CN104731809A - Processing method and device of attribute information of objects - Google Patents

Abstract

The invention provides a processing method and device of attribute information of objects. The method provided by the embodiment includes: acquiring the attribute information of each of the objects and operational information of each object so that an attribute relational network is established according to the attribute information and the operational information, with the attribute relational network being composed of at least two nodes each corresponding to one piece of attribute information; if the attribute information of two nodes belongs to the same object, determining that the two nodes are related, selecting at least one piece of attribute information as an attribute combination according to the attribute relational network. The problem that the single-attribute statistical results of the prior art allow no precise positioning of specific objects can be solved; attribute analysis reliability is improved.

Description

The disposal route of the attribute information of object and device

[technical field]

The application relates to the information processing technology, particularly relates to a kind of disposal route and device of attribute information of object.

[background technology]

Along with the development of the network information technology and universal, internet penetrates into the life of people, the every field of study and work, really brings the mankind into the information age.But the quantity of information on internet is very large, the object that user is not easy to find oneself to need and products & services, and be not easy object that issue needs other users and products & services.In order to help user, needing fully to excavate historical data information, analyzing from the attribute dimensions of object, help user's object search or recommended targetedly.In prior art, object-based single attribute, is mapped to object operation information on its attribute, thus obtains the statistics of its attribute according to the operation information statistics comprised in historical data information.

But the statistics of the single attribute obtained in prior art cannot accurately navigate to concrete object, thus result in the reduction of the reliability of attributive analysis.

[summary of the invention]

The many aspects of the application provide a kind of disposal route and device of attribute information of object, in order to improve the reliability of attributive analysis.

The one side of the application, provides a kind of disposal route of attribute information of object, comprising:

Obtain the attribute information of each object at least one object and the operation information of described each object;

According to the attribute information of described each object and the operation information of described each object, set up relation on attributes network; Wherein,

Described relation on attributes network is made up of at least two nodes, the corresponding attribute information of each node; If the attribute information that two nodes are corresponding belongs to same object, between described two nodes, there is incidence relation; The operation information of the weight of the described incidence relation each object belonging to attribute information corresponding to described two nodes, determines;

According to described relation on attributes network, select at least one attribute information, using as combinations of attributes.

Aspect as above and arbitrary possible implementation, provide a kind of implementation further, the operation information of the described attribute information according to described each object and described each object, sets up relation on attributes network, comprising:

According to the attribute information of described each object, determine the incidence relation between corresponding each node and described each node;

According to the operation information of described each object, determine the weight of described incidence relation;

According to the weight of the incidence relation between described each node, described each node and described incidence relation, set up described relation on attributes network.

Aspect as above and arbitrary possible implementation, provide a kind of implementation further, the operation information of the described attribute information according to described each object and described each object, sets up relation on attributes network, also comprise:

If attribute information belong to pre-set can delete property information, delete the node that described attribute information is corresponding;

If the weight of incidence relation is less than the weight threshold pre-set, delete described incidence relation;

If the quantity sum with node with other nodes of incidence relation is less than the amount threshold pre-set, delete described node and the incidence relation between described node and other nodes.

Aspect as above and arbitrary possible implementation, provide a kind of implementation further, describedly selects at least one attribute information according to described relation on attributes network, using as combinations of attributes, comprising:

Utilize corporations' detection algorithm, the node in described relation on attributes network is divided into groups, to obtain at least two groupings; Wherein, the compactedness between the node in described each grouping be greater than node in described each grouping and other divide into groups in node between compactedness;

Utilize key extraction algorithm, at least one incidence relation at least two groupings described in deletion;

If there is not incidence relation between node and any node, from the grouping belonging to described node, delete described node.

Aspect as above and arbitrary possible implementation, provide a kind of implementation further, describedly utilizes corporations' detection algorithm, divided into groups by the node in described relation on attributes network, to obtain at least two groupings, comprising:

By each node division to a grouping, utilize obtain packet parameters; Wherein,

Q is packet parameters;

W _i,jfor the weight of the incidence relation between node i and node j, node j is other nodes in described relation on attributes network except node i;

$m=\frac{1}{2}\underset{i,j}{\mathrm{\Σ}}{w}_{i,j};$

$Q=\frac{1}{2m}\underset{i,j}{\mathrm{\Σ}}(w_{i,j}-\frac{s_i{s}_j}{2m})\mathrm{\δ}(c_i,c_j);$

S _ifor the intensity of node i,

C _igrouping belonging to node i, c _jfor the grouping belonging to node j, if c _i=c _j, then δ (c _i, c _j)=1; If c _i≠ c _j, then δ (c _i, c _j)=0;

To each node i, respectively node i and node j are carried out union operation, to be divided into same grouping, obtain the variation delta Q of the packet parameters after each merging;

Variation delta Q is selected to be greater than 0 and the maximum group result of value, using grouping each in described group result as a new node; Wherein, this new node and the weight of the incidence relation of other nodes are the weight sum of the incidence relation of all nodes and these other nodes in corresponding grouping;

Repeatedly perform described union operation, until the variation delta Q of packet parameters is less than or equal to 0, record group result now, as described at least two groupings.

Aspect as above and arbitrary possible implementation, provide a kind of implementation further, and described operation information comprises at least one item in click information, Information on Collection and purchase information.

The another aspect of the application, provides a kind for the treatment of apparatus of attribute information of object, comprising:

Acquiring unit, for the operation information of the attribute information and described each object that obtain each object at least one object;

Set up unit, for according to the attribute information of described each object and the operation information of described each object, set up relation on attributes network; Wherein,

Described relation on attributes network is made up of at least two nodes, the corresponding attribute information of each node; If the attribute information that two nodes are corresponding belongs to same object, between described two nodes, there is incidence relation; The operation information of the weight of the described incidence relation each object belonging to attribute information corresponding to described two nodes, determines;

Selection unit, for according to described relation on attributes network, selects at least one attribute information, using as combinations of attributes.

