CN104731867B

CN104731867B - A kind of method and apparatus that object is clustered

Info

Publication number: CN104731867B
Application number: CN201510090184.XA
Authority: CN
Inventors: 周泽伟; 程涛远
Original assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Current assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Priority date: 2015-02-27
Filing date: 2015-02-27
Publication date: 2018-09-07
Anticipated expiration: 2035-02-27
Also published as: CN104731867A

Abstract

The present invention designs the method clustered to object in computer equipment, and this method includes：The transfer case information of multiple objects is obtained, the transfer case information is used to indicate transfer case based on object information acquisition behavior, that user is in the multiple object；According to the transfer case information, the multiple object is clustered, obtains the cluster result of the multiple object.The present invention can cluster object by analyzing transfer case information of the user in object, therefore identified object classification is more objective, accurate.

Description

A kind of method and apparatus that object is clustered

Technical field

The present invention relates to field of computer technology more particularly to a kind of method and apparatus clustered to object.

Background technology

In the prior art, natural language analysis is usually carried out by the description text to object, to classify to object. Particularly, when object is related to commercial use, such as when object is brand, in addition to brand name carry out natural language analysis with Outside, can also be in conjunction with the data from object angle, industry and region, the sales situation of brand and market as belonging to brand need The factors such as seek, to classify to brand.

Invention content

The mesh of the present invention includes providing a kind of method and apparatus clustered to object.

According to an aspect of the present invention, a kind of method for being clustered to object in computer equipment is provided, Wherein, this method includes：

The transfer case information of multiple objects is obtained, the transfer case information is used to indicate to be obtained based on object information and be gone For, transfer case that user is in the multiple object；

According to the transfer case information, the multiple object is clustered, obtains the cluster knot of the multiple object Fruit.

According to another aspect of the present invention, it additionally provides a kind of for being clustered to object in computer equipment Device, wherein the device includes：

Device for the transfer case information for obtaining multiple objects, the transfer case information are used to indicate based on object Information acquirement behavior, transfer case that user is in the multiple object；

For according to the transfer case information, being clustered to the multiple object, the poly- of the multiple object is obtained The device of class result.

Compared with prior art, the present invention has the following advantages：1) the solution of the present invention has broken the prejudice of this field, energy It is enough that object is clustered by analyzing transfer case information of the user in object；2) number from object angle is compared According to the scheme carried out to object in the present invention by analyzing transfer case of the user in multiple objects, closer to user angle Degree, more can intuitively reflect understanding of the user to object, therefore, object classification determined by the solution of the present invention is more objective, Accurately；Even if 3) in the data from user perspective, transfer case information of the invention is not common data yet, true On, if clearly referring to the data from user perspective, those skilled in the art are easier it is envisioned that from the user direct Evaluation (such as gives a mark, comments on word).

Description of the drawings

By reading a detailed description of non-restrictive embodiments in the light of the attached drawings below, of the invention other Feature, objects and advantages will become more apparent upon：

Fig. 1 is the method flow schematic diagram clustered to object of a preferred embodiment of the invention；

Fig. 2 is the structural schematic diagram of the clustering apparatus clustered to object of a preferred embodiment of the invention；

Fig. 3 shows the schematic diagram of transfer path of the user of a preferred embodiment in multiple objects；

Fig. 4 shows the schematic diagram of transfer path of the user of a preferred embodiment in multiple keywords；

Fig. 5 shows the transfer path of a reticular structure from keyword to the transfer path of the reticular structure of object The specific example of conversion；

Fig. 6 show preferred embodiment, from an object to the schematic diagram of the transfer of multiple objects；

Fig. 7 shows a specific example of Fig. 6.

Same or analogous reference numeral represents same or analogous component in attached drawing.

Specific implementation mode

Present invention is further described in detail below in conjunction with the accompanying drawings.

Fig. 1 is the method flow schematic diagram clustered to object of a preferred embodiment of the invention.Wherein, this implementation The method of example mainly realizes that the computer equipment includes the network equipment and user equipment by computer equipment.The network Equipment includes but not limited to single network server, the server group of multiple network servers composition or is based on cloud computing (Cloud Computing the cloud being made of a large amount of computers or network server), wherein cloud computing is one kind of Distributed Calculation, by One super virtual computer of the computer collection composition of a group loose couplings；Network residing for the network equipment includes but not It is limited to internet, wide area network, Metropolitan Area Network (MAN), LAN, VPN network etc..The user equipment includes but not limited to PC machine, tablet electricity Brain, smart mobile phone, PDA, IPTV etc..

It should be noted that the computer equipment and network are only for example, other are existing or are likely to occur from now on Computing device or network are such as applicable to the present invention, should also be included within the scope of the present invention, and include by reference In this.

Method according to the present embodiment includes step S1 and step S2.

In step sl, computer equipment obtains the transfer case information of multiple objects.

Wherein, the object may include any object that can be clustered.Preferably, the object has commerciality Matter.It is highly preferred that the object includes brand.

Wherein, the transfer case information is used to indicate obtains behavior, user the multiple right based on object information Transfer case as in.Wherein, the object information obtains the behavior that behavior includes any information that can be used in obtaining object； For example, it includes obtaining the behavior of object information by searching for keyword related with object that the object information, which obtains behavior,； In another example it includes obtaining object information by clicking and browsing content related with object that the object information, which obtains behavior, Behavior.Wherein, described " obtaining behavior based on object information " indicates that the transfer case reflects user and obtained in object information The transfer case generated in behavior, it is preferable that the transfer case needs to obtain behavior based on object information to be determined；Example Such as, the object search changed in search behavior by counting multiple users, or by counting multiple users in search behavior Change with the associated search key of object, to determine transfer case information etc. of the user in middle object.

Preferably, the transfer case information of the multiple object includes but not limited at least one of following：

1) transfer path information of the user in multiple objects.

Wherein, the transfer path information indicates transfer path of the user in multiple objects.For example, right there are three As Object1, Object2 and Object3, transfer path information indicates transfer path packet of multiple users in three objects It includes：It is transferred to Object2 from Object1, and, it is transferred to Object3 from Object1.

2) transfer number information of the user between each object.

Wherein, the transfer number information indicates transfer number of the user between each object.For example, there are three Object Object1, Object2 and Object3, transfer number information indicate transfer of multiple users between three objects Number includes：It is transferred to Object2 five times from Object1, and, it is transferred to Object3 eight times from Object1.

