CN103176995B - A kind of method of information navigation, equipment and system - Google Patents
A kind of method of information navigation, equipment and system Download PDFInfo
- Publication number
- CN103176995B CN103176995B CN201110432357.3A CN201110432357A CN103176995B CN 103176995 B CN103176995 B CN 103176995B CN 201110432357 A CN201110432357 A CN 201110432357A CN 103176995 B CN103176995 B CN 103176995B
- Authority
- CN
- China
- Prior art keywords
- attribute
- property value
- public
- public attribute
- query word
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the present application provides a kind of method, equipment and system of information navigation, comprise: when carrying out information navigation, no longer be confined to leaf class attribute information now, but when determining the query word that client provides, determine at least one public attribute that this query word is corresponding, and the property value that each described public attribute is corresponding, and the property value of public attribute corresponding for this query word and each public attribute is pushed to client, thus solve prior art Problems existing, and meet the Content Selection demand of user, reduce the complexity of screening.
Description
Technical field
The application relates to field of information processing, particularly relates to a kind of method of information navigation, equipment and system.
Background technology
In field of information processing, user is at client input inquiry word, wish to obtain the content relevant to the query word inputted, navigation server is after obtaining the query word that provides of client, identify the query intention of this user, provide the information relevant to the query word that user inputs to client, thus reduce the query context of user, the relevant information that user can be provided according to navigation server, finds required content as early as possible.
In prior art, navigation server provides the information relevant to the query word that user inputs to comprise following three kinds of modes to client:
Mode one, provide the information (hereinafter referred to as pure classification navigate) relevant to the query word that user inputs in pure classification mode, under pure classification navigate mode, according to the query word of user's input, provide the category information relevant to this query word.Described classification refers to the classification of commodity, have foreground category and backstage classification point.Foreground category is used for user interface (UI, UserInterface) and shows, backstage classification is used for merchandise control, and foreground category and background class object mapping relations are described by rule.The bibliography system of current main flow represents with tree structure, each parent order has multiple subcategory, but each subcategory only has a parent order, the increasing extent that represents of classification is little from top to bottom, wherein, the parent order (this parent order does not have parent order) of most higher level can be called one-level classification, and the subcategory (this subcategory does not have subcategory) of most subordinate can be called leaf classification.The navigation of pure classification be the earliest according to search class now subcategory commodity amount carry out classification recommendation, develop into subsequently adopt classification click accounting carry out classification fold, current pure classification navigation is mixed the information such as commodity amount, classification click and purchase and is carried out integrated navigation, and ways of presentation also changes into classification by the tiling of single classification and tiles with father and son's classification and deposit.
Mode two, to provide the information relevant to the query word that user inputs (navigating hereinafter referred to as sheerly property) in sheerly property mode, under sheerly property navigate mode, according to the query word of user's input, provide the attribute information relevant to this query word.Described attribute is for describing the characteristic of commodity, and attribute depends on leaf classification, namely only has leaf classification just can have attribute.A leaf classification can have multiple attribute, and an attribute can have multiple property value.Such as, brand, material, pattern, price etc. are the attribute that t sympathizes (leaf classification), and wherein, for brand generic, " A Yilian " brand is a property value of brand generic, and " Nike " brand is also a property value of brand generic.
Because attribute depends on leaf classification, therefore, only when the query word of user's input is the keyword of a leaf classification or user have input query word in client and have selected certain leaf classification, navigation server just can provide sheerly property navigate mode.Sheerly property navigation to represent form rich and varied, sheerly property navigation can be carried out with the ways of presentation of What You See Is What You Get, and user can carry out the operations such as attribute multiselect.
Mode three, provide the information (hereinafter referred to as classification attribute navigate) relevant to the query word that user inputs with classification and attribute mode, under classification attribute navigate mode, according to the query word of user's input, the category information (non-leaf classification) relevant to this query word is not only provided, the attribute information relevant to this query word is also provided.
Relative to the first and the second navigate mode, classification attribute navigate mode provides diversified relevant information, the classification that user not only can provide according to classification attribute navigate mode carries out Content Selection, and the attribute that also can provide according to classification attribute navigate mode carries out Content Selection.
