CN103729362B

CN103729362B - The determination method and apparatus of navigation content

Info

Publication number: CN103729362B
Application number: CN201210387331.6A
Authority: CN
Inventors: 宁伟; 顾湘余
Original assignee: Alibaba Group Holding Ltd
Current assignee: Alibaba Group Holding Ltd
Priority date: 2012-10-12
Filing date: 2012-10-12
Publication date: 2017-07-21
Anticipated expiration: 2032-10-12
Also published as: CN103729362A

Abstract

This application discloses a kind of determination method and apparatus of navigation content, wherein, this method includes：Obtain the search term of input；The clicking rate of search result corresponding in each classification of search term in classification tree on each layer is obtained according to history access record；The classification shown according to required for being determined the clicking rate got, the classification that the above-mentioned needs determined are shown is shown as navigation content on webpage, and shows the search result of the above-mentioned class determined now.Present application addresses the low technical problem of the search efficiency that directly the corresponding classification of the corresponding all leaf nodes of search term is all shown and caused in the prior art, the technique effect for improving search efficiency has been reached.

Description

The determination method and apparatus of navigation content

Technical field

The application is related to internet arena, in particular to a kind of determination method and apparatus of navigation content.

Background technology

At present, in the application platforms such as search engine, portal website and e-commerce platform, user searches for convenience Related content, typically can be by search result according to certain criteria for classification progress tissue, and typically conventional is according to classification Form is shown to result, and user is further screened to classification, to reduce the scope of search result, so as to help user The object content for meeting itself search need is found faster more accurately.However, due to the limitation in user knowledge field, it is initial defeated The query word entered can not express the real or definite intention of user authentic and validly sometimes, be navigated by above-mentioned classification Mode can just point out the demand that user more accurately expresses oneself to be intended to, so as to guide user to enter by clicking on navigation information The clear and definite query demand of oneself in one step ground.

In common e-commerce website, website sets up the bibliography system of oneself according to the different classifications of merchandise news, The bibliography system of foundation is usually to carry out tissue with the structure of tree, such as, all merchandise newss related to mobile phone are all included into hand In machine classification, all merchandise newss related to mp3 are all included in mp3 classifications.Search engine is scanned for merchandise news Before, index can be initially set up, and all merchandise news to be searched for difference are tagged, and the label is used for identifying this business Product information is to belong to which or the corresponding classification of multiple leaf nodes, wherein, the corresponding classification of leaf node is exactly leaf class Mesh, wherein, not subdivisible classification unit in tree construction of the corresponding classification of leaf node for representing bibliography system, citing comes Say：Digital, computer-＞ mobile communication-＞ mobile phones, for such classification tree, mobile phone therein is not subdivisible classification list Member, mobile phone is the corresponding classification of leaf node of the classification tree.

At present, common navigation is typically all and, based on the corresponding classification of leaf node, asked in the inquiry for receiving user After asking, user can be presented to so that user enters according to the corresponding classification of leaf node belonging to all results that search is returned Row screening and navigate, the purpose of navigation can be realized in this way, but this mode is only applicable to user's request compares Specifically, clearly search.For example, user searches for one-piece dress, this search term has clearly indicated the search meaning of user very much Figure, and it is relatively more specific, the corresponding classification of search result corresponding to this search term is also generally all leaf node Corresponding classification.In this case, directly the corresponding classification of displaying leaf node is rational to user, however, user passes through Some more wide in range search terms can be often inputted, for example：User inputs " household " as search term, wide in range for this comparison Search term, can be shown that user thinks that search residence furniture may also think a variety of search intentions such as search bedding, therefore, obtain Search result it is also more wide in range, sometimes including multiple industries or multiple classifications.

If all inquiries determine that following ask occurs in navigation content all in the way of the corresponding classification of leaf node Topic：

, can if being directly targeted to the corresponding class of a point-device leaf node now firstly, for wide in range inquiry Can occur display the corresponding classification of leaf node and user desired by search result between there is no the situation of any relation, So that search result is inaccurate, the Experience Degree of user is reduced.Now, user can re-enter other query words and carry out again Search, because client to search engine server repeatedly sends inquiry request, the meeting in the case where search number of users is numerous Huge inquiry pressure is caused to search engine server.

Secondly, if navigating to the corresponding class of multiple leaf nodes now for wide in range inquiry, the search knot recalled Fruit number is generally on the high side, and result may be across the corresponding classification of multiple leaf nodes and the corresponding class of multiple non-leaf nodes Mesh, the industry being distributed sometimes may also have multiple.The quantity of one side search result can be relatively more, limited in display location In the case of, displaying result is difficult to cover being all intended to for user；On the other hand, the corresponding classification of the leaf node of multiple industries is mixed The sensation for being shown and easily being produced to user and comparing entanglement is combined, Consumer's Experience can be poor.Another further aspect, due to needing Match query operation is carried out now in the corresponding class of multiple leaf nodes, cause the data volume of match query huge, cause to search The efficiency reduction of hitch fruit, and during for sending a fairly large number of search result got to subscription client, also will Due to the huge of data volume network transfer speeds can be caused to decline, further influence search efficiency.

In view of the above-mentioned problems, not yet proposing effective solution.

The content of the invention

This application provides a kind of determination method and apparatus of navigation content, at least to solve directly search in the prior art The low technical problem of search efficiency that the corresponding classification of the corresponding all leaf nodes of rope word is all shown and caused.

According to the one side of the application there is provided a kind of determination method of navigation content, including：Obtain the search of input Word；Search result corresponding in each classification of above-mentioned search term in classification tree on each layer is obtained according to history access record Clicking rate；The classification shown according to required for being determined the clicking rate got, the classification that the above-mentioned needs determined are shown It is shown as navigation content on webpage, and shows the search result of the above-mentioned class determined now.

Preferably, according to required for being determined the clicking rate that gets displaying classification, above-mentioned need to show by what is determined Classification be shown as navigation content on webpage, and show that the search result of the above-mentioned class determined now includes：When The clicking rate of first leaf node corresponding classification of the search result in above-mentioned classification tree corresponding to above-mentioned search term is more than Or during equal to the first predetermined threshold, the corresponding classification of above-mentioned first leaf node is shown as above-mentioned navigation content；Or Person is pre- more than first in the clicking rate of the corresponding classification of each leaf node when the search result corresponding to above-mentioned search term Determine threshold value, but the clicking rate of the corresponding classification of one or more leaf nodes be more than the second predetermined threshold and said one or When the corresponding classification of multiple leaf nodes belongs to same father node, the corresponding classification of leaf node under above-mentioned father node is made It is shown for above-mentioned navigation content.