Aspect as above and arbitrary possible implementation, provide a kind of implementation further, describedly sets up unit, specifically for

According to the attribute information of described each object, determine the incidence relation between corresponding each node and described each node;

According to the operation information of described each object, determine the weight of described incidence relation; And

According to the weight of the incidence relation between described each node, described each node and described incidence relation, set up described relation on attributes network.

Aspect as above and arbitrary possible implementation, provide a kind of implementation further, describedly sets up unit, also for

If attribute information belong to pre-set can delete property information, delete the node that described attribute information is corresponding;

If the weight of incidence relation is less than the weight threshold pre-set, delete described incidence relation;

If the quantity sum with node with other nodes of incidence relation is less than the amount threshold pre-set, delete described node and the incidence relation between described node and other nodes.

Aspect as above and arbitrary possible implementation, provide a kind of implementation, described selection unit further, specifically for

Utilize corporations' detection algorithm, the node in described relation on attributes network is divided into groups, to obtain at least two groupings; Wherein, the compactedness between the node in described each grouping be greater than node in described each grouping and other divide into groups in node between compactedness;

Utilize key extraction algorithm, at least one incidence relation at least two groupings described in deletion;

If there is not incidence relation between node and any node, from the grouping belonging to described node, delete described node.

Aspect as above and arbitrary possible implementation, provide a kind of implementation, described selection unit further, specifically for

By each node division to a grouping, utilize obtain packet parameters; Wherein,

Q is packet parameters;

W _i,jfor the weight of the incidence relation between node i and node j, node j is other nodes in described relation on attributes network except node i;

$m=\frac{1}{2}\underset{i,j}{\mathrm{\Σ}}{w}_{i,j};$

$Q=\frac{1}{2m}\underset{i,j}{\mathrm{\Σ}}(w_{i,j}-\frac{s_i{s}_j}{2m})\mathrm{\δ}(c_i,c_j);$

S _ifor the intensity of node i,

C _igrouping belonging to node i, c _jfor the grouping belonging to node j, if c _i=c _j, then δ (c _i, c _j)=1; If c _i≠ c _j, then δ (c _i, c _j)=0;

To each node i, respectively node i and node j are carried out union operation, to be divided into same grouping, obtain the variation delta Q of the packet parameters after each merging;

Variation delta Q is selected to be greater than 0 and the maximum group result of value, using grouping each in described group result as a new node; Wherein, this new node and the weight of the incidence relation of other nodes are the weight sum of the incidence relation of all nodes and these other nodes in corresponding grouping;

Repeatedly perform described union operation, until the variation delta Q of packet parameters is less than or equal to 0, record group result now, as described at least two groupings.

Aspect as above and arbitrary possible implementation, provide a kind of implementation further, and the described operation information that described acquiring unit obtains comprises at least one item in click information, Information on Collection and purchase information.

As shown from the above technical solution, the embodiment of the present application is by obtaining the attribute information of each object and the operation information of described each object at least one object, and then according to the attribute information of described each object and the operation information of described each object, set up relation on attributes network, described relation on attributes network is made up of at least two nodes, the corresponding attribute information of each node; If the attribute information that two nodes are corresponding belongs to same object, between described two nodes, there is incidence relation, make it possible to according to described relation on attributes network, select at least one attribute information, using as combinations of attributes, the statistics of single attribute in prior art can be avoided accurately cannot to navigate to the problem of concrete object, thus improve the reliability of attributive analysis.

In addition, adopt the technical scheme that the application provides, due to the attribute information of the object of the current all categories in website can be got, therefore, based on the object of all categories of a website attribute information selected by combinations of attributes, effectively can improve the coverage rate of attributive analysis.

[accompanying drawing explanation]

In order to be illustrated more clearly in the technical scheme in the embodiment of the present application, be briefly described to the accompanying drawing used required in embodiment or description of the prior art below, apparently, accompanying drawing in the following describes is some embodiments of the application, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.

The schematic flow sheet of the disposal route of the attribute information of the object that Fig. 1 provides for the application one embodiment;

Fig. 2 is the topological diagram of the relation on attributes network in the embodiment that Fig. 1 is corresponding;

The structural representation of the treating apparatus of the attribute information of the object that Fig. 3 provides for another embodiment of the application.

[embodiment]

For making the object of the embodiment of the present application, technical scheme and advantage clearly, below in conjunction with the accompanying drawing in the embodiment of the present application, technical scheme in the embodiment of the present application is clearly and completely described, obviously, described embodiment is some embodiments of the present application, instead of whole embodiments.Based on the embodiment in the application, those of ordinary skill in the art are not making other embodiments whole obtained under creative work prerequisite, all belong to the scope of the application's protection.

In addition, term "and/or" herein, being only a kind of incidence relation describing associated images, can there are three kinds of relations in expression, and such as, A and/or B, can represent: individualism A, exists A and B simultaneously, these three kinds of situations of individualism B.In addition, character "/" herein, general expression forward-backward correlation is to the relation similarly being a kind of "or".

The schematic flow sheet of the disposal route of the attribute information of the object that Fig. 1 provides for the application one embodiment, as shown in Figure 1.

101, the attribute information of each object at least one object and the operation information of described each object is obtained.

Alternatively, in one of the present embodiment possible implementation, in 101, the described operation information obtained can include but not limited at least one item in click information, Information on Collection and purchase information, and the embodiment of the present application is not particularly limited this.

