Address before: 100088 room 112, block D, 28 new street, new street, Xicheng District, Beijing (Desheng Park) Patentee before: BEIJING QIHOO TECHNOLOGY Co.,Ltd. Patentee before: Qizhi software (Beijing) Co.,Ltd. |
|
| TR01 | Transfer of patent right |