Preferably, according to required for being determined the clicking rate that gets displaying classification, above-mentioned need to show by what is determined Classification be shown as navigation content on webpage, and show that the search result of the above-mentioned class determined now includes：When Search result corresponding to above-mentioned search term the clicking rate of the corresponding classification of each leaf node be both less than the first predetermined threshold, And the search result corresponding to above-mentioned search term is more than the second predetermined threshold in the clicking rate of the corresponding classification of each leaf node Multiple leaf nodes corresponding classification when being not belonging to same father node, layer is carried out according to the set membership between each classification Level display.

Preferably, carrying out level displaying according to the set membership between each classification includes：It is determined that each leaf node pair The clicking rate of the father node for the classification answered, wherein, the clicking rate of above-mentioned father node is all child nodes pair under above-mentioned father node The sum of the clicking rate for the classification answered；The father node that rate is more than or equal to the 3rd predetermined threshold is will click on, and above-mentioned clicking rate is big The corresponding classification of all child nodes under the father node of the 3rd predetermined threshold is used as the first navigation content to be presented； By above-mentioned father node in the above-mentioned first navigation content to be presented and the corresponding classification of all child nodes under above-mentioned father node Level displaying is carried out according to the relation between father and son's node.

Preferably, it is determined that each after the clicking rate of the father node of the corresponding classification of leaf node, the above method is also wrapped Include：It is determined that the clicking rate of the upper layer node less than the father node of above-mentioned 3rd predetermined threshold, wherein, the click of above-mentioned upper layer node Rate is the sum of the clicking rate of the corresponding classification of all child nodes under above-mentioned upper layer node；It will click on rate pre- more than or equal to the 4th Determine the upper layer node of threshold value, and all child nodes that above-mentioned clicking rate is more than or equal under the upper layer node of the 4th predetermined threshold Corresponding classification is used as the second navigation content to be presented；By the above-mentioned first navigation content to be presented and above-mentioned second to be presented Navigation content according to each other set membership carry out level displaying.

Preferably, above-mentioned history access record includes：The clicking rate of the corresponding search result of above-mentioned search term, and/or on State the clicking rate of each classification in the corresponding navigation content of search term.

Preferably, according to history access record obtain in each classification of above-mentioned search term in classification tree on each layer it is right The clicking rate for the search result answered includes：It is each class assignment weighted value according to above-mentioned history access record, wherein, click on secondary The big weighted value of many classification correspondences of number, the small weighted value of the few correspondence of number of clicks；According to the weight for each class assignment Value determines the clicking rate of each above-mentioned classification.

According to further aspect of the application there is provided a kind of exhibiting device of navigation content, including：First obtains single Member, the search term for obtaining input；Second acquisition unit, for obtaining above-mentioned search term in classification according to history access record The clicking rate of corresponding search result in each classification in tree on each layer；Determining unit, for according to the click got The classification of displaying required for rate is determined, the classification that the above-mentioned needs determined are shown carries out exhibition as navigation content on webpage Show, and show the search result of the above-mentioned class determined now.

Preferably, above-mentioned determining unit includes：First display module, for when the search result corresponding to above-mentioned search term , will be above-mentioned when the clicking rate of the corresponding classification of the first leaf node in above-mentioned classification tree is more than or equal to the first predetermined threshold The corresponding classification of first leaf node is shown as above-mentioned navigation content；Or second display module, for that above-mentioned ought search Search result corresponding to rope word is not more than the first predetermined threshold in the clicking rate of the corresponding classification of each leaf node, still It is more than the second predetermined threshold and said one or multiple leaf sections in the clicking rate of the corresponding classification of one or more leaf nodes When the corresponding classification of point belongs to same father node, the corresponding classification of the leaf node under above-mentioned father node is regard as above-mentioned navigation Content is shown.

Preferably, above-mentioned determining unit includes：3rd display module, for when the search result corresponding to above-mentioned search term The first predetermined threshold and the search knot corresponding to above-mentioned search term are both less than in the clicking rate of the corresponding classification of each leaf node Fruit is more than the corresponding classification of multiple leaf nodes of the second predetermined threshold not in the clicking rate of the corresponding classification of each leaf node When belonging to same father node, level displaying is carried out according to the set membership between each classification.

In this application, behavior is clicked on by analyzing the history of user, to obtain the associated class point of destination of current search word Situation is hit, so as to determine that different navigation contents are shown according to different situations so that the navigation content of displaying is more accurate Really, it is to avoid user due to the navigation content of display is inaccurate and re-searching for for producing, reduce by re-searching for being made Into the inquiry pressure to search engine server；Simultaneously as what is returned is accurate navigation content (leaf node pair The classification answered) rather than the corresponding classification of all leaf nodes corresponding with search term, therefore, in the case where ensureing accuracy The quantity of the search result under the navigation content (the corresponding classification of leaf node) of return is reduced, searching for network transmission is reduced The quantity of hitch fruit, improves the return speed and search efficiency of search result.Efficiently solve through the above way existing The search efficiency that directly the corresponding classification of the corresponding all leaf nodes of search term is all shown and caused in technology is low Technical problem, reached improve search efficiency technique effect.

Brief description of the drawings

Accompanying drawing described herein is used for providing further understanding of the present application, constitutes the part of the application, this Shen Schematic description and description please is used to explain the application, does not constitute the improper restriction to the application.In the accompanying drawings：

Fig. 1 is a kind of preferred structure block diagram of the determination system according to the navigation content of the embodiment of the present application；

Fig. 2 is a kind of preferred structure block diagram of the determining device of the navigation content according to the embodiment of the present application；

Fig. 3 is a kind of preferred schematic diagram of the webpage to be presented with navigation content according to the embodiment of the present application；

Fig. 4 is another preferred structure block diagram of the determining device of the navigation content according to the embodiment of the present application；

Fig. 5 is a kind of preferred schematic diagram of the classification tree according to the embodiment of the present application；

Fig. 6 is another preferred schematic diagram of the classification tree according to the embodiment of the present application；

Fig. 7 is a kind of preferred flow charts of the determination method according to the navigation content of the embodiment of the present application；

Fig. 8 is another preferred flow charts of the determination method according to the navigation content of the embodiment of the present application；

Fig. 9 is a kind of preferred architecture schematic diagram of the determination system according to the navigation content of the embodiment of the present application；

Figure 10 is another preferred flow charts of the determination method according to the navigation content of the embodiment of the present application.