3) transition probability information of the user between each object.

Wherein, the transition probability information indicates transition probability of the user between each object.For example, there are three Object Object1, Object2 and Object3, transition probability information indicate transfer of multiple users between three objects Probability includes：The probability that Object2 is transferred to from Object1 is 38.46%, and, it is transferred to Object3's from Object1 Probability is 61.54%.

It should be noted that transfer path may be not present between partial objects in multiple objects, (i.e. user is in object Do not carried out transfer in information acquirement behavior between such partial objects), then the transfer number between such partial objects and Transition probability is zero.Furthermore, it is possible to the case where in the presence of the object itself is transferred to from an object；For example, user is searching for Possible continuous several times search for the information of the same object using different search keys in behavior, to generate from an object The case where being transferred to the object itself.

Preferably, which can be used a variety of storage modes.

For example, the transfer case information storage is table, and transfer road of the user in multiple objects is had recorded in table The transfer number and transition probability of diameter and user between each object, as shown in table 1 below.

Transfer path	Transfer number	Transition probability
			Object1→Object2	5	38.46%
Object1→Object3	8	61.54%

Table 1

In another example the transfer case information includes：It is stored as the transfer path of reticular structure, and, in the reticular structure Transfer number and/or transition probability between each node (i.e. between each object).Such as extremely for 9 object Object1 Object9, the transfer case information of 9 objects include transfer path as shown in Figure 3, and, there is arrow connection in Fig. 3 Each node between (such as from Object1 to Object2, from Object1 to Object3, from Object1 to Object4 Deng) transfer number and/or transition probability.

It should be noted that the above-mentioned examples are merely illustrative of the technical solutions of the present invention, rather than to the limit of the present invention System, it should be appreciated by those skilled in the art that any be used to indicate obtains behavior, user the multiple right based on object information The transfer case information of transfer case as in, should be included in the scope of the present invention.

Specifically, the mode that computer equipment obtains the transfer case information of multiple objects includes but not limited to：

1) computer equipment directly acquires the transfer case information of pre-determining, multiple object.

For example, computer equipment reads the transfer case of pre-determining, multiple object from local or other equipment Information.

2) step S1 further comprises step S11 and step S12.

In step s 11, computer equipment obtains the transfer case information of multiple keywords.

Wherein, the transfer case information of the multiple keyword is used to indicate obtains behavior, user based on object information Transfer case in multiple keywords.Preferably, the multiple keyword is related to the object in object information acquisition behavior Connection；If for example, object information obtain behavior be object search behavior, keyword can be the search behavior in it is input by user or The search key etc. of selection.

Preferably, the transfer case information of the multiple keyword includes at least one of following：

A) transfer path information of the user in the multiple keyword.

Wherein, transfer path information of the user in the multiple keyword indicates user and turns in multiple keywords Move path.For example, there are three keywords Query1, Query2 and Query3, transfer path information indicates multiple users at this Transfer path in three keywords includes：It is transferred to Query2 from Query1, and, it is transferred to Query3 from Query1.

B) transfer number information of the user between each keyword.

Wherein, transfer number information of the user between each keyword indicates user and turns between each keyword Move number.For example, there are three keywords Query1, Query2 and Query3, transfer number information indicates multiple users at this Transfer number between three keywords includes：It is transferred to Query2 five times from Query1, and, it is transferred to from Query1 Query3 eight times.

It should be noted that transfer path may be not present between Partial key word in multiple keywords, (i.e. user exists In object information acquisition behavior transfer was not carried out between such Partial key word), then between such Partial key word Transfer number is zero.

Preferably, a variety of storage modes can be used in the transfer case information of multiple keyword.

For example, the transfer case information storage is table, and transfer of the user in multiple keywords is had recorded in table The transfer number of path and user between each keyword, as shown in table 2 below.

Transfer path	Transfer number
		Query1→Query2	5
Query1→Query3	8

Table 2

In another example the transfer case information includes：It is stored as the transfer path of reticular structure, and, in the reticular structure Transfer number between each node (between i.e. each keyword).Such as 9 keyword Query1 to Query9, this 9 The transfer case information of object includes transfer path as shown in Figure 4, and, each node with arrow connection in Fig. 4 it Between transfer number.

It should be noted that the above-mentioned examples are merely illustrative of the technical solutions of the present invention, rather than to the limit of the present invention System, it should be appreciated by those skilled in the art that any be used to indicate obtains behavior, user in multiple keywords based on object information In transfer case transfer case information, should be included in the scope of the present invention.

Specifically, the realization method that computer equipment obtains the transfer case information of multiple keywords includes but not limited to：

A) computer equipment directly acquires the transfer case information of pre-determining, multiple keyword.

For example, computer equipment reads the transfer feelings of pre-determining, multiple keyword from local or other equipment Condition information.

B) computer equipment obtains the keyword concern record of at least one user, and is paid close attention to and remembered according to the keyword Record, determines the transfer case information of the multiple keyword.

Wherein, keyword concern record includes the key that the multiple user paid close attention in object information acquisition behavior The temporal information that word and the keyword are concerned.Preferably, it includes search behavior that object information, which obtains behavior, the concern The keyword crossed includes searched keyword；Preferably, it includes navigation patterns that object information, which obtains behavior, described to pay close attention to Keyword includes being clicked come browsing the keyword of contents of object.

Record is paid close attention to preferably for the keyword of each user, computer equipment is according in keyword concern record Including the temporal information that is concerned of keyword, determine between transfer path and each keyword of the user in keyword Transfer number；Also, computer equipment is by merging transfer path and each keyword of each user in keyword Between transfer number, determine the transfer case information of the multiple keyword.

For example, computer equipment obtains the keyword concern record of user A and user B；Wherein, user A and user B Keyword concern record is respectively as shown in the following table 3 and table 4：

The keyword of concern	The time that keyword is concerned
		Query1	2014-12-13-10:40
Query3	2014-12-13-10:36

Table 3

The keyword of concern	The time that keyword is concerned
		Query1	2014-11-10-00:14
Query2	2014-11-10-00:23

Table 4

Then the keyword of user A is paid close attention to and is recorded, computer equipment determines transfer path packets of the user A in keyword It includes " Query1 → Query3 ", and the transfer number of " Query1 → Query3 " is 1；Similarly, computer equipment determines user B Transfer path in keyword includes " Query1 → Query2 ", and the transfer number of " Query1 → Query2 " is 1.It connects It, computer equipment merges the transfer number between the transfer path and each keyword of user A and B in keyword, really The transfer case information of fixed the multiple keyword is as shown in table 5 below.