Such as, specifically can extract in the fixed time list object having operation behavior from database, this list object can comprise following field:

The identification information of object, such as, the commodity sign product_ID of commodity;

The operation information of object, such as, the sales volume product_num of commodity;

Classification belonging to object, such as, the classification logotype category_ID of commodity;

The attribute information of object, such as, the attribute information attribute1 of commodity, attribute2 ..., attributeN, wherein, N is natural number.

102, according to the attribute information of described each object and the operation information of described each object, relation on attributes network is set up.

Wherein, described relation on attributes network is made up of at least two nodes, the corresponding attribute information of each node; If the attribute information that two nodes are corresponding belongs to same object, between described two nodes, there is incidence relation; The operation information of the weight of the described incidence relation each object belonging to attribute information corresponding to described two nodes is determined.

Alternatively, in one of the present embodiment possible implementation, in 102, specifically according to the attribute information of described each object, the incidence relation between corresponding each node and described each node can be determined.Like this, owing to considering the incidence relation between attribute information, thus more effective information can be provided for attributive analysis.

Then, according to the operation information of described each object, the weight of described incidence relation is determined.Particularly, if attribute information corresponding to two nodes does not belong to same object, illustrate not have incidence relation between described two nodes, so, the weight of corresponding described incidence relation then can be set to 0; If the attribute information that two nodes are corresponding belongs to same object, illustrate to have incidence relation between described two nodes, so, the weight of corresponding described incidence relation then can be set to the numerical value being greater than 0.

As for the size of the weight value of related information, specifically can the operation information of the object belonging to two nodes corresponding to this incidence relation determine.If two nodes corresponding to incidence relation only belong to an object, so, the weight of this incidence relation can only be determined according to the operation information of this object.Such as, the weight of incidence relation is set to the purchase volume in the fixed time of object.If two nodes corresponding to incidence relation belong to two objects or two or more object, so, the weight of this incidence relation can be determined according to the operation information of whole object.Such as, the weight of incidence relation is set to the purchase volume sum in the fixed time of whole object.

Finally, then according to the weight of the incidence relation between described each node, described each node and described incidence relation, described relation on attributes network can be set up.

Like this, this relation on attributes network that is undirected, Weight (weight) has just established.The tables of data storing this relation on attributes network can be as follows: (attributeM, attributeN, weight), and wherein, N and M is mutually different natural number.

Be understandable that, in the topological diagram of relation on attributes network, between two nodes with incidence relation, can couple together with line, as a limit of relation on attributes network, between two nodes without incidence relation, without any line, as shown in Figure 2.In fig. 2, between node A and Node B, there is incidence relation, between node A and node C, not there is incidence relation, between Node B and node C, there is incidence relation.Node A and Node B can be called mutually the adjacent node of the other side, and Node B and node C can be called mutually the adjacent node of the other side.

Alternatively, in one of the present embodiment possible implementation, after 102, or in the process of execution 102, execution can also adjust operation as follows further, more there is statistical significance to make set up relation on attributes network.

Such as, knot modification operates, even attribute information belong to pre-set can delete property information, delete the node that described attribute information is corresponding.As, the size of model, clothes.

Or, more such as, the incidence relation operation between knot modification, even the weight of incidence relation is less than the weight threshold pre-set, and deletes described incidence relation.

Or, again such as, incidence relation operation between knot modification operation and knot modification, the quantity sum even with node with other nodes of incidence relation is less than the amount threshold pre-set, and deletes described node and the incidence relation between described node and other nodes.Wherein, there is with node the quantity of other nodes of incidence relation, the angle value of node can also be called.

Be understandable that, if the incidence relation between a node and other nodes deleted after, this node and any node do not have incidence relation, so, then can by this knot removal.That is, can not there is isolated node in relation on attributes network, any node namely in relation on attributes network at least has incidence relation with other nodes arbitrary in this relation on attributes network.

103, according to described relation on attributes network, at least one attribute information is selected, using as combinations of attributes.

Alternatively, in one of the present embodiment possible implementation, in 103, specifically can utilize corporations' detection algorithm, the node in described relation on attributes network is divided into groups, to obtain at least two groupings; Wherein, the compactedness between the node in described each grouping be greater than node in described each grouping and other divide into groups in node between compactedness.

Particularly, based on the relation on attributes network that execution 102 is set up, in an initial condition, by each node division to a grouping, utilize obtain packet parameters, be denoted as Q.

Wherein,

Q is packet parameters;

W _i,jfor the weight of the incidence relation between node i and node j, node j is other nodes in described relation on attributes network except node i;

$m=\frac{1}{2}\underset{i,j}{\mathrm{\Σ}}{w}_{i,j};$

$Q=\frac{1}{2m}\underset{i,j}{\mathrm{\Σ}}(w_{i,j}-\frac{s_i{s}_j}{2m})\mathrm{\δ}(c_i,c_j);$

S _ifor the intensity of node i,

C _igrouping belonging to node i, c _jfor the grouping belonging to node j, if c _i=c _j, then δ (c _i, c _j)=1; If c _i≠ c _j, then δ (c _i, c _j)=0;

To each node i, respectively node i and node j are carried out union operation, to be divided into same grouping, obtain the variation delta Q of the packet parameters after each merging.

Variation delta Q is selected to be greater than 0 and the maximum group result of value, using grouping each in described group result as a new node; Wherein, this new node and the weight of the incidence relation of other nodes are the weight sum of the incidence relation of all nodes and these other nodes in corresponding grouping.

Repeatedly perform described union operation, until the variation delta Q of packet parameters is less than or equal to 0, record group result now, as described at least two groupings.

For commodity, namely this division community structure out divides into groups to can be good at the incidence relation between characterization attributes information, be under the jurisdiction of the attribute information of a grouping more easily simultaneously in order to describe class I goods, and the weight of incidence relation between attribute information, also the related content that the operation information having quantized this kind of commodity reflects, such as, to sell fast degree.