Embodiment

Describe the application in detail below with reference to accompanying drawing and in conjunction with the embodiments.It should be noted that not conflicting In the case of, the feature in embodiment and embodiment in the application can be mutually combined.

Before the further detail below of description present embodiments, it will be described with reference to Figure 1 and can be used for realizing this Shen The suitable counting system structure of one of principle please.In the following description, except as otherwise noted, otherwise will refer to by one or The action and the symbol of operation that multiple computers are performed represent to describe present embodiments.As such, it will be understood that sometimes Being referred to as this kind of action and operation of computer execution includes the processing unit of computer to representing data with structured form The manipulation of electric signal.It is safeguarded on this manipulation transforms data or position in the accumulator system of computer, this is with this The mode that the technical staff in field understands reconfigures or changed the operation of computer.The data structure for safeguarding data is that have The physical location of the memory of particular community defined in the form of data.Although however, this Shen described in above-mentioned context Please, but its being not intended to limit property, as understood by those skilled in the art, action described below and operation Each side can also be realized with hardware.

Accompanying drawing is turned to, wherein identical reference number refers to identical element, and the principle of the application is shown in a conjunction Realized in suitable computing environment.Describe based on described embodiments herein, and be not considered as on herein not below The alternative embodiment that is expressly recited and limit the application.

Fig. 1 shows the schematic diagram of an example computer architecture available for these equipment.For the mesh of description , the architecture painted is only an example of proper environment, and not use scope or the function proposition to the application are any Limitation.The computing system should not be also construed to have any dependence or demand to any component shown in Fig. 1 or its combination.

The principle of the application can use other universal or special calculating or communication environment or configure to operate.Suitable for this Well-known computing system, environment and the example of configuration of application include but is not limited to, personal computer, server, many places Manage device system, the system based on microprocessor, minicomputer, mainframe computer and the distribution including any said system or equipment Formula computing environment.

In its most basic configuration, the determination system 100 of the navigation content in Fig. 1 at least includes：The clothes of one website Business device 102 and one or more clients 104.Server 102 can include but is not limited to Micro-processor MCV or may be programmed patrol Collect processing unit, the storage device for data storage and the transmitting device with client communication of device FPGA etc.；Client End 104 can include：Micro-processor MCV, the transmitting device with server communication, the display device with user mutual.In this theory In bright book and claims, " system " can also be defined as being able to carry out software, firmware or microcode realizing appointing for function The display systems 100 of the integrated navigation content of what nextport hardware component NextPort or nextport hardware component NextPort can even is that it is distributed, to realize distribution Formula function.

As used in this application, term " module ", " component " or " unit " can refer to the display systems in navigation content The software object or routine performed on 100.Different components described herein, module, unit, engine and service can be implemented as The object or process of (for example, being used as single thread) are performed on the display systems 100 of navigation content.Although described herein System and method preferably realized with software, but the realization of the combination of hardware or software and hardware be also may and by structure Think.

Embodiment 1

Fig. 2 is a kind of preferred structural representation of the determining device 200 according to the embodiment of the present application navigation content, preferably , the device in the present embodiment is located on server 102 or in client 104.

In the embodiment of the present application, the webpage of display is mainly made up of two parts, as shown in figure 3, including：Navigation area and search Rope fruiting area, wherein, the related navigation information of navigation area displaying, for showing that the classification that can be clicked on is classified to user, so that So that user more easily scans for；Search results area is used to show the search result corresponding to the search term of user's input, Either displaying user clicks on the search result triggered after some classification in navigation information.

It is preferred that, the determining device 200 of navigation content as shown in Figure 2 can include：First acquisition unit 202, second Acquiring unit 204 and determining unit 206.These structures are specifically described below.

First acquisition unit 202, the search term for obtaining input；

Second acquisition unit 204, is coupled with first acquisition unit 202, for obtaining classification tree according to history access record In search result corresponding in each classification on each layer clicking rate；

For example, user's input search term " household ", corresponding just to obtain corresponding search during using " household " as search term As a result history access record, it is preferable that history access record includes：The clicking rate of the corresponding search result of search term, and/or The clicking rate of each classification in the corresponding navigation content of search term.In the preferred embodiment, the history access record of user Refer to all users stored in site databases and do not refer to going through for the user singly based on the history access record of the search term History accesses record.Meanwhile, above-mentioned history access record can correspond to the click of the user of each specific search result Rate or corresponding to the clicking rate corresponding to each classification in the navigation content of " household ".For example, during search " household ", User is 30% to the clicking rate of " sofa ", and user is 20% etc. to the clicking rate of " dining table ", then is accessed and remembered by history Record the clicking rate corresponding to search term obtained required for these data counted are exactly.Wherein, above-mentioned navigation can be Guided accordingly for the search under certain bibliography system to user or browse, classification is exactly and text or product The related taxonomic hierarchies of theme, can play a part of navigation in the search, and auxiliary user scans for.

Determining unit 206, is coupled with second acquisition unit 204, for being opened up required for being determined according to the clicking rate got The classification shown, the classification that the needs determined are shown is shown as navigation content on webpage, and is shown described The search result of the class determined now.

In the above-described embodiments, behavior is clicked on by analyzing the history of user, to obtain the related classification of current search word Click situation, so as to determine that different navigation contents are shown according to different situations so that the navigation content of displaying is more Plus it is accurate, it is to avoid user due to the navigation content of display is inaccurate and re-searching for of producing, reduce due to re-searching for The caused inquiry pressure to search engine server；Simultaneously as what is returned is accurate navigation content (leaf section The corresponding classification of point) rather than the corresponding classification of all leaf nodes corresponding with search term, therefore, ensureing the feelings of accuracy The quantity of the search result under the navigation content (the corresponding classification of leaf node) of return is reduced under condition, network transmission is reduced Search result quantity, improve the return speed and search efficiency of search result.Efficiently solve through the above way The search efficiency that directly the corresponding classification of the corresponding all leaf nodes of search term is all shown and caused in the prior art Low technical problem, has reached the technique effect for improving search efficiency.