Transfer path	Transfer number
		Query1→Query2	1
Query1→Query3	1

Table 5

It should be noted that the above-mentioned examples are merely illustrative of the technical solutions of the present invention, rather than to the limit of the present invention System, it should be appreciated by those skilled in the art that the realization method of any transfer case information for obtaining multiple keywords, should all include Within the scope of the invention.

In step s 12, computer equipment according to multiple keywords be respectively associated to object and the multiple key The transfer case information of word determines the transfer case information of multiple objects.

Specifically, computer equipment can be according to transfer path information of the user in multiple keywords and multiple keywords Be respectively associated to object, determine transfer path information of the user in multiple objects, also, according to user in each keyword Between transfer number information and multiple keywords be respectively associated to object, determine transfer of the user between each object Number and/or transition probability information.

For example, the transfer case information of multiple keywords is as shown in Table 2 above, and Query1, Query2 and Query3 points It is not associated with to Object1, Object2 and Object3；Computer equipment is according to the transfer case information of keyword shown in table 2 And aforementioned incidence relation, determine that transfer path of the user in object Object1, Object2 and Object3 includes " Object1 → Object2 " and " Object1 → Object3 ", and the transfer number of 2 transfer paths is respectively 5 and 8；It connects It, computer equipment calculates transition probability=5/ of " Object1 → Object2 " according to the transfer number of 2 transfer paths (5+8)=38.46%, transition probability=8/ (5+8)=61.54% of " Object1 → Object3 ", that is, computer equipment Obtain transfer case information as shown in Table 1.

It should be noted that being associated with to same since a user may pay close attention in continuous several times object information acquisition behavior The different keywords (plain keyword is such as searched using the difference of the corresponding same object in multiple search) of one object, therefore, There may be the object transition probability of itself is transferred to from an object, as p shown in Fig. 6 may be present₀₀Deng.The one of Fig. 6 A specific example can be found in Fig. 7.As shown in Figure 7, " Gymboree " probability of itself is transferred to from " Gymboree " may be up to 71.86%.

It should be noted that preferably, computer equipment can be based on following formula come computing object i to the transfer of object j Probability p_ij：

Wherein, a_ijIndicate the transfer number of object i to object j,Transfer numbers of the expression object i to all objects.

For example, as shown in fig. 6, object Object0 is transferred to its own and other multiple object Object1 extremely Object13；For being transferred to Object8 from object Object1, transition probabilities of the object Objectp0 to object Object8Wherein,Indicate from object Object0 to Object0 itself and object Object1 to All transfer numbers of Object13.

It should be noted that multiple keywords be respectively associated to object can be determined in advance, Fig. 5 is shown from keyword Reticular structure transfer path to object reticular structure transfer path transform instances.In Fig. 5, the reticular structure of top In each node be keyword, each node in the reticular structure of lower section is and the keyword pair in the respective nodes of top The object answered.

It should be noted that the above-mentioned examples are merely illustrative of the technical solutions of the present invention, rather than to the limit of the present invention System, it should be appreciated by those skilled in the art that the realization method of any transfer case information for obtaining multiple objects, should be included in In the scope of the present invention.

In step s 2, computer equipment is according to the transfer case information of the multiple objects obtained in step sl, to this Multiple objects are clustered, and the cluster result of the multiple object is obtained.

Wherein, the cluster result of the multiple object can behave as diversified forms；For example, the cluster result includes multiple collection It closes, each gathers the object for including and belong to a kind of；In another example the cluster result includes：Object ID and corresponding with the object ID Category IDs, then can determine the classification belonging to the object by the corresponding category IDs of each object ID.

Specifically, computer equipment clusters multiple object according to the transfer case information of multiple objects, obtains The realization method of the cluster result of the multiple object includes but not limited to：

1) computer equipment directly clusters multiple objects according to the transfer case information of multiple objects, obtains more The cluster result of a object.Wherein, the transition probability between two objects or transfer number are higher, then two objects, which are gathered, is A kind of possibility is higher.

For example, the transfer case information that computer equipment obtains in step sl is as shown in aforementioned table 1, then computer equipment Be less than predetermined threshold 60% according to the transition probability between Object1 and Object2 38.46%, determine Object1 and Object2 cannot gather for one kind, also, computer equipment is according to the transition probability 61.54% between Object1 and Object3 More than predetermined threshold 60%, determine that Object1 and Object3 gather for one kind.Then computer equipment acquisition shows as two set Cluster result [Object1, Object3], [Object2]；Wherein, which belongs to Same category, Object2 individually belong to a classification.

It should be noted that in the multiple object it is existing gathered for a kind of object in the case of it is (such as multiple right As it is middle may include to have been gathered for a kind of object by artificial or computer equipment operation), in fact it could happen that judgement, which has gathered, is Whether a kind of multiple objects and other one or more objects can gather for a kind of situation, then：One object is with having gathered Transition probability or transfer number are higher between one or more of a kind of multiple objects object, then an object with gathered It is that a kind of possibility is higher to be gathered for a kind of multiple objects；Gather for one or more of a kind of multiple objects object Transition probability between one or more of a kind of multiple objects object is gathered with other or transfer number is higher, then should Gathered for a kind of multiple objects gathered with other be a kind of multiple objects by gathered be one kind possibility it is higher.

2) computer equipment is by based on the transfer distance between the transfer case information acquisition object, to described more A object is clustered, and the cluster result of the multiple object is obtained.

Specifically, computer equipment can first obtain the transfer distance between all objects, further according to transfer distance to multiple Object is clustered, and the cluster result of the multiple object is obtained；Alternatively, computer equipment can perform multiple cluster operation to obtain Multiple objects cluster result, such as in each cluster operation from multiple objects selected section object, and determination the part Transfer distance needed between object, to carry out cluster operation to the partial objects.

Preferably, the transfer distance between the object includes but not limited at least one of following：

A) transfer distance between an object in the multiple object and another object in the multiple object. Wherein, the transfer distance is smaller, and in the multiple object a object is gathered with another object in the multiple object It is bigger for a kind of possibility.