The classification that classification attribute navigate mode provides is at least one classification relevant to query word, and depend on leaf classification due to attribute, the attribute that classification attribute navigate mode provides belongs to a leaf classification of certain classification at least one classification described, make user when selecting certain attribute to carry out Content Selection, make the context too small (only for a leaf classification of a classification) filtered out, the larger query context (at least one classification described) that query word that user provides is corresponding can not be embodied, the query intention of user can not be fully demonstrated, cause the uncomplete content face inquired, accuracy is lower.
Further, when classification attribute navigate mode provides attribute information, the click accounting determining leaf classification is needed, when the click accounting of a setting leaf classification reaches threshold value, there is provided the attribute information of this leaf classification to user, now, also there is the problem that threshold value is difficult to determine.When the threshold value of click accounting of the leaf classification set higher (be generally 85% and more than), this requirement can not be met by making the click of a lot of leaf classification, abundant attribute information cannot be provided, user cannot carry out Content Selection according to attribute information, if and the threshold value of the click accounting of the leaf classification of setting is lower, too much attribute information is provided by causing, cause the load of system heavier, there is provided the speed of relevant information comparatively slow, and user screen the complexity increase of content.
Summary of the invention
The embodiment of the present application provides a kind of method, equipment and system of information navigation, comprehensive for solving the attribute information that existing information navigation method provides, and is difficult to the problem of the click accounting threshold value determining leaf classification.
A method for information navigation, described method comprises:
The query word that navigation server determination client provides;
Navigation server extracts at least one public attribute corresponding to this query word, and the property value that each described public attribute is corresponding;
Property value corresponding to the public attribute extracted and this public attribute is pushed to client by navigation server.
A device for information navigation, described device comprises:
Determination module, for determining the query word that client provides;
First extraction module, for extracting at least one public attribute corresponding to this query word, and the property value that each described public attribute is corresponding;
Pushing module, for pushing to client by property value corresponding to the public attribute extracted and this public attribute.
A system for information navigation, described system comprises client and navigation server, wherein:
Client, provides query word for navigation server;
Navigation server, for extracting at least one public attribute corresponding to this query word, and the property value that each described public attribute is corresponding, and property value corresponding to the public attribute extracted and this public attribute is pushed to client.
According to the scheme that the embodiment of the present application provides, when carrying out information navigation, no longer be confined to leaf class attribute information now, but when determining the query word that client provides, determine at least one public attribute that this query word is corresponding, and the property value that each described public attribute is corresponding, and the property value of public attribute corresponding for this query word and each public attribute is pushed to client, thus solve prior art Problems existing, and meet the Content Selection demand of user, reduce the complexity of screening.
Accompanying drawing explanation
The flow chart of steps of the method for the information navigation that Fig. 1 provides for the embodiment of the present application one;
The flow chart of steps of the method for the information navigation that Fig. 2 provides for the embodiment of the present application two;
The flow chart of steps of the method for the information navigation that Fig. 3 provides for the embodiment of the present application three;
The structural representation of the classification tree that Fig. 4 provides for the embodiment of the present application three;
The structural representation of the device of the information navigation that Fig. 5 provides for the embodiment of the present application four;
The structural representation of the system of the information navigation that Fig. 6 provides for the embodiment of the present application five.
Embodiment
The classification attribute navigate mode that prior art provides launches certain class attribute now, just requires different to the click accounting launching classification, can not recommend inhomogeneity object public attribute.In addition, the query word (query) that classification attribute navigate mode covers is less, concentrates on the query of main fitting type, can not solve the demand of most attribute selection.
In the scheme that the embodiment of the present application provides, by attribute is carried, recommend relevant to query and across the set of properties of classification, several property values corresponding below each set of properties, and classification and attribute selection entrance can be provided for needing the wide in range class query of attribute selection function simultaneously.Because the attribute above carried is the publicly-owned attribute of each associated class object, meets the screening requirements of public users, shorten the searching route of user.
Below by Figure of description and each embodiment, the application's scheme is described.
Embodiment one,
The embodiment of the present application one provides a kind of method of information navigation, and the steps flow chart of the method as shown in Figure 1, comprising:
The query word that step 101, navigation server determination client provide.
When user needs to carry out content search, by the query word that client input is relevant to query contents, in this step, client navigation server provides this query word, makes navigation server can determine the query word that client provides.
Step 102, navigation server extract public attribute and property value corresponding to public attribute.