The Crack cause of this corporations phenomenon has two classes:

The first kind is that these attribute informations portray the style describing a kind of commodity jointly, such as, this commodity class of dress now, Style:Sexy & Club, Silhouette:Sheath, Dresses Length:Above Knee & Mini tri-attribute informations are divided in a grouping, because these words all describe the specific features of " sexy skirt " this category, and also demonstrate the fast sale of this kind of commodity in sales volume data.

Equations of The Second Kind is the common element that these attribute informations reflect commodity of selling fast jointly, such as, the combination of these three attributes of Waistline:Nature, Sleeve Length:Short, Sleeve Style:Puff Sleeve, any combination of two that sales volume data show between these three attributes all achieves good sales volume, the commodity namely having these combinations of attributes more easily obtain the favor of buyer, and this is for instructing seller to replenish, guiding commodity to move towards all have practical significance.

It should be noted that, in the community structure mining process of attribute information, all properties information in same grouping might not be connected completely between two, this declared attribute combined information can break through existing commodity limitation, obtain the attribute with potential incidence relation, provide information of forecasting to the commodity that future may sell fast.

Then, utilize key extraction algorithm, at least one incidence relation at least two groupings described in deletion.

For playing exercisable directive function, need to screen the group result of final attribute information, remove the content easily causing misleading as far as possible, export the combinations of attributes wherein determined most, and limit the quantity of the attribute information in each combinations of attributes, make result have ubiquity and operability.

For achieving the above object, key extraction algorithm is adopted finally to process group result.Key extraction algorithm is that a kind of network proposed based on this network characteristic with community structure simplifies algorithm.It can according to the statistical property of nodes and limit (incidence relation namely between node) and interaction relationship, analyze and obtain the annexation with important statistical significance, thus the secondary annexation in removal network, extract the diaphyseal portion in network, both keep connectivity and the architectural characteristic of legacy network, highlight again the important annexation in network.For each grouping, adopt key extraction algorithm, specifically can perform following operation:

Operation one, a setting number of nodes threshold value N ^*, after ensureing adopting key extraction algorithm, the number of nodes in each grouping is less than or equal to number of nodes threshold value N ^*.Number of nodes in the grouping that operation five exports is less than or equal to number of nodes threshold value N ^*time, the screening of this grouping is terminated, enters the screening step of next grouping.

Operation two, a setting heterogeneous coefficient threshold set Ψ, comprise multiple heterogeneous coefficient threshold α in this heterogeneous coefficient threshold set Ψ ^*.In general, heterogeneous coefficient threshold α ^*for less real number, such as, 10 ^-3, 10 ^-6, 10 ^-10deng.For each heterogeneous coefficient threshold α ^*, can executable operations three, operation four and operation five.For ensureing that the number of nodes in grouping can reach full number of nodes foot threshold value, the heterogeneous coefficient threshold α comprised in heterogeneous coefficient threshold set Ψ ^*can be set as by 10 ^-3to 10 ^-10successively decrease.

Operation three, utilization obtain the significance level of each node in whole incidence relations of this node, be denoted as p _i,j;

Wherein,

P _i,jfor the significance level of node i in whole incidence relations of this node i;

W _i,jfor the weight of the incidence relation between node i and node j, node j is other nodes in described relation on attributes network except node i;

S _ifor the intensity of node i,

Operate four, according to the significance level of each node in whole incidence relations of this node, utilize obtain the heterogeneous coefficient of incidence relation, be denoted as α _i,j.

Wherein,

α _i,jfor the heterogeneous coefficient of the incidence relation between node i and node j;

K is the angle value of node i, namely has the quantity of other nodes of incidence relation with node i.

Operation five, according to the heterogeneous coefficient of incidence relation and heterogeneous coefficient threshold, reservation process or delete processing are carried out to described incidence relation.

Particularly, if the heterogeneous coefficient of incidence relation is less than or equal to heterogeneous coefficient threshold, illustrates that this incidence relation is dependence edge in statistical significance, should be retained; If the heterogeneous coefficient of incidence relation is greater than heterogeneous coefficient threshold, illustrates that this incidence relation is not dependence edge in statistical significance, should be deleted.Can know thus, heterogeneous coefficient threshold is less, the limit comprised in relation on attributes network and the quantity of node fewer.

Finally, if there is not incidence relation between node and any node, then can delete described node from the grouping belonging to described node.

Be understandable that, if the incidence relation between a node and other nodes deleted after, this node and any node do not have incidence relation, so, then can by this knot removal.That is, can not there is isolated node in relation on attributes network, any node namely in relation on attributes network at least has incidence relation with other nodes arbitrary in this relation on attributes network.

In the present embodiment, by the operation information of the attribute information and described each object that obtain each object at least one object, and then according to the attribute information of described each object and the operation information of described each object, set up relation on attributes network, described relation on attributes network is made up of at least two nodes, the corresponding attribute information of each node; If the attribute information that two nodes are corresponding belongs to same object, between described two nodes, there is incidence relation, make it possible to according to described relation on attributes network, select at least one attribute information, using as combinations of attributes, the statistics of single attribute in prior art can be avoided accurately cannot to navigate to the problem of concrete object, thus improve the reliability of attributive analysis.

In addition, adopt the technical scheme that the application provides, due to the attribute information of the object of the current all categories in website can be got, therefore, based on the object of all categories of a website attribute information selected by combinations of attributes, effectively can improve the coverage rate of attributive analysis.