Wherein, the transfer distance between two objects by transfer number information in transfer case information and/or can turn Probabilistic information is moved to determine.

For example, the transfer distance between two objects can be determined by following formula：

Wherein, d_ijIndicate the transfer distance between object i and object j, p_ijIndicate that the transfer between object i to object j is general Rate, p_jiIndicate that the transition probability between object j to object i, r expression parameters, the parameter can manually be set.

It should be noted that above-mentioned formula can adjust as needed, such as by the (p in formula_ij+p_ji)/2 are adjusted toDeng.

B) transfer distance between an object in the multiple object and multiple objects in the multiple object.Its In, the transfer distance is smaller, and it is one that in the multiple object a object is gathered with multiple objects in the multiple object The possibility of class is bigger.Preferably, multiple objects in the multiple object are usually gathered for one kind.

Wherein, the transfer between an object in the multiple object and multiple objects in the multiple object away from From, can according to the transfer distance between one or more of multiple objects in an object and multiple object object come It determines, it also can be according to the transfer time between one or more of multiple objects in an object to multiple object object Number/transition probability determines.

For example, co-existing in 9 object Object1 to Object9, wherein it is one that an object Object1 and three, which have gathered, Transfer distance between object Object4, Object7 and Object8 of class can be determined by following any mode：

The first：By between Object1 and Object4, between Object1 and Object7, Object1 and Object8 it Between transfer distance in minimum transfer distance as the transfer between Object1 and Object4, Object7 and Object8 away from From.

Second：By between Object1 and Object4, between Object1 and Object7, Object1 and Object8 it Between transfer distance in maximum transfer distance as the transfer between Object1 and Object4, Object7 and Object8 away from From.

The third：Between Object1 and Object4, between Object1 and Object7, Object1 and Object8 it Between three transfer distances calculated, such as averaged, and using result of calculation as Object1 and Object4, Transfer distance between Object7 and Object8.

4th kind：Determine between Object1 and Object4, between Object1 and Object7, Object1 and Object8 Between transfer number/transition probability in maximum transfer number/transition probability, and according to the maximum transfer number/transfer Probability seeks transfer distance, as the transfer distance between Object1 and Object4, Object7 and Object8.

5th kind：Determine between Object1 and Object4, between Object1 and Object7, Object1 and Object8 Between transfer number/transition probability in minimum transfer number/transition probability, and according to transfer number/transfer of the minimum Probability seeks transfer distance, as the transfer distance between Object1 and Object4, Object7 and Object8.

6th kind：Between Object1 and Object4, between Object1 and Object7, Object1 and Object8 it Between transfer number/transition probability calculated, such as averaged, and transfer distance is sought according to result of calculation, as Transfer distance between Object1 and Object4, Object7 and Object8.

C) transfer between other multiple objects in the multiple objects and the multiple object in the multiple object away from From.Wherein, the transfer distance is smaller, multiple objects in the multiple object and other multiple objects in the multiple object It is bigger for the possibility of one kind by being gathered.

Wherein, the transfer between other multiple objects in the multiple objects and the multiple object in the multiple object Distance, can be according between one or more of one or more of multiple object object and other multiple objects object Transfer distance determine, also can be according to one in one or more of multiple object object and other multiple objects Or between multiple objects between transfer number/transition probability determine.

For example, co-existing in 9 object Object1 to Object9, wherein gathered for a kind of two object Object1 and Object3 and two transfer distance gathered between a kind of object Object4 and Object8, can be by following any Mode determines：

The first：By between Object1 and Object4, between Object1 and Object8, Object3 and Object4 it Between, transfer distance minimum in the transfer distance between Object3 and Object8 as Object1 and Object3 with Transfer distance between Object4 and Object8.

Second：By between Object1 and Object4, between Object1 and Object8, Object3 and Object4 it Between, in the transfer distance between Object3 and Object8 maximum transfer distance as Object1 and Object3 with Transfer distance between Object4 and Object8.

The third：Between Object1 and Object4, between Object1 and Object8, Object3 and Object4 it Between, the shifting science and technology in four directions distance between Object3 and Object8 calculated, such as averaged, and using result of calculation as Transfer distance between Object1 and Object3 and Object4 and Object8.

4th kind：Determine between Object1 and Object4, between Object1 and Object8, Object3 and Object4 Between, maximum transfer number/transition probability in transfer number/transition probability between Object3 and Object8, and according to Maximum transfer number/the transition probability seeks transfer distance, as Object1 and Object3 and Object4 and Object8 Between transfer distance.

5th kind：Determine between Object1 and Object4, between Object1 and Object8, Object3 and Object4 Between, transfer number/transition probability minimum in transfer number/transition probability between Object3 and Object8, and according to Transfer number/transition probability of the minimum seeks transfer distance, as Object1 and Object3 and Object4 and Object8 Between transfer distance.

6th kind：Between Object1 and Object4, between Object1 and Object8, Object3 and Object4 it Between, transfer number/transition probability between Object3 and Object8 calculated, such as averaged, and according to calculating As a result transfer distance is sought, as the transfer distance between Object1 and Object3 and Object4 and Object8.

It should be noted that the above-mentioned examples are merely illustrative of the technical solutions of the present invention, rather than to the limit of the present invention System, it should be appreciated by those skilled in the art that the transfer distance between any object, should be included in the scope of the present invention.

Realization method 2 as step S2) one of preferred embodiment, step S2 further comprise step S21, step S22, Step S23, step S24 and step S25.

In the step s 21, computer equipment selects first part's object and second part object in the multiple object.

Wherein, first part's object, which can be one or more of the multiple object object, second part object, to be The one or more objects different from first part's object in the multiple object.Preferably, when first part's object or second Should include that first part's object of multiple objects or second part object belong to a kind of when partial objects are multiple.

In step S22, computer equipment, which obtains, to be based on having with first part's object and the second part object Transfer distance between the transfer case information of pass determines, first part's object and second part object.

It should be noted that before step S22, the transfer distance between first part's object and second part object can It can have existed；For example, transfer distance between first part's object and second part object may be previous in step Determined by computer equipment etc..

Preferably, when the transfer distance between first part's object and second part object is existing, computer Equipment directly reads the transfer distance between first part's object and second part object.As computer equipment is directly read Transfer distance between local already present first part's object and second part object.