In this step, navigation server extracts at least one public attribute corresponding to this query word, and the property value that each described public attribute is corresponding.
Because background class object stability is higher relative to foreground category, therefore, in the present embodiment, the property value of public attribute and public attribute can be extracted according to backstage classification.
In this step, navigation server can be determined in setting duration, in the number of clicks for described query word, to the number of clicks of each backstage leaf classification; For each backstage leaf classification, determine that number of clicks exceedes the backstage leaf classification of threshold value (such as, described threshold settings is 80%) with the ratio for the number of clicks of described query word; The backstage leaf classification of threshold value is exceeded for each number of clicks and the ratio for the number of clicks of described query word, determine the attribute of this backstage leaf classification, certainly, now can also verify the legitimacy of the attribute determined further, and subsequent operation is performed to the attribute by legitimate verification;
According at least one in following two kinds of modes, the property value of public attribute and public attribute can be extracted:
Mode one, by getting the mode determination public attribute of common factor.
According to the mark of attribute, get the common factor of the attribute determined, such as, determine that 5 numbers of clicks exceed the backstage leaf classification of threshold value with the ratio for the number of clicks of described query word, be respectively backstage leaf classification 1, backstage leaf classification 2, backstage leaf classification 3, backstage leaf classification 4 and backstage leaf classification 5, the corresponding attribute-bit of backstage leaf classification 1 is PID1, PID2, PID3, 4 attributes of PID5, the corresponding attribute-bit of backstage leaf classification 2 is PID1, PID3, PID5, 4 attributes of PID7, the corresponding attribute-bit of backstage leaf classification 3 is PID1, PID2, PID5, PID9, 5 attributes of PID10, the corresponding attribute-bit of backstage leaf classification 4 is PID1, PID11, 3 attributes of PID13, the corresponding attribute-bit of backstage leaf classification 5 is PID1, PID15, PID16, PID17, PID18, 6 attributes of PID19, the common factor then getting attribute can obtain the attribute that attribute-bit is " PID1 ".
And can using each attribute in described common factor as the public attribute determined, using the property value of the property value of this attribute as this public attribute, certainly, also can carry out legitimate verification to each property value, and will the property value of property value as this public attribute of legitimate verification be passed through.
Mode two, by getting the mode determination public attribute of union.
Owing to determining the public attribute for query word in the present embodiment, this public attribute across classification, therefore, may cause all kinds ofly identifying different attributes now, there is identical meaning, therefore, in the present embodiment, a kind of mode determining public attribute according to attribute-name is additionally provided.
Concrete, the attribute with same alike result name can be determined, and the attribute with same alike result name is merged into a public attribute, and due to the property value of the attribute with same alike result name may not be identical, can using the property value of the described union with the property value of the attribute of same alike result name as this public attribute after merging, certainly, in mode two, also can carry out legitimate verification to each property value, and will the property value of property value as this public attribute of legitimate verification be passed through.
Such as, determine that 5 numbers of clicks exceed the backstage leaf classification of threshold value with the ratio for the number of clicks of described query word, be respectively backstage leaf classification 1, backstage leaf classification 2, backstage leaf classification 3, backstage leaf classification 4 and backstage leaf classification 5, wherein:
The corresponding attribute of backstage leaf classification 1 is called 4 attributes of PIDVID1, PIDVID2, PIDVID3, PIDVID5;
The corresponding attribute of backstage leaf classification 2 is called 4 attributes of PIDVID1, PIDVID3, PIDVID5, PIDVID7;
The corresponding attribute of backstage leaf classification 3 is called 5 attributes of PIDVID1, PIDVID2, PIDVID5, PIDVID9, PIDVID10;
The corresponding attribute of backstage leaf classification 4 is called 3 attributes of PIDVID1, PIDVID11, PIDVID13;
The corresponding attribute of backstage leaf classification 5 is called 6 attributes of PIDVID1, PIDVID15, PIDVID16, PIDVID17, PIDVID18, PIDVID19;
Then the attribute with same alike result name " PIDVID1 " can be merged into a public attribute, and can using the property value of the union of the property value of the attribute of attribute " PIDVID1 " by name corresponding respectively to backstage leaf classification 1, backstage leaf classification 2, backstage leaf classification 3, backstage leaf classification 4 and backstage leaf classification 5 as the public attribute after merging.