, can be based on clicking rate point after the clicking rate of search term search result corresponding in each classification is got The difference of cloth determines the mode of its specific navigation displaying.In a preferred embodiment, as shown in figure 4, determining unit 206 It can include：First display module 402 and the second display module 404.

First display module 402, for when the search result corresponding to search term is in the corresponding classification of the first leaf node Clicking rate be more than or equal to the first predetermined threshold when, using the corresponding classification of the first leaf node be used as navigation content carry out exhibition Show；If, can be with for example, the clicking rate of some leaf node is 99% in history access record corresponding to some search term Determine that the search need of user can be obtained by satisfaction now in the corresponding class of the leaf node, therefore the leaf can be directly displayed Search result corresponding to the corresponding classification of child node and the corresponding classification of the leaf node, user is by once searching for The thing for oneself wanting search is can be obtained by, it is through referred to as classification.Wherein, the corresponding classification of leaf node is classification Minimum and not subdivisible classification unit, these more specific search terms such as " one-piece dress " in tree.

However, the corresponding search result of not all search term all concentrates on the corresponding classification of a leaf node, It may also concentrate in the corresponding classification of multiple leaf nodes corresponding to the corresponding classification of some non-leaf nodes, for this The situation of kind, can be focused on referred to as classification, and its corresponding navigation exhibition method can be that the non-leaf nodes is corresponding The corresponding classification of all leaf nodes of class now is all shown, can also be only by the corresponding class of the non-leaf nodes now The corresponding classification of several leaf nodes more concentrated is shown as navigation content, wherein, so-called concentration is one relative Concept, can according to the history of each node click on behavior judge click on intensity, it is preferable that intensity here Refer to the point in classification tree level of the node obtained after being counted to the clicking rate of all users where itself Rate size is hit, it is then determined that whether it is more than a default predetermined threshold, so that it is determined that the classification more concentrated in classification. In one preferred embodiment, the second display module 404 is used for when the search result corresponding to search term is in each leaf node The clicking rate of corresponding classification is not all more than the first predetermined threshold, but in the corresponding class point of destination of one or more leaf nodes When hitting rate and belonging to same father node more than the second predetermined threshold and the corresponding classification of the one or more leaf nodes, father is saved The corresponding all leaf nodes of point are shown as navigation content.Classification is focused on a specific embodiment below and carried out Illustrate：

It is a kind of preferred schematic diagram of classification tree as shown in Figure 5, from the content shown by Fig. 5, in all leaves In the corresponding classification of node, only B21 and B23 are to concentrate, and find out the node of all concentrations, are used as the node finally shown. For example, it can be set to clicking rate just be can be regarded as the classification more concentrated in this layer more than 30% and (may be used in the corresponding classification of leaf node again To be referred to as node).Simultaneously it can be found that B21 and B23 is the B2 corresponding classification of leaf node, it is therefore contemplated that user Click demand also just can be almost met under B2.It therefore, it can the B2 corresponding classification of direct child node： B21, B22 and B23 are shown as navigation content to be shown.Preferably, B21 and B23 can also only be shown either B2 can also be also served as into navigation content to be shown, here it is so-called classification is focused on.

In actual applications, it is also possible to clicking rate in the corresponding history access record of some corresponding search term occur higher The corresponding classification of several leaf nodes be not belonging to the situation of the corresponding classification of same non-leaf nodes, so pass through above-mentioned class The mode that mesh is focused on, which show the content that needs show, will also compare many, in this case, it is possible to according to classification it Between set membership carry out level displaying.That is, determine to show which non-leaf nodes is corresponding according to the clicking rate of each classification Classification, and its corresponding classification of corresponding leaf node, can be called classification level for this ways of presentation.At one preferably In embodiment, determining unit includes：3rd display module, for when the search result corresponding to search term is in each leaf section The clicking rate of the corresponding classification of point is both less than the first predetermined threshold, and clicking rate is more than multiple leaf nodes of the second predetermined threshold When corresponding classification is not belonging to same father node, level displaying is carried out according to the set membership between each classification.

Specifically, when the displaying of classification level is carried out, it may be determined that the father node of the corresponding classification of leaf node Clicking rate whether Relatively centralized, it is preferable if the clicking rate of father node do not concentrate can continue up layer push away until correspondence leaf The former generation of the corresponding classification of child node or the node more concentrated in its corresponding layer that belongs to of ancestors either shift root section onto Untill point.That is, if corresponding bottom one layer at its of the corresponding classification of the corresponding leaf node of one search term is more to divide Scattered, i.e., its corresponding clicking rate is relatively low, and whether the clicking rate for determining that its father node is relative in the corresponding layer of father node Concentrate, i.e., whether corresponding clicking rate is higher, if it is, all child nodes under its corresponding father node are corresponding Classification is all shown as final navigation content, if the clicking rate of the corresponding father node of leaf classification is not also high, base Layer is continued up in its father node to judge until finding untill respective layer belongs to the node more concentrated or root node.It is excellent Selection of land, the clicking rate of each upper layer node can carry out accumulating operation by the clicking rate of its corresponding lower level node and obtain. The displaying of classification level is specifically described with a specific example below：

It is another preferred schematic diagram of classification tree as shown in Figure 6, classification level is according to from the corresponding class of leaf node What the mode that mesh is searched upwards was carried out, until it is to concentrate in its respective layer to find corresponding father node either ancestors' node In the case of, just stop searching judging.In classification tree as shown in Figure 6, because B12 in its respective layer is scattered, it is thus determined that Whether its father node (B1) is to concentrate, it is determined that in the case of being, by the corresponding classification of B1 all child nodes corresponding with its (B11, B12) is all shown as navigation content, likewise, itself being scattered, its corresponding father's section for C11 and C12 Point (C1) is also scattered, but its corresponding ancestors' node (C) is to concentrate, therefore, and its ancestors' node C and its ancestors are saved The corresponding classification of child node (C11 and C12) of point is shown as navigation content.Therefore, for the classification tree shown in Fig. 6, It is exactly eventually as what navigation content carried out level displaying：C：C1、C2；B1：B11、B12；B2：B21、B22、B23；E3：E31、 E32.Wherein, the mode of level displaying can be exhibition method as shown in Figure 3, i.e. enter according to set membership each other Row level is shown.