In the absence of the transfer distance between first part's object and second part object, computer equipment according to Determined based on the transfer case information between first part's object and the second part object, first part's object One or more of transfer distance between one or more of object and the second part object object, determine described Transfer distance between a part of object and second part object.Wherein, how to determine between two objects, an object with it is more The mode of transfer distance between a object, between multiple objects and multiple objects is directed to the " transfer between object aforementioned Distance " is described in detail in illustrating, details are not described herein.If in addition, one or more of first part's object object with this Transfer distance between one or more of two partial objects object has existed before the execution of this step, then in this step In directly read, if not yet obtaining the transfer distance when this step executes, need be based on first part's object and institute The transfer case information between second part object is stated to determine the transfer distance.

In step S23, computer equipment according to the transfer distance between first part's object and second part object, Determine whether first part's object gathers with second part object for one kind.

Wherein, the transfer distance is smaller, and first part's object is gathered higher for the possibility of one kind with second part object； The transfer distance is bigger, and first part's object is gathered smaller for the possibility of one kind with second part object.

In step s 24, computer equipment reselects first part's object and second part object, wherein selects again It was not carried out cluster operation between the first part's object selected and second part object.

In step s 25, computer equipment repeats step S22, step S23 and step S24, until acquisition is the multiple right The cluster result of elephant.Preferably, various ways can be used in computer equipment, to determine whether having obtained the poly- of the multiple object Class result；For example, whether number of repetition has been more than predetermined repetition threshold value, if be not present and be not carried out the first of cluster operation Partial objects and second part object etc..

It gives an example below, this preferred embodiment is better described：

For example, co-existing in 5 objects Object1, Object2, Object3, Object4, Object5.

In the step s 21, computer equipment selects Object1 as first part's object, selects Object2 as second Partial objects.Then, in step S22, computer equipment is according to the transfer case information between Object1 and Object2, really Determine transfer distance between Object1 and Object2.Then, in step S23, computer equipment according to Object1 with Transfer distance between Object2 determines that Object1 and Object2 gathers for one kind；Then, in step s 24, computer equipment It selects to have gathered for a kind of Object1 and Object2, as first part's object, to select Object3 as second part object.

Then, computer equipment repeats step S22 to step S23, determines that Object1 and Object2 and Object3 cannot Gather for one kind, and repeat step S24, select to have gathered for a kind of Object1 and Object2, as first part's object, to select Object4 is as second part object.

Then, computer equipment repeats step S22 to step S23, determines that Object1 and Object2 and Object4 cannot Gather for one kind, and repeat step S24, select to have gathered for a kind of Object1 and Object2, as first part's object, to select Object5 is as second part object.

Then, computer equipment repeats step S22 to step S23, determines that Object1 and Object2 and Object5 cannot Gather for one kind, and repeat step S24, selects Object3 as first part's object, select Object4 as second part pair As.

Then, computer equipment repeats step S22 to step S23, determines that Object3 and Object4 gather for one kind, lays equal stress on Multiple step S24 select to have gathered Object3 and Object4 for one kind as first part's object, selects Object5 as the Two partial objects.

Then, computer equipment repeats step S22 to step S23, determines that Object3 and Object4 gather with Object5 and is One kind, and step S24 is repeated, it selects to have gathered for a kind of Object1 and Object2 as first part's object, selection has gathered It is a kind of Object3, Object4 and Object5 as second part object.

Then, computer equipment repeats step S22 to step S23, determine Object1 and Object2 and Object3, Object4, Object5 cannot gather for one kind.Also, computer equipment, which judges currently to be not present, did not carried out the first of cluster Partial objects and second part object stop cluster operation.Then object Object1, Object2, Object3, Object4, The cluster result of Object5 is：[Object1, Object2], [Object3, Object4, Object5].

In the prior art, natural language analysis is usually carried out by the description text to object, to classify to object. Particularly, when object is related to commercial use, such as when object is brand, the influence that artificial supervisor judges is received, in addition to object Title carries out other than natural language analysis, can also in conjunction with from the data of object angle, industry and region as belonging to object, The factors such as the sales situation and the market demand of object, to classify to object.That is, to design commercial use object into When row classification, those skilled in the art can have such prejudice：According to the commercial data from object angle, to right As classifying.

The solution of the present invention has broken above-mentioned prejudice, can by analyze transfer case information of the user in object come pair Object is clustered；Also, compared to the data from object angle, by analyzing user in multiple objects in the present invention Transfer case more can intuitively reflect understanding of the user to object come the scheme clustered to object closer to user perspective, Therefore, object classification determined by the solution of the present invention is more objective, accurate；Even if in addition, in the number from user perspective In, transfer case information of the invention is not common data yet, if in fact, clearly referring to the number from user perspective According to those skilled in the art are easier it is envisioned that direct evaluation from the user (such as give a mark, comment on word).

Fig. 2 is the structural schematic diagram of the clustering apparatus clustered to object of a preferred embodiment of the invention.This is poly- Class device can mount in computer equipment, which includes：Dress for the transfer case information for obtaining multiple objects It sets (hereinafter referred to as " acquisition device 1 "), for according to the transfer case information, being clustered to the multiple object, obtains The device (hereinafter referred to as " sub- clustering apparatus 2 ") of the cluster result of the multiple object.

Acquisition device 1 obtains the transfer case information of multiple objects.

1) transfer path information of the user in multiple objects.

2) transfer number information of the user between each object.

3) transition probability information of the user between each object.

Preferably, which can be used a variety of storage modes.

For example, the transfer case information storage is table, and transfer road of the user in multiple objects is had recorded in table The transfer number and transition probability of diameter and user between each object, as shown in aforementioned table 1.

Specifically, the mode that acquisition device 1 obtains the transfer case information of multiple objects includes but not limited to：

1) acquisition device 1 is standby directly acquires pre-determining, multiple object transfer case information.

For example, acquisition device 1 reads the transfer case letter of pre-determining, multiple object from local or other equipment Breath.

2) acquisition device 1 further comprises the device of the transfer case information for obtaining multiple keywords (hereinafter referred to as It is " the first sub- acquisition device ", not shown) and for according to multiple keywords be respectively associated to object and the multiple key The transfer case information of word, determine the transfer case information of multiple objects device (hereinafter referred to as " the first determining device ", figure not Show).

First sub- acquisition device obtains the transfer case information of multiple keywords.