It should be noted that, for aforesaid each embodiment of the method, in order to simple description, therefore it is all expressed as a series of combination of actions, but those skilled in the art should know, the application is not by the restriction of described sequence of movement, because according to the application, some step can adopt other orders or carry out simultaneously.Secondly, those skilled in the art also should know, the embodiment described in instructions all belongs to preferred embodiment, and involved action and module might not be that the application is necessary.

In the above-described embodiments, the description of each embodiment is all emphasized particularly on different fields, in certain embodiment, there is no the part described in detail, can see the associated description of other embodiments.

The structural representation of the treating apparatus of the attribute information of the object that Fig. 3 provides for another embodiment of the application, as shown in Figure 3.The treating apparatus of the attribute information of the object of the present embodiment can comprise acquiring unit 31, set up unit 32 and selection unit 33.Wherein, acquiring unit 31, for the operation information of the attribute information and described each object that obtain each object at least one object; Set up unit 32, for according to the attribute information of described each object and the operation information of described each object, set up relation on attributes network; Wherein, described relation on attributes network is made up of at least two nodes, the corresponding attribute information of each node; If the attribute information that two nodes are corresponding belongs to same object, between described two nodes, there is incidence relation; The operation information of the weight of the described incidence relation each object belonging to attribute information corresponding to described two nodes, determines; Selection unit 33, for according to described relation on attributes network, selects at least one attribute information, using as combinations of attributes.

Alternatively, in one of the present embodiment possible implementation, the described operation information that described acquiring unit 31 obtains can include but not limited at least one item in click information, Information on Collection and purchase information, and the embodiment of the present application is not particularly limited this.

Such as, described acquiring unit 31 specifically can extract in the fixed time list object having operation behavior from database, and this list object can comprise following field:

The identification information of object, such as, the commodity sign product_ID of commodity;

The operation information of object, such as, the sales volume product_num of commodity;

Classification belonging to object, such as, the classification logotype category_ID of commodity;

The attribute information of object, such as, the attribute information attribute1 of commodity, attribute2 ..., attributeN, wherein, N is natural number.

Alternatively, in one of the present embodiment possible implementation, describedly set up unit 32, specifically for the attribute information according to described each object, determine the incidence relation between corresponding each node and described each node; According to the operation information of described each object, determine the weight of described incidence relation; And according to the weight of the incidence relation between described each node, described each node and described incidence relation, set up described relation on attributes network.Like this, owing to considering the incidence relation between attribute information, thus more effective information can be provided for attributive analysis.

Particularly, if attribute information corresponding to two nodes does not belong to same object, illustrate not have incidence relation between described two nodes, so, the weight of corresponding described incidence relation then can be set to 0; If the attribute information that two nodes are corresponding belongs to same object, illustrate to have incidence relation between described two nodes, so, the weight of corresponding described incidence relation then can be set to the numerical value being greater than 0.

As for the size of the weight value of related information, specifically can the operation information of the object belonging to two nodes corresponding to this incidence relation, determine.If two nodes corresponding to incidence relation only belong to an object, so, the weight of this incidence relation can only be determined according to the operation information of this object.Such as, the weight of incidence relation is set to the purchase volume in the fixed time of object.If two nodes corresponding to incidence relation belong to two objects or two or more object, so, the weight of this incidence relation can be determined according to the operation information of whole object.Such as, the weight of incidence relation is set to the purchase volume sum in the fixed time of whole object.

Like this, this relation on attributes network that is undirected, Weight (weight) has just established.The tables of data storing this relation on attributes network can be as follows: (attributeM, attributeN, weight), and wherein, N and M is mutually different natural number.

Be understandable that, in the topological diagram of relation on attributes network, between two nodes with incidence relation, can couple together with line, as a limit of relation on attributes network, between two nodes without incidence relation, without any line, as shown in Figure 2.In fig. 2, between node A and Node B, there is incidence relation, between node A and node C, not there is incidence relation, between Node B and node C, there is incidence relation.Node A and Node B can be called mutually the adjacent node of the other side, and Node B and node C can be called mutually the adjacent node of the other side.

Alternatively, in one of the present embodiment possible implementation, describedly set up unit 32, after performing corresponding operation, or performing in corresponding operating process, execution can also adjust operation as follows further, more there is statistical significance to make set up relation on attributes network.

Such as, set up if described unit 32 can also be further used for attribute information belong to pre-set can delete property information, delete the node that described attribute information is corresponding.As, the size of model, clothes.

Or, more such as, set up the weight that unit 32 can also be further used for incidence relation if described and be less than the weight threshold pre-set, delete described incidence relation.

Or, more such as, set up the quantity sum that unit 32 can also be further used for other nodes with node with incidence relation if described and be less than the amount threshold pre-set, delete described node and the incidence relation between described node and other nodes.Wherein, there is with node the quantity of other nodes of incidence relation, the angle value of node can also be called.

Be understandable that, if the incidence relation between a node and other nodes deleted after, this node and any node do not have incidence relation, so, then can by this knot removal.That is, can not there is isolated node in relation on attributes network, any node namely in relation on attributes network at least has incidence relation with other nodes arbitrary in this relation on attributes network.

Alternatively, in one of the present embodiment possible implementation, described selection unit 33, specifically may be used for utilizing corporations' detection algorithm, is divided into groups by the node in described relation on attributes network, to obtain at least two groupings; Wherein, the compactedness between the node in described each grouping be greater than node in described each grouping and other divide into groups in node between compactedness; Utilize key extraction algorithm, at least one incidence relation at least two groupings described in deletion; If there is not incidence relation between node and any node, from the grouping belonging to described node, delete described node.

Particularly, based on the relation on attributes network set up unit 32 and set up, in an initial condition, described selection unit 33 specifically may be used for by each node division to a grouping, utilizes $Q=\frac{1}{2m}\underset{i,j}{\mathrm{\Σ}}(w_{i,j}-\frac{s_i{s}_j}{2m})\mathrm{\δ}(c_i,c_j),$ Obtain packet parameters, be denoted as Q.