In each above-mentioned preferred embodiment, above-mentioned history access record includes but is not limited to the corresponding search of search term As a result the clicking rate of each classification in clicking rate, and/or the corresponding navigation content of search term.

Can be each class assignment weight to preferably confirm the clicking rate of each classification node in classification structure Value, the click situation that the weighted value of corresponding setting can be in history access record determines that clicking rate is high can just to be set Put big weight, clicking rate it is low small weight can be just set.In a preferred embodiment, visited according to the history of user Ask that the clicking rate that record obtains search term search result corresponding in each classification includes：It is each according to history access record Individual class assignment weighted value, wherein, the corresponding big weighted value of classification more than number of clicks, the small weight of the few correspondence of number of clicks Value；The clicking rate of each classification is determined according to the weighted value for each class assignment.

Embodiment 2

On the basis of the device shown in Fig. 2 and Fig. 4, this application provides a kind of determination side of preferred navigation content Method.As shown in fig. 7, this method may include steps of：

Step S702：Obtain the search term of input；

Step S704：According to corresponding to obtaining in each classification of search term in classification tree on each layer history access record Search result clicking rate；

For example, user's input search term " household ", corresponding just to obtain corresponding search during using " household " as search term As a result history access record, it is preferable that history access record includes：The clicking rate of the corresponding search result of search term, and/or The clicking rate of each classification in the corresponding navigation content of search term.In the preferred embodiment, the history access record of user Refer to all users stored in site databases and do not refer to going through for the user singly based on the history access record of the search term History accesses record.Meanwhile, above-mentioned history access record can correspond to the click of the user of each specific search result Rate or corresponding to the clicking rate corresponding to each classification in the navigation content of " household ".For example, during search " household ", All users are 30% to the clicking rate of " sofa ", and the clicking rate to " dining table " is 20% etc., then accessed and remembered by history Record the clicking rate corresponding to search term obtained required for these data counted are exactly.Wherein, above-mentioned navigation can be Guided accordingly for the search under certain bibliography system to user or browse, classification is exactly and text or product The related taxonomic hierarchies of theme, can play a part of navigation in the search, and auxiliary user scans for.

Step S706：The classification of displaying according to required for being determined the clicking rate that gets, described needs to open up by what is determined The classification shown is shown as navigation content on webpage, and the search result of the class determined described in displaying now.

, can be based on clicking rate point after the clicking rate of search term search result corresponding in each classification is got The difference of cloth determines the mode of its specific navigation displaying.Preferably, navigation content can be carried out according to following several ways Displaying：

When the search result corresponding to search term in the clicking rate of the corresponding classification of the first leaf node is more than or equal to the During one predetermined threshold, the corresponding classification of the first leaf node is shown as navigation content；That is, during this time can be Only show that the clicking rate is more than or equal to the corresponding classification of leaf node of the first predetermined threshold in navigation area, and in result exhibition Show area's directly search result of the displaying corresponding to the search term.If for example, the history access record corresponding to some search term In the clicking rate of some leaf node be 99%, then can determine the search need of user in the corresponding class of the leaf node now Satisfaction is can be obtained by, therefore the corresponding classification of the leaf node and the corresponding classification institute of the leaf node can be directly displayed Corresponding search result, the thing that user can be obtained by oneself and want search by once searching for, referred to as class Mesh goes directly.Wherein, the corresponding classification of leaf node is minimum in classification tree and not subdivisible classification unit, for example " even clothing These more specific search terms such as skirt ".

However, the corresponding search result of not all search term all concentrates on the corresponding classification of a leaf node, It may also concentrate in the corresponding classification of multiple leaf nodes corresponding to the corresponding classification of some non-leaf nodes, for this The situation of kind, can be focused on referred to as classification, and its corresponding navigation exhibition method can be that the non-leaf nodes is corresponding The corresponding classification of all leaf nodes of class now is all shown, can also be only by the corresponding class of the non-leaf nodes now The corresponding classification of the higher leaf node of several clicking rates is shown as navigation content, wherein, so-called concentration is a phase To concept, can according to the history of each node click on behavior judge click on intensity, it is preferable that collection intermediate range here Degree refers in classification tree level of the node obtained after being counted to the clicking rate of all users where itself Clicking rate size, it is then determined that whether it is more than a default predetermined threshold, so that it is determined that the classification more concentrated in classification. In a preferred embodiment, when the search result corresponding to search term is in the clicking rate of the corresponding classification of each leaf node All it is not more than the first predetermined threshold, but is more than the second predetermined threshold in the clicking rate of the corresponding classification of one or more leaf nodes Value and when the corresponding classification of one or more leaf nodes belongs to same father node, by all leaf nodes under the father node Corresponding classification is shown as navigation content.

Classification is focused on a specific embodiment below and is specifically described：

It is a kind of preferred schematic diagram of classification tree as shown in Figure 5, from the content shown by Fig. 5, in all leaves In the corresponding classification of node, only B21 and B23 are to concentrate, and find out the node of all concentrations, are used as the node finally shown. For example, it can be set to clicking rate just be can be regarded as the classification more concentrated in this layer more than 30% and (may be used in the corresponding classification of leaf node again To be referred to as node).Simultaneously it can be found that B21 and B23 is the B2 corresponding classification of leaf node, it is therefore contemplated that user Click demand also just can be almost met under B2.It therefore, it can the B2 corresponding classification of direct child node： B21, B22 and B23 are shown as navigation content to be shown.Preferably, B21 and B23 can also only be shown either B2 can also be also served as into navigation content to be shown.Here it is so-called classification is focused on.