A) transfer path information of the user in the multiple keyword.

B) transfer number information of the user between each keyword.

For example, the transfer case information storage is table, and transfer of the user in multiple keywords is had recorded in table The transfer number of path and user between each keyword, as shown in Table 2 above.

Specifically, the realization method that the first sub- acquisition device obtains the transfer case information of multiple keywords includes but unlimited In：

A) the first sub- acquisition device directly acquires the transfer case information of pre-determining, multiple keyword.

For example, the first sub- acquisition device reads turn from local or other equipment Condition of shifting one's love information.

B) the first sub- acquisition device obtains the keyword concern record of at least one user, and is paid close attention to according to the keyword Record, determines the transfer case information of the multiple keyword.

Record is paid close attention to preferably for the keyword of each user, the first sub- acquisition device is paid close attention to according to the keyword to be remembered The temporal information that the keyword for including in record is concerned determines transfer path and each keyword of the user in keyword Between transfer number；Also, the first sub- acquisition device is by merging transfer path in keyword of each user and each Transfer number between a keyword determines the transfer case information of the multiple keyword.

For example, the first sub- acquisition device obtains the keyword concern record of user A and user B；Wherein, user A and use The keyword concern record of family B is respectively as shown in aforementioned table 3 and table 4.

Then the keyword of user A is paid close attention to and is recorded, the first sub- acquisition device determines transfer roads of the user A in keyword Diameter includes " Query1 → Query3 ", and the transfer number of " Query1 → Query3 " is 1；Similarly, the first sub- acquisition device Determine that transfer paths of the user B in keyword includes " Query1 → Query2 ", and the transfer of " Query1 → Query2 " is secondary Number is 1.Then, between transfer paths and each keyword of first sub- acquisition device merging the user A and B in keyword Transfer number determines the transfer case information of the multiple keyword as shown in aforementioned table 5.

First determining device according to multiple keywords be respectively associated to object and the multiple keyword transfer feelings Condition information determines the transfer case information of multiple objects.

Specifically, the first determining device can be according to transfer path information of the user in multiple keywords and multiple keys Word be respectively associated to object, determine transfer path information of the user in multiple objects, also, the first determining device according to Transfer number information and multiple keyword of the family between each keyword be respectively associated to object, determine user each Transfer number between object and/or transition probability information.

For example, the transfer case information of multiple keywords is as shown in Table 2 above, and Query1, Query2 and Query3 points It is not associated with to Object1, Object2 and Object3；First determining device is believed according to the transfer case of keyword shown in table 2 Breath and aforementioned incidence relation, determine that transfer path of the user in object Object1, Object2 and Object3 includes " Object1 → Object2 " and " Object1 → Object3 ", and the transfer number of 2 transfer paths is respectively 5 and 8；It connects , the first determining device according to the transfer numbers of 2 transfer paths, calculate the transition probability of " Object1 → Object2 "= 5/ (5+8)=38.46%, transition probability=8/ (5+8)=61.54% of " Object1 → Object3 ", that is, first determines Device obtains the transfer case information as shown in aforementioned table 1.

It should be noted that preferably, the first determining device can be turned based on following formula come computing object i to object j Move Probability p_ij：

The transfer case information for multiple objects that sub- clustering apparatus 2 is obtained according to acquisition device 1 carries out multiple object Cluster, obtains the cluster result of the multiple object.

Specifically, sub- clustering apparatus 2 clusters multiple object, obtains according to the transfer case information of multiple objects The realization method of cluster result of the multiple object includes but not limited to：

1) sub- clustering apparatus 2 directly clusters multiple objects according to the transfer case information of multiple objects, obtains The cluster result of multiple objects.Wherein, the transition probability between two objects or transfer number are higher, then two objects are gathered It is higher for a kind of possibility.

For example, the transfer case information that acquisition device 1 obtains is as shown in aforementioned table 1, then 2 basis of sub- clustering apparatus Transition probability 38.46% between Object1 and Object2 is less than predetermined threshold 60%, determines Object1 and Object2 It cannot gather for one kind, also, sub- clustering apparatus 2 is more than pre- according to the transition probability 61.54% between Object1 and Object3 Determine threshold value 60%, determines that Object1 and Object3 gather for one kind.Then sub- clustering apparatus 2 obtains the cluster for showing as two set As a result [Object1, Object3], [Object2]；Wherein, which belongs to same class Not, Object2 individually belongs to a classification.

It should be noted that in the multiple object it is existing gathered for a kind of object in the case of it is (such as multiple right As it is middle may include to have been gathered for a kind of object by artificial or clustering apparatus operation), in fact it could happen that it is one to judge to have gathered Whether multiple objects of class and other one or more objects can gather for a kind of situation, then：One object and to have gathered be one Transition probability or transfer number are higher between one or more of multiple objects of class object, then an object is with having gathered A kind of multiple objects are gathered higher for the possibility of one kind；Gathered for one or more of a kind of multiple objects object with Other have gathered transition probability between one or more of a kind of multiple objects object or transfer number is higher, then this has been Gather for a kind of multiple objects gathered with other be a kind of multiple objects by gathered be one kind possibility it is higher.

2) sub- clustering apparatus 2 is by based on the transfer distance between the transfer case information acquisition object, to described Multiple objects are clustered, and the cluster result of the multiple object is obtained.

Specifically, sub- clustering apparatus 2 can first obtain the transfer distance between all objects, further according to transfer distance to multiple Object is clustered, and the cluster result of the multiple object is obtained；Alternatively, sub- clustering apparatus 2 can perform multiple cluster operation Obtain the cluster result of multiple objects, such as in each cluster operation from multiple objects selected section object, and determine the portion Transfer distance needed between point object, to carry out cluster operation to the partial objects.

Realization method 2 as sub- clustering apparatus 2) one of preferred embodiment, sub- clustering apparatus 2 further comprise for The device of first part's object and second part object is selected in multiple objects, and (hereinafter referred to as " first choice device ", figure is not Show), for obtain it is being determined based on transfer case information related with first part's object and the second part object, Device (hereinafter referred to as " the second sub- acquisition device ", the figure of transfer distance between first part's object and second part object Do not show), for according to the transfer distance between first part's object and second part object, determine first part's object with Whether second part object gather for a kind of device (hereinafter referred to as " the second determining device ", not shown), for reselecting the The device (hereinafter referred to as " the second selection device ", not shown) of a part of object and second part object and for triggering described Two sub- acquisition device, second determining device, second selection device repeat operation, until acquisition is the multiple right The device (hereinafter referred to as " trigger device ", not shown) of the cluster result of elephant.