Wherein,

Q is packet parameters;

W _i,jfor the weight of the incidence relation between node i and node j, node j is other nodes in described relation on attributes network except node i;

$m=\frac{1}{2}\underset{i,j}{\mathrm{\Σ}}{w}_{i,j};$

$Q=\frac{1}{2m}\underset{i,j}{\mathrm{\Σ}}(w_{i,j}-\frac{s_i{s}_j}{2m})\mathrm{\δ}(c_i,c_j);$

S _ifor the intensity of node i,

C _igrouping belonging to node i, c _jfor the grouping belonging to node j, if c _i=c _j, then δ (c _i, c _j)=1; If c _i≠ c _j, then δ (c _i, c _j)=0;

To each node i, respectively node i and node j are carried out union operation, to be divided into same grouping, obtain the variation delta Q of the packet parameters after each merging; Variation delta Q is selected to be greater than 0 and the maximum group result of value, using grouping each in described group result as a new node; Wherein, this new node and the weight of the incidence relation of other nodes are the weight sum of the incidence relation of all nodes and these other nodes in corresponding grouping; And repeatedly perform described union operation, until the variation delta Q of packet parameters is less than or equal to 0, record group result now, as described at least two groupings.

For commodity, namely this division community structure out divides into groups to can be good at the incidence relation between characterization attributes information, be under the jurisdiction of the attribute information of a grouping more easily simultaneously in order to describe class I goods, and the weight of incidence relation between attribute information, also the related content that the operation information having quantized this kind of commodity reflects, such as, to sell fast degree.

The Crack cause of this corporations phenomenon has two classes:

The first kind is that these attribute informations portray the style describing a kind of commodity jointly, such as, this commodity class of dress now, Style:Sexy & Club, Silhouette:Sheath, Dresses Length:Above Knee & Mini tri-attribute informations are divided in a grouping, because these words all describe the specific features of " sexy skirt " this category, and also demonstrate the fast sale of this kind of commodity in sales volume data.

Equations of The Second Kind is the common element that these attribute informations reflect commodity of selling fast jointly, such as, the combination of these three attributes of Waistline:Nature, Sleeve Length:Short, Sleeve Style:Puff Sleeve, any combination of two that sales volume data show between these three attributes all achieves good sales volume, the commodity namely having these combinations of attributes more easily obtain the favor of buyer, and this is for instructing seller to replenish, guiding commodity to move towards all have practical significance.

It should be noted that, in the community structure mining process of attribute information, all properties information in same grouping might not be connected completely between two, this declared attribute combined information can break through existing commodity limitation, obtain the attribute with potential incidence relation, provide information of forecasting to the commodity that future may sell fast.

For playing exercisable directive function, need to screen the group result of final attribute information, remove the content easily causing misleading as far as possible, export the combinations of attributes wherein determined most, and limit the quantity of the attribute information in each combinations of attributes, make result have ubiquity and operability.For achieving the above object, described selection unit 33 adopts key extraction algorithm finally to process group result.Key extraction algorithm is that a kind of network proposed based on this network characteristic with community structure simplifies algorithm.It can according to the statistical property of nodes and limit (incidence relation namely between node) and interaction relationship, analyze and obtain the annexation with important statistical significance, thus the secondary annexation in removal network, extract the diaphyseal portion in network, both keep connectivity and the architectural characteristic of legacy network, highlight again the important annexation in network.Described selection unit 33 for each grouping, can adopt key extraction algorithm, specifically can perform following operation:

Operation one, described selection unit 33 set a number of nodes threshold value N ^*, after ensureing adopting key extraction algorithm, the number of nodes in each grouping is less than or equal to number of nodes threshold value N ^*.Number of nodes in the grouping that operation five exports is less than or equal to number of nodes threshold value N ^*time, the screening of this grouping is terminated, enters the screening step of next grouping.

Operation two, described selection unit 33 set a heterogeneous coefficient threshold set Ψ, comprise multiple heterogeneous coefficient threshold α in this heterogeneous coefficient threshold set Ψ ^*.In general, heterogeneous coefficient threshold α ^*for less real number, such as, 10 ^-3, 10 ^-6, 10 ^-10deng.For each heterogeneous coefficient threshold α ^*, can executable operations three, operation four and operation five.For ensureing that the number of nodes in grouping can reach full number of nodes foot threshold value, the heterogeneous coefficient threshold α comprised in heterogeneous coefficient threshold set Ψ ^*can be set as by 10 ^-3to 10 ^-10successively decrease.

Operation three, described selection unit 33 utilize obtain the significance level of each node in whole incidence relations of this node, be denoted as p _i,j;

Wherein,

P _i,jfor the significance level of node i in whole incidence relations of this node i;

W _i,jfor the weight of the incidence relation between node i and node j, node j is other nodes in described relation on attributes network except node i;

S _ifor the intensity of node i,

Operation four, described selection unit 33, according to the significance level of each node in whole incidence relations of this node, utilize obtain the heterogeneous coefficient of incidence relation, be denoted as α _i,j.

Wherein,

α _i,jfor the heterogeneous coefficient of the incidence relation between node i and node j;

K is the angle value of node i, namely has the quantity of other nodes of incidence relation with node i.

Operation five, described selection unit 33, according to the heterogeneous coefficient of incidence relation and heterogeneous coefficient threshold, carry out reservation process or delete processing to described incidence relation.