In actual applications, it is also possible to clicking rate in the corresponding history access record of some corresponding search term occur higher The corresponding classification of several leaf nodes be not belonging to the situation of the corresponding classification of same non-leaf nodes, so pass through above-mentioned class The mode that mesh is focused on, which show the content that needs show, will also compare many, in this case, it is possible to according to classification it Between set membership carry out level displaying.That is, determine to show which non-leaf nodes is corresponding according to the clicking rate of each classification Classification, and its corresponding classification of corresponding leaf node, can be called classification level for this ways of presentation.At one preferably In embodiment, recalling for search term search result corresponding in each classification is obtained according to the history access record of user Rate includes：Make a reservation for when the search result corresponding to search term is both less than first in the clicking rate of the corresponding classification of each leaf node Threshold value, and clicking rate is more than multiple leaf classifications of the second predetermined threshold when being not belonging to same father node, according to each classification Between set membership carry out level displaying.

Specifically, when the displaying of classification level is carried out, can be realized according to mode as shown in Figure 8, including it is following Step：

Step S802：It is determined that the clicking rate of the father node of the corresponding classification of each leaf node, wherein, the click of father node Rate is the sum of the clicking rate of the corresponding classification of all child nodes under father node；

Step S804：The father node that rate is more than or equal to the 3rd predetermined threshold is will click on, and it is all under the father node The corresponding classification of child node is used as the first navigation content to be presented；

Step S806：It is determined that the clicking rate of the upper layer node less than the father node of the 3rd predetermined threshold, wherein, upper layer node Clicking rate be all child nodes under upper layer node clicking rate sum；

Step S808：The upper layer node that rate is more than or equal to the 4th predetermined threshold is will click on, and under the upper layer node All child nodes are used as the second navigation content to be presented, i.e. by the upper layer node of determination and its corresponding direct sub- section The corresponding classification of point is all shown as navigation content；Can be preferably for the 3rd above-mentioned threshold value and the 4th threshold value The numerical value determined according to the characteristic of each level, does not have inevitable contact between the first and second threshold values.

Step S810：By the first navigation content to be presented and the second navigation content to be presented according to father each other Subrelation carries out level displaying.

That is, determine the corresponding classification of leaf node father node clicking rate whether Relatively centralized, it is preferable if father save The clicking rate of point, which is not concentrated, can continue up the category that layer pushes away former generation or ancestors up to the corresponding classification of correspondence leaf node Untill the node more concentrated in its corresponding layer either shifts root node onto.That is, if the corresponding leaf of one search term Corresponding bottom one layer at its of the corresponding classification of node is more scattered, i.e., its corresponding clicking rate is relatively low, determines that it Whether the clicking rate of father node is Relatively centralized in the corresponding layer of father node, i.e., whether corresponding clicking rate is higher, such as Fruit is then all to be shown all child nodes under its corresponding father node as final navigation content, if the leaf section The clicking rate of the corresponding father node of point is not also high, then continuing up layer based on its father node judges in respective layer to belong to until finding Untill the node more concentrated either root node.Preferably, the clicking rate of each upper layer node can be corresponding by its The clicking rate of lower level node carries out accumulating operation and obtained.The displaying of classification level is carried out specifically with a specific example below It is bright：

Embodiment 3

The embodiment of the present application provides a kind of preferred embodiment further to explain the application, but is worth note Meaning, the preferred embodiment is intended merely to preferably describe the application, does not constitute and the application is improperly limited.

For following not enough present in prior art：First, the corresponding classification of displaying leaf node completely is led to user Boat form is excessively single；Secondly, the navigation content for wide in range inquiry is unfriendly, and navigation classification displaying is more chaotic, is unfavorable for Result class destination aggregation (mda) needed for user finds oneself.The navigation for directly carrying out the corresponding classification of leaf node is difficult that accurate determination is used The true intention of family search, sometimes to the positioning made mistake.

, being capable of the analysis by the search term to user and the history according to user in method provided herein Access record and determine the distribution situation of the clicking rate of each classification in search result, and different lead is provided according to different demands Boat form, i.e. provided for wide in range and accurate inquiry based on varigrained navigation, it is preferable that can include following several Mode：Classification is through, classification is focused on, classification level and classification tiling etc., thus effectively determine the search intention of user so as to More accurately determine search result.Wherein,

Classification is gone directly, and the search term of user exactly is directly targeted into some maximally related classification, is carried out so as to obtain user Primary screening can be obtained by the result for wanting search.

Classification is focused on, and several maximally related classifications are directly shown aiming at search term, and automatic shield falls other non-phases The classification of pass, screening efficiency is lifted with the screening scope for reducing user.

Classification level, be namely based on user to search term according to classification tree construction show different subtrees, provide class The path of mesh set membership, helps user to determine search intention with faster speed when user carries out wide in range search.

Classification tile, be exactly user search intention it is especially wide in range when, directly show all related to search term The corresponding classification of leaf node.

Above-mentioned several ways are specifically described below：

The click situation of the corresponding classification of each leaf node is determined according to the distribution of the clicking rate in history access record To provide different form of navigation：

1) for wide in range inquiry, the searching structure of classification includes the corresponding classification of multiple non-leaf nodes, in this case Can according to classification architecture, layering displaying is carried out to result, referred to as classification level can also be called level navigation；

2) for more accurate specific inquiry, the corresponding classification Relatively centralized of leaf node belonging to search result, one As generally concentrate on the corresponding classification of leaf node of the corresponding class of some non-leaf nodes now, such case can directly determine Position is to the leaf node under the non-leaf nodes so that user is screened, and referred to as classification is focused on；

3) for very accurate specific search term, and corresponding search result largely belongs to a leaf node correspondence Classification when, search result directly can be navigated into the corresponding classification of this leaf node, and automatically by search engine mistake The result of the corresponding classification of other leaf nodes is filtered, referred to as classification goes directly；

If 4) distribution of searching structure can be covered by multiple leaf nodes (the corresponding classification of leaf node i.e. above) Lid, but directly can not can then make the several leaf nodes for covering search result specific to some specific upper layer nodes Shown for navigation content, this mode tiles as classification.

I.e. according to different query intentions, the different form of navigation of user is provided, specifically there are four kinds：Classification goes directly, classification Focus on, classification level, and classification tiling.

In aforesaid way, first according to the correlation between text and large-scale historic user behavioural analysis user Original query and the classification that is clicked between correlation, so that the query intention of user is navigated into all related leaves The corresponding classification of node, then judges specific navigation display form, on the one hand analyzes the intended scope of query word, that is, determine to use The query intention at family is wide in range or specific, on the one hand by bibliography system structure, by the corresponding class of all related leaf nodes Mesh, which is built, claims its corresponding classification tree construction, and all tree nodes can assign different weights according to the situation of click, finally lead to Cross the level that Threshold Analysis judges displaying.