First choice device selects first part's object and second part object in multiple objects.

Second sub- acquisition device, which obtains, is based on transfer related with first part's object and the second part object Transfer distance between situation information determines, first part's object and second part object.

It should be noted that before the second sub- acquisition device executes operation, first part's object and second part object Between transfer distance may have existed；For example, the transfer distance between first part's object and second part object may Device determination etc. is clustered in operating previous.

Preferably, when the transfer distance between first part's object and second part object is existing, the second son Acquisition device directly reads the transfer distance between first part's object and second part object.Such as the second sub- acquisition device Directly read the transfer distance between local already present first part's object and second part object.

In the absence of the transfer distance between first part's object and second part object, the second sub- acquisition device According to determined based on the transfer case information between the first part's object and the second part object, first part Transfer distance between one or more of one or more of object object and the second part object object, determines institute State the transfer distance between first part's object and second part object.Wherein, how to determine between two objects, an object The mode of transfer distance between multiple objects, between multiple objects and multiple objects, aforementioned for " between object Transfer distance " is described in detail in illustrating, details are not described herein.If in addition, one or more of first part's object object with Transfer distance between one or more of second part object object before the second sub- acquisition device executes operation Through existing, then the second sub- acquisition device is directly read, if not yet obtaining this turn when the second sub- acquisition device executes operation Distance is moved, then needs to determine the transfer based on the transfer case information between first part's object and the second part object Distance.

Second determining device according to the transfer distance between first part's object and second part object, determine this first Whether partial objects gather with second part object for one kind.

Second selection device reselects first part's object and second part object, wherein first reselected Divide and was not carried out cluster operation between object and second part object.

Trigger device triggers the second sub- acquisition device, second determining device, second selection device and repeats Operation is executed, until obtaining the cluster result of the multiple object.Preferably, various ways can be used in trigger device, to judge Whether the cluster result of the multiple object has been obtained；For example, whether number of repetition has been more than predetermined repetition threshold value, if not In the presence of the first part's object and second part object etc. for being not carried out cluster operation.

It gives an example below, this preferred embodiment is better described：

First choice device selects Object1 as first part's object, selects Object2 as second part object. Then, the second sub- acquisition device is according to the transfer case information between Object1 and Object2, determine Object1 with Transfer distance between Object2.Then, the second determining device is determined according to transfer distance between Object1 and Object2 Object1 and Object2 gathers for one kind；Then, the second selection device selects to have gathered and make for a kind of Object1 and Object2 For first part's object, select Object3 as second part object.

Then, trigger device triggers the second sub- acquisition device and the second determining device repeats operation, to determine Object1 and Object2 and Object3 cannot gather for one kind, and trigger device triggers the second selection device and repeats operation, It selects to have gathered as a kind of Object1 and Object2 as first part's object, selects Object4 as second part pair As.

Then, trigger device triggers the second sub- acquisition device and the second determining device repeats operation, to determine Object1 and Object2 and Object4 cannot gather for one kind, and trigger device triggers the second selection device and repeats operation, It selects to have gathered as a kind of Object1 and Object2 as first part's object, selects Object5 as second part pair As.

Then, trigger device triggers the second sub- acquisition device and the second determining device repeats operation, to determine Object1 and Object2 and Object5 cannot gather for one kind, and trigger device triggers the second selection device and repeats operation, It selects the Object3 as first part's object, selects Object4 as second part object.

Then, trigger device triggers the second sub- acquisition device and the second determining device repeats operation, to determine Object3 and Object4 gathers for one kind, and trigger device triggers the second selection device and repeats operation, come select to have gathered for A kind of Object3 and Object4 selects Object5 as second part object as first part's object.

Then, trigger device triggers the second sub- acquisition device and the second determining device repeats operation, to determine Object3 and Object4 and Object5 gather for one kind, and trigger device triggers the second selection device and repeats operation, to select Select and gathered for a kind of Object1 and Object2 as first part's object, select to have gathered the Object3 for one kind, Object4 and Object5 is as second part object.

Then, trigger device triggers the second sub- acquisition device and the second determining device repeats operation, to determine Object1 and Object2 and Object3, Object4, Object5 cannot gather for one kind.Also, trigger device judges currently There is no the first part's objects and second part object that did not carried out cluster, stop cluster operation.Then object Object1, The cluster result of Object2, Object3, Object4, Object5 is：[Object1, Object2], [Object3, Object4, Object5].

The solution of the present invention has broken above-mentioned prejudice, can by analyze transfer case information of the user in object come pair Object is clustered；Also, compared to the data from object angle, by analyzing user in multiple objects in the present invention Transfer case carrys out the scheme carried out to object, closer to user perspective, more can intuitively reflect understanding of the user to object, because This, object classification determined by the solution of the present invention is more objective, accurate；Even if in addition, in the data from user perspective In, transfer case information of the invention is not common data yet, if in fact, clearly refer to the data from user perspective, Those skilled in the art are easier it is envisioned that direct evaluation from the user (such as give a mark, comment on word).

It should be noted that the present invention can be carried out in the assembly of software and/or software and hardware, for example, this hair Application-specific integrated circuit (ASIC) can be used in bright each device or any other is realized similar to hardware device.In one embodiment In, software program of the invention can be executed by processor to realize steps described above or function.Similarly, of the invention Software program (including relevant data structure) can be stored in computer readable recording medium storing program for performing, for example, RAM memory, Magnetic or optical driver or floppy disc and similar devices.In addition, hardware can be used to realize in some steps or function of the present invention, example Such as, coordinate to execute the circuit of each step or function as with processor.

It is obvious to a person skilled in the art that invention is not limited to the details of the above exemplary embodiments, Er Qie In the case of without departing substantially from spirit or essential attributes of the invention, the present invention can be realized in other specific forms.Therefore, no matter From the point of view of which point, the present embodiments are to be considered as illustrative and not restrictive, and the scope of the present invention is by appended power Profit requires rather than above description limits, it is intended that all by what is fallen within the meaning and scope of the equivalent requirements of the claims Variation includes within the present invention.Any reference signs in the claims should not be construed as limiting the involved claims.This Outside, it is clear that one word of " comprising " is not excluded for other units or step, and odd number is not excluded for plural number.That is stated in system claims is multiple Unit or device can also be realized by a unit or device by software or hardware.The first, the second equal words are used for table Show title, and does not represent any particular order.