Particularly, if the heterogeneous coefficient of incidence relation is less than or equal to heterogeneous coefficient threshold, illustrates that this incidence relation is dependence edge in statistical significance, should be retained; If the heterogeneous coefficient of incidence relation is greater than heterogeneous coefficient threshold, illustrates that this incidence relation is not dependence edge in statistical significance, should be deleted.Can know thus, heterogeneous coefficient threshold is less, the limit comprised in relation on attributes network and the quantity of node fewer.

Be understandable that, if the incidence relation between a node and other nodes deleted after, this node and any node do not have incidence relation, so, then can by this knot removal.That is, can not there is isolated node in relation on attributes network, any node namely in relation on attributes network at least has incidence relation with other nodes arbitrary in this relation on attributes network.

In the present embodiment, the attribute information of each object at least one object and the operation information of described each object is obtained by acquiring unit, and then by setting up the operation information of unit according to the attribute information of described each object and described each object, set up relation on attributes network, described relation on attributes network is made up of at least two nodes, the corresponding attribute information of each node; If the attribute information that two nodes are corresponding belongs to same object, between described two nodes, there is incidence relation, make selection unit can according to described relation on attributes network, select at least one attribute information, using as combinations of attributes, the statistics of single attribute in prior art can be avoided accurately cannot to navigate to the problem of concrete object, thus improve the reliability of attributive analysis.

In addition, adopt the technical scheme that the application provides, due to the attribute information of the object of the current all categories in website can be got, therefore, based on the object of all categories of a website attribute information selected by combinations of attributes, effectively can improve the coverage rate of attributive analysis.

Be understandable that, adopt the technical scheme that the application provides, the combinations of attributes exported, can play a directive function to the user of the operator of website and website.The node of the relation on attributes network set up is the attribute information of object, while be the incidence relation between attribute information, because the weight of incidence relation is determined by the operation information of object, therefore, can select operation information according to Nutrition guide needs.

Such as, if need to carry out getting ready the goods, selecting the guidances such as goods, or need to carry out the guidances such as advertisement putting, or need to carry out the guidances such as purchase, etc., operation information can be set as the purchase information of object and the marketing information of object.

Or, more such as, if need to carry out the guidances such as potential user, operation information can be set as the Information on Collection of object.

Or, more such as, if need to carry out the guidances such as the page improves, operation information can be set as the click information of object.

The setting of the application to operation information is not particularly limited.

Those skilled in the art can be well understood to, and for convenience and simplicity of description, the system of foregoing description, the specific works process of device and unit, with reference to the corresponding process in preceding method embodiment, can not repeat them here.

In several embodiments that the application provides, should be understood that, disclosed system, apparatus and method, can realize by another way.Such as, device embodiment described above is only schematic, such as, the division of described unit, be only a kind of logic function to divide, actual can have other dividing mode when realizing, such as multiple unit or assembly can in conjunction with or another system can be integrated into, or some features can be ignored, or do not perform.Another point, the coupling each other representing or discuss or direct-coupling or communication connection can be by some interfaces, and the indirect coupling of device or unit or communication connection can be electrical, machinery or other form.

The described unit illustrated as separating component or can may not be and physically separates, and the parts represented as unit can be or may not be physical location, namely can be positioned at a place, or also can be distributed in multiple network element.Some or all of unit wherein can be selected according to the actual needs to realize the object of the present embodiment scheme.

In addition, each functional unit in each embodiment of the application can be integrated in a processing unit, also can be that the independent physics of unit exists, also can two or more unit in a unit integrated.Above-mentioned integrated unit both can adopt the form of hardware to realize, and the form that hardware also can be adopted to add SFU software functional unit realizes.

The above-mentioned integrated unit realized with the form of SFU software functional unit, can be stored in a computer read/write memory medium.Above-mentioned SFU software functional unit is stored in a storage medium, comprising some instructions in order to make a computer installation (can be personal computer, server, or network equipment etc.) or processor (processor) perform the part steps of method described in each embodiment of the application.And aforesaid storage medium comprises: USB flash disk, portable hard drive, ROM (read-only memory) (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), magnetic disc or CD etc. various can be program code stored medium.

Last it is noted that above embodiment is only in order to illustrate the technical scheme of the application, be not intended to limit; Although with reference to previous embodiment to present application has been detailed description, those of ordinary skill in the art is to be understood that: it still can be modified to the technical scheme described in foregoing embodiments, or carries out equivalent replacement to wherein portion of techniques feature; And these amendments or replacement, do not make the essence of appropriate technical solution depart from the spirit and scope of each embodiment technical scheme of the application.

Claims (12)

1. a disposal route for the attribute information of object, is characterized in that, comprising:

Obtain the attribute information of each object at least one object and the operation information of described each object;

According to the attribute information of described each object and the operation information of described each object, set up relation on attributes network; Wherein,

Described relation on attributes network is made up of at least two nodes, the corresponding attribute information of each node; If the attribute information that two nodes are corresponding belongs to same object, between described two nodes, there is incidence relation; The operation information of the weight of the described incidence relation each object belonging to attribute information corresponding to described two nodes is determined;

According to described relation on attributes network, select at least one attribute information, using as combinations of attributes.

2. method according to claim 1, is characterized in that, the operation information of the described attribute information according to described each object and described each object, sets up relation on attributes network, comprising:

According to the attribute information of described each object, determine the incidence relation between corresponding each node and described each node;

According to the operation information of described each object, determine the weight of described incidence relation;

According to the weight of the incidence relation between described each node, described each node and described incidence relation, set up described relation on attributes network.

3. method according to claim 2, is characterized in that, the operation information of the described attribute information according to described each object and described each object, sets up relation on attributes network, also comprises:

If attribute information belong to pre-set can delete property information, delete the node that described attribute information is corresponding;

If the weight of incidence relation is less than the weight threshold pre-set, delete described incidence relation;

If the quantity sum with node with other nodes of incidence relation is less than the amount threshold pre-set, delete described node and the incidence relation between described node and other nodes.