The embodiment of the present application additionally provides a kind of system general frame for realizing above-mentioned way of search, as shown in figure 9, this is System includes four parts：

Front-end server, the inquiry for receiving user, distribution request (includes navigation algorithm to smartquery platforms Platform) and search engine；

Smartquery platforms, are an algorithm platforms for being used to handle inquiry, above-mentioned navigation display method service is just It can be deployed on the machine of this platform；

Search engine, for Query Result；

Tool server is runed, for manually being runed.

Based on above-mentioned system architecture, the embodiment of the present application additionally provides a kind of preferred handling process, as shown in Figure 10, Including (step S1002 to step S1012)：

Step S1002：Obtain the requested next search term in search engine front end.

Step S1004：By analyzing history access record in large-scale user's history daily record, the classification of search term is judged Tendency, what classification in this step was referred to is the corresponding classification of leaf node in classification structure.

Step S1006：Judge whether it is that classification goes directly, it is preferable that through classification is to search for very clear and definite for user Query word, and the user for analyzing almost all (99%) by the historical behavior of website clicks on behavior and concentrates on this class Mesh, that is, user demand situation about can be met substantially now in this class, what navigation will be gone directly with classification Mode is showed, and the search of one query word, the knot actually obtained have only been carried out although the effect finally showed is user Fruit is result of the obtained result of this search term after classification filtering, so that the result that user's inquiry is obtained is more Accurately, the path of user's search is effectively shortened, the experience effect of user is improved.

Step S1008：Judge whether it is that classification is focused on, classification focuses on the search term referred to for user, and search engine is returned Return result and be likely distributed in multiple classifications, but the history of user clicks on behavior and generally concentrates on certain several classification (such as 3), Namely the demand of user can be all met substantially now in these classes.In this case, navigation can be poly- with classification Burnt mode is shown.The final distribution of results returned to some search term that the effect of user is user that embodies is in multiple classes Mesh, but navigation only shows certain several classification therein, and the screening scope of user can be effectively reduced in this way, helps User's automatic shield falls other irrelevant classifications, has reached reduction screening cost, improves the effect of Consumer's Experience.

Step S1010：Judge whether it is classification level, classification level is usually more wide in range in user's search term, is returned Search result and the history of user click in the case that behavior also comparatively disperses and use, one side leaf node pair The classification distributed quantity answered is more, and on the other hand the father node of the corresponding classification of these leaf nodes also has multiple, that is, is distributed in It is difficult that the search intention of user is directly navigated into some specific leaves for this search situation on multiple different classifications The corresponding classification of node, it is necessary to the leaf of displaying if all corresponding classifications of leaf node are exposed directly into user The quantity of the corresponding classification of child node is too many and very numerous and diverse, without logic, so as to it is difficult to allow user to be looked for less cost To the classification navigation oneself needed, Consumer's Experience can be poor, influences the click behavior of navigation area user, in this case, Just can be in the way of level be navigated, corresponding classification of leaf node that will be all is according to the structure in classification tree System, carries out a point heap, and click on probability etc. according to user extracts suitable classification subtree as classification sorting technique in classification tree, The effect finally shown is to click on that classification is organized into different subtrees according to high-rise father node with preference according to the history of user Structure, the navigation orderliness shown in this way is more clear, while it is easy to different users to select different sub-tree structures, Improve search efficiency.

I.e., as shown in Figure 6 above, illustrate, built according to bibliography system structure so that classification tree is three-level classification as an example One classification tree, further according to the intensity that behavior judges to click on of clicking on of each classification.It is according to from leaf in classification level What the mode that the corresponding classification of child node is searched upwards was carried out, until it is right at its to find corresponding father node either ancestors' node Answer in the case of being concentration in layer, just stop searching judgement.

In the classification tree shown in Fig. 6, because B12 in its respective layer is scattered, it is thus determined that its father node (B1) whether It is to concentrate, it is determined that in the case of being, B1 all child nodes corresponding with its are all shown as navigation content, equally , itself it is scattered, its corresponding father node (C1) is also scattered for C11 and C12, but its corresponding ancestors' section Point (C) is to concentrate, and therefore, its ancestors' node C He its ancestors' node child node (C11 and C12) are entered as navigation content Row displaying.Therefore, for Fig. 6, finally show：C：C1、C2；B1：B11、B12；B2：B21、B22、B23：E3：E31、 E32。

Step S1012：If being unsatisfactory for all of the above scene, show that the corresponding search result of the search term can be It is covered to, but can not be reasonably assigned under certain scene, this feelings in the corresponding classification of a number of leaf node Condition just can using classification tile exhibition method, i.e. using the corresponding classification of all leaf nodes that can be covered all as Final navigation content is shown.

In the above-described embodiment, behavior is clicked on by analyzing the history of user, to obtain all phases of current search word The corresponding classification of leaf node is closed, through according to classification then according to different scenes, classification is focused on, classification level and class The order of mesh tiling is judged, when it is determined that some scene met is just exited, as the output result of final navigation content, Suitable navigation content can be provided according to the difference of search term through the above way to be shown to improve search efficiency, Also Consumer's Experience is improved.

Description more than, the application has the advantage that：Behavior is clicked on by the history for analyzing user, to obtain The associated class purpose for obtaining current search term clicks on situation, so as to determine that different navigation contents carry out exhibitions according to different situations Show so that the navigation content of displaying is more accurate, it is to avoid user is because the navigation content of display is inaccurate and produces again Search, reduces the inquiry pressure to search engine server caused by re-searching for；Simultaneously as return be compared with For accurate navigation content (the corresponding classification of leaf node) rather than the corresponding class of all leaf nodes corresponding with search term Mesh, therefore, the search under the navigation content (the corresponding classification of leaf node) of return is reduced in the case where ensureing accuracy As a result quantity, reduces the quantity of the search result of network transmission, improves the return speed and search efficiency of search result. Efficiently solve through the above way in the prior art directly by the corresponding classification of the corresponding all leaf nodes of search term all The low technical problem of the search efficiency that is shown and causes, has reached the technique effect for improving search efficiency.