Claims

1. a kind of method for being clustered to object in computer equipment, wherein this method includes：

The transfer case information of multiple objects is obtained, the transfer case information is used to indicate obtains behavior based on object information , the transfer case that user is in the multiple object；

According to the transfer case information, the multiple object is clustered, obtains the cluster result of the multiple object；

Wherein, the step of obtaining the transfer case information include：

Obtain the transfer case information of multiple keywords, wherein the transfer case information of the multiple keyword is used to indicate base Behavior, transfer case that user is in multiple keywords are obtained in object information；

According to the multiple keyword be respectively associated to multiple objects and the multiple keyword transfer case information, really Determine transfer case information of the user in the multiple object；

Wherein, the step of transfer case information for obtaining the multiple keyword includes：

The keyword concern record of at least one user is obtained, keyword concern record includes that the multiple user believes in object The temporal information that the keyword and the keyword paid close attention in breath acquisition behavior are concerned；

It is paid close attention to and is recorded according to the keyword, determine the transfer case information of the multiple keyword.

2. according to the method described in claim 1, wherein, described the step of being clustered, includes：

By based on the transfer distance between the transfer case information acquisition object, to be clustered to the multiple object, Obtain the cluster result of the multiple object.

3. according to the method described in claim 2, wherein, described the step of being clustered, includes：

First part's object and second part object are selected in the multiple object；

Obtain it is being determined based on transfer case information related with first part's object and the second part object, this Transfer distance between a part of object and second part object；

According to the transfer distance between first part's object and second part object, first part's object and second are determined Divide whether object gathers for one kind；

Reselect first part's object and second part object, wherein the first part's object and second part reselected Cluster operation was not carried out between object；

Repeat the transfer distance obtained between first part's object and second part object, the determination first part pair As whether gathering for a kind of, described the step of reselecting first part's object and second part object with second part object, directly To the cluster result for obtaining the multiple object.

4. according to the method described in claim 3, wherein, obtaining and turning between first part's object and second part object Move apart from the step of include：

When the transfer distance between first part's object and second part object is existing, described first is directly read Divide the transfer distance between object and second part object；

In the absence of the transfer distance between first part's object and second part object, according to based on described first One or more of the transfer case information divided between object and the second part object determines, first part's object Transfer distance between one or more of object and the second part object object determines first part's object and Transfer distance between two partial objects.

5. method according to any one of claim 2 to 4, wherein the transfer distance between the object includes following At least one of：

Transfer distance between an object in the multiple object and another object in the multiple object；

Transfer distance between an object in the multiple object and multiple objects in the multiple object；

The transfer distance between other multiple objects in multiple objects and the multiple object in the multiple object.

6. according to the method described in claim 1, wherein, the transfer case information of the multiple keyword includes following at least one ：

Transfer path information of the user in the multiple keyword；

Transfer number information of the user between each keyword.

7. method according to claim 1 to 4, wherein the transfer case information of the multiple object includes At least one of below：

Transfer path information of the user in the multiple object；

Transfer number information of the user between each object；

Transition probability information of the user between each object.

8. method according to claim 1 to 4, wherein the object includes brand.

9. a kind of device for being clustered to object in computer equipment, wherein the device includes：

Device for the transfer case information for obtaining multiple objects, the transfer case information are used to indicate based on object information Acquisition behavior, transfer case that user is in the multiple object；

For according to the transfer case information, being clustered to the multiple object, the cluster knot of the multiple object is obtained The device of fruit；

Wherein, the device for obtaining the transfer case information includes：

Device for the transfer case information for obtaining multiple keywords, wherein the transfer case information of the multiple keyword It is used to indicate transfer cases based on object information acquisition behavior, that user is in multiple keywords；

For according to the multiple keyword be respectively associated to multiple objects and the multiple keyword transfer case believe Breath, determines the device of transfer case information of the user in the multiple object；

Wherein, the device of the transfer case information for obtaining the multiple keyword includes：

Keyword for obtaining at least one user pays close attention to the device of record, and keyword concern record includes the multiple use The temporal information that the keyword and the keyword that family was paid close attention in object information acquisition behavior are concerned；

It is recorded for being paid close attention to according to the keyword, determines the device of the transfer case information of the multiple keyword.

10. device according to claim 9, wherein include for the device clustered：

For by based on the transfer distance between the transfer case information acquisition object, to gather to the multiple object Class obtains the device of the cluster result of the multiple object.

11. device according to claim 10, wherein include for the device clustered：

Device for selecting first part's object and second part object in the multiple object；

For obtain it is being determined based on transfer case information related with first part's object and the second part object, The device of transfer distance between first part's object and second part object；

For according to the transfer distance between first part's object and second part object, determining first part's object and Whether two partial objects gather for a kind of device；

Device for reselecting first part's object and second part object, wherein the first part's object reselected It was not carried out cluster operation between second part object；

For triggering device for obtaining the transfer distance between first part's object and second part object, for determine should Whether first part's object gathers with second part object for a kind of device, for reselecting first part's object and second The device of object is divided to repeat operation, until obtaining the device of the cluster result of the multiple object.

12. according to the devices described in claim 11, wherein for obtain first part's object and second part object it Between the device of transfer distance include：

For when the transfer distance between first part's object and second part object is existing, directly reading described The device of transfer distance between a part of object and second part object；

For in the absence of the transfer distance between first part's object and second part object, according to based on described One that transfer case information between a part of object and the second part object determines, in first part's object or Transfer distance between one or more of multiple objects and the second part object object determines first part's object The device of transfer distance between second part object.

13. device according to any one of claims 10 to 12, wherein the transfer distance between the object include with It is at least one of lower：

14. device according to claim 9, wherein the transfer case information of the multiple keyword include it is following at least One：

Transfer path information of the user in the multiple keyword；

Transfer number information of the user between each keyword.

15. the device according to any one of claim 9 to 12, wherein the transfer case packet of the multiple object It includes at least one of following：

Transfer path information of the user in the multiple object；

Transfer number information of the user between each object；

Transition probability information of the user between each object.

16. the device according to any one of claim 9 to 12, wherein the object includes brand.