4. the method according to the arbitrary claim of claims 1 to 3, is characterized in that, describedly selects at least one attribute information according to described relation on attributes network, using as combinations of attributes, comprising:

Utilize corporations' detection algorithm, the node in described relation on attributes network is divided into groups, to obtain at least two groupings; Wherein, the compactedness between the node in described each grouping be greater than node in described each grouping and other divide into groups in node between compactedness;

Utilize key extraction algorithm, at least one incidence relation at least two groupings described in deletion;

If there is not incidence relation between node and any node, from the grouping belonging to described node, delete described node.

5. method according to claim 4, is characterized in that, describedly utilizes corporations' detection algorithm, is divided into groups by the node in described relation on attributes network, to obtain at least two groupings, comprising:

By each node division to a grouping, utilize obtain packet parameters; Wherein,

Q is packet parameters;

W _i,jfor the weight of the incidence relation between node i and node j, node j is other nodes in described relation on attributes network except node i;

$m=\frac{1}{2}\underset{i,j}{\mathrm{\Σ}}{w}_{i,j};$

$Q=\frac{1}{2m}\underset{i,j}{\mathrm{\Σ}}(w_{i,j}-\frac{s_i{s}_j}{2m})\mathrm{\δ}(c_i,c_j);$

S _ifor the intensity of node i,

C _igrouping belonging to node i, c _jfor the grouping belonging to node j, if c _i=c _j, then δ (c _i, c _j)=1; If c _i≠ c _j, then δ (c _i, c _j)=0;

To each node i, respectively node i and node j are carried out union operation, to be divided into same grouping, obtain the variation delta Q of the packet parameters after each merging;

Variation delta Q is selected to be greater than 0 and the maximum group result of value, using grouping each in described group result as a new node; Wherein, this new node and the weight of the incidence relation of other nodes are the weight sum of the incidence relation of all nodes and these other nodes in corresponding grouping;

Repeatedly perform described union operation, until the variation delta Q of packet parameters is less than or equal to 0, record group result now, as described at least two groupings.

6. the method according to the arbitrary claim of Claims 1 to 4, is characterized in that, described operation information comprises at least one item in click information, Information on Collection and purchase information.

7. a treating apparatus for the attribute information of object, is characterized in that, comprising:

Acquiring unit, for the operation information of the attribute information and described each object that obtain each object at least one object;

Set up unit, for according to the attribute information of described each object and the operation information of described each object, set up relation on attributes network; Wherein,

Described relation on attributes network is made up of at least two nodes, the corresponding attribute information of each node; If the attribute information that two nodes are corresponding belongs to same object, between described two nodes, there is incidence relation; The operation information of the weight of the described incidence relation each object belonging to attribute information corresponding to described two nodes is determined;

Selection unit, for according to described relation on attributes network, selects at least one attribute information, using as combinations of attributes.

8. device according to claim 7, is characterized in that, describedly sets up unit, specifically for

According to the attribute information of described each object, determine the incidence relation between corresponding each node and described each node;

According to the operation information of described each object, determine the weight of described incidence relation; And

According to the weight of the incidence relation between described each node, described each node and described incidence relation, set up described relation on attributes network.

9. device according to claim 8, is characterized in that, describedly sets up unit, also for

If attribute information belong to pre-set can delete property information, delete the node that described attribute information is corresponding;

If the weight of incidence relation is less than the weight threshold pre-set, delete described incidence relation;

If the quantity sum with node with other nodes of incidence relation is less than the amount threshold pre-set, delete described node and the incidence relation between described node and other nodes.

10. the device according to the arbitrary claim of claim 7 ~ 9, is characterized in that, described selection unit, specifically for

Utilize corporations' detection algorithm, the node in described relation on attributes network is divided into groups, to obtain at least two groupings; Wherein, the compactedness between the node in described each grouping be greater than node in described each grouping and other divide into groups in node between compactedness;

Utilize key extraction algorithm, at least one incidence relation at least two groupings described in deletion;

If there is not incidence relation between node and any node, from the grouping belonging to described node, delete described node.

11. devices according to claim 10, is characterized in that, described selection unit, specifically for

By each node division to a grouping, utilize obtain packet parameters; Wherein,

Q is packet parameters;

W _i,jfor the weight of the incidence relation between node i and node j, node j is other nodes in described relation on attributes network except node i;

$m=\frac{1}{2}\underset{i,j}{\mathrm{\Σ}}{w}_{i,j};$

$Q=\frac{1}{2m}\underset{i,j}{\mathrm{\Σ}}(w_{i,j}-\frac{s_i{s}_j}{2m})\mathrm{\δ}(c_i,c_j);$

S _ifor the intensity of node i,

C _igrouping belonging to node i, c _jfor the grouping belonging to node j, if c _i=c _j, then δ (c _i, c _j)=1; If c _i≠ c _j, then δ (c _i, c _j)=0;

To each node i, respectively node i and node j are carried out union operation, to be divided into same grouping, obtain the variation delta Q of the packet parameters after each merging;

Variation delta Q is selected to be greater than 0 and the maximum group result of value, using grouping each in described group result as a new node; Wherein, this new node and the weight of the incidence relation of other nodes are the weight sum of the incidence relation of all nodes and these other nodes in corresponding grouping;

Repeatedly perform described union operation, until the variation delta Q of packet parameters is less than or equal to 0, record group result now, as described at least two groupings.

12. devices according to the arbitrary claim of claim 7 ~ 11, it is characterized in that, the described operation information that described acquiring unit obtains comprises at least one item in click information, Information on Collection and purchase information.