Obviously, those skilled in the art should be understood that each module or each step of above-mentioned the application can be with general Computing device realize that they can be concentrated on single computing device, or be distributed in multiple computing devices and constituted Network on, alternatively, the program code that they can be can perform with computing device be realized, it is thus possible to they are stored Performed in the storage device by computing device, and in some cases, can be shown to be performed different from order herein The step of going out or describe, they are either fabricated to each integrated circuit modules respectively or by multiple modules in them or Step is fabricated to single integrated circuit module to realize.So, the application is not restricted to any specific hardware and software combination.

The preferred embodiment of the application is the foregoing is only, the application is not limited to, for the skill of this area For art personnel, the application can have various modifications and variations.It is all within spirit herein and principle, made any repair Change, equivalent substitution, improvement etc., should be included within the protection domain of the application.

Claims

1. a kind of determination method of navigation content, it is characterised in that including：

Obtain the search term of input；

Search knot corresponding in each classification of the search term in classification tree on each layer is obtained according to history access record The clicking rate of fruit；

The classification of displaying according to required for being determined the clicking rate that gets, using the classification of the needs determined displaying as leading Boat content is shown on webpage, and the search result of the class determined described in displaying now；

Wherein, after the distribution situation of the clicking rate of each classification in determining the search result, provided according to different demands Different form of navigation.

2. according to the method described in claim 1, it is characterised in that the class shown according to required for being determined the clicking rate got Mesh, the classification that the needs determined are shown be shown as navigation content on webpage including：

When the click of first leaf node corresponding classification of the search result corresponding to the search term in the classification tree When rate is more than or equal to the first predetermined threshold, exhibition is carried out using the corresponding classification of first leaf node as the navigation content Show；Or

When the search result corresponding to the search term the corresponding classification of each leaf node clicking rate be not more than first Predetermined threshold, but it is more than the second predetermined threshold and one in the clicking rate of the corresponding classification of one or more leaf nodes Or multiple leaf nodes corresponding classification is when belonging to same father node, by the corresponding classification of leaf node under the father node It is shown as the navigation content.

3. according to the method described in claim 1, it is characterised in that the class shown according to required for being determined the clicking rate got Mesh, the classification that the needs determined are shown be shown as navigation content on webpage including：

When the search result corresponding to the search term, in the clicking rate of the corresponding classification of each leaf node to be both less than first pre- Determine threshold value and search result corresponding to the search term in the clicking rate of the corresponding classification of each leaf node to be more than second pre- When the corresponding classification of multiple leaf nodes for determining threshold value is not belonging to same father node, according to the set membership between each classification Carry out level displaying.

4. method according to claim 3, it is characterised in that level exhibition is carried out according to the set membership between each classification Show including：

It is determined that the clicking rate of the father node of the corresponding classification of each leaf node, wherein, the clicking rate of the father node is described The sum of the clicking rate of the corresponding classification of all child nodes under father node；

The father node that rate is more than or equal to the 3rd predetermined threshold is will click on, and the clicking rate is more than or equal to the 3rd predetermined threshold The corresponding classification of all child nodes of the father node of value is used as the first navigation content to be presented；

Father node described in described first navigation content to be presented and all child nodes under the father node is corresponding Classification carries out level displaying according to the relation between father and son's node.

5. method according to claim 4, it is characterised in that it is determined that the father node of the corresponding classification of each leaf node Clicking rate after, methods described also includes：

It is determined that the clicking rate of the upper layer node less than the father node of the 3rd predetermined threshold, wherein, the point of the upper layer node The rate of hitting is the sum of the clicking rate of all subcategories under the upper layer node；

The upper layer node that rate is more than or equal to the 4th predetermined threshold is will click on, and the clicking rate is more than or equal to the 4th and made a reservation for The corresponding classification of all child nodes under the upper layer node of threshold value is used as the second navigation content to be presented；

Described first navigation content to be presented and second navigation content to be presented are closed according to father and son each other System carries out level displaying.

6. method according to any one of claim 1 to 5, it is characterised in that the history access record includes：It is described The clicking rate of each classification in the clicking rate of the corresponding search result of search term, and/or the corresponding navigation content of the search term.

7. method according to any one of claim 1 to 5, it is characterised in that according to being obtained history access record The clicking rate of corresponding search result includes in each classification of search term in classification tree on each layer：

It is each class assignment weighted value according to the history access record, wherein, the corresponding big power of classification more than number of clicks Weight values, the small weighted value of the few correspondence of number of clicks；

The clicking rate of each classification according to being determined the weighted value for each class assignment.

8. a kind of determining device of navigation content, it is characterised in that including：

First acquisition unit, the search term for obtaining input；

Second acquisition unit, for obtaining each classification of the search term in classification tree on each layer according to history access record In corresponding search result clicking rate；

Determining unit, the classification of displaying required for for being determined according to the clicking rate that gets described needs exhibition by what is determined The classification shown is shown as navigation content on webpage, and the search result of the class determined described in displaying now；

Wherein, the determining unit is after the distribution situation of the clicking rate of each classification in determining the search result, according to Different demands provide different form of navigation.

9. device according to claim 8, it is characterised in that the determining unit includes：

First display module, for the first leaf node when the search result corresponding to the search term in the classification tree When the clicking rate of corresponding classification is more than or equal to the first predetermined threshold, the corresponding classification of first leaf node is regard as institute Navigation content is stated to be shown；Or

Second display module, for when the search result corresponding to the search term is in the corresponding class point of destination of each leaf node Hit rate and be not all more than the first predetermined threshold, but it is pre- more than second in the clicking rate of the corresponding classification of one or more leaf nodes Threshold value is determined and when the corresponding classification of one or more of leaf nodes belongs to same father node, by the leaf under the father node The corresponding classification of child node is shown as the navigation content.

10. device according to claim 8, it is characterised in that the determining unit includes：

3rd display module, for when the search result corresponding to the search term is in the corresponding class point of destination of each leaf node The rate of hitting is both less than the first predetermined threshold and search result corresponding to the search term is in the corresponding classification of each leaf node When multiple leaf classifications that clicking rate is more than the second predetermined threshold are not belonging to same father node, according to the father between each classification Subrelation carries out level displaying.