After pass-through mode one and/or mode two extract the property value of public attribute and public attribute, can also screen the public attribute extracted further:
Because user selects query contents according to foreground category, therefore, can further according to foreground category and background class object mapping ruler, determine that public attribute that pass-through mode one or mode two obtain (now, this public attribute can be considered as the first attribute) whether belong to the attribute of foreground leaf classification, and the first attribute only obtained at pass-through mode one or mode two is when belonging to the attribute of foreground leaf classification, just the first attribute that pass-through mode one or mode two obtain is defined as public attribute, thus facilitate subsequent user to understand this public attribute, otherwise, the first attribute that pass-through mode one or mode two obtain is not defined as public attribute.
Therefore, for each the first attribute determined, when determining that this first attribute belongs to the attribute of foreground leaf classification, this first attribute can be defined as a public attribute, certainly, now can using the property value of the property value of this first attribute as this public attribute.
The public attribute extracted and property value corresponding to this public attribute are pushed to client by step 103, navigation server.
In this step, property value corresponding to the public attribute extracted and this public attribute is pushed to client, can specifically comprise:
According in setting duration, in the number of clicks for described query word, to the number of clicks order from high to low of each public attribute, select the public attribute of rank top N, wherein, N is positive integer.Such as, can according within a week of setting, 50 points for described query word hit, it is 25 times to the click of public attribute 1 (attribute-bit is PID1), it is 20 times to the click of public attribute 2 (attribute-bit is PID2), be 5 times to the click of public attribute 3 (attribute-bit is PID3), be followed successively by according to the public attribute of number of clicks first 2 of select progressively rank from high to low: public attribute 1 and public attribute 2.Wherein, if public attribute is determined by the mode one in step 102, each backstage leaf class that then can exceed threshold value for number of clicks and the ratio for the number of clicks of described query word to the number of clicks of public attribute is now to the number of clicks sum of attribute of mark being designated this public attribute, if public attribute is determined by the mode two in step 102, each backstage leaf class that then can exceed threshold value for number of clicks and the ratio for the number of clicks of described query word to the number of clicks of public attribute is now to the number of clicks sum of attribute of attribute-name with this public attribute.
For each public attribute selected, when the property value of this public attribute is discrete values type, determine the number of clicks of each property value in the number of clicks to this public attribute, according to the number of clicks order from high to low of each property value, the property value of M position before selection rank, wherein, M is positive integer.Such as, for the public attribute selected being designated PID1,3 corresponding property values are respectively: PIDVID1, PIDVID3, PIDVID7, hitting 25 points of PID1,5 times, 12 times and 8 times are respectively to the number of clicks of PIDVID1, PIDVID3 and PIDVID7, are then followed successively by according to the property value of number of clicks first 3 of select progressively rank from high to low: PIDVID3, PIDVID7 and PIDVID1.When the property value of this public attribute is serial number type, the order descending or ascending according to property value arranges, such as, for the public attribute selected being designated PID2,6 corresponding property values are respectively: 39,38,37,36,35,34, then can obtain the property value after sorting is 39,38,37,36,35,34 or be 34,35,36,37,38,39.
By the public attribute of rank top N selected with for each public attribute, the property value of M position before the rank selected, or the property value after descending or ascending order arrangement pushes to client, concrete, can according to public attribute 1, PIDVID3, PIDVID7, PIDVID1; Public attribute 2,39,38,37,36,35, the form of 34, pushes to client by the public attribute selected (first 2 of rank) and the property value of this public attribute.
When carrying out the display of navigation information (public attribute extracted and property value corresponding to public attribute), can by a hurdle show navigator information, each data that one hurdle display is corresponding with query word, navigation information and each data corresponding with query word are divided two hurdle displays, thus make display interface clearly simple and clear, be convenient to user and check display interface.
Preferably, after step 102, before step 103, before property value corresponding to the public attribute extracted and this public attribute is pushed to client, the property value selected in advance can also be determined further, concrete, the property value selected in advance can be determined by the following method:
Step 103 ', determine the property value selected in advance.
Property value corresponding for each public attribute extracted is carried out text matches with described query word or synonym mates, using the property value that matches with described query word as the property value selected in advance.
Then now, in step 103, property value corresponding to the public attribute extracted and this public attribute is pushed to client by navigation server, specifically comprises: the described property value selected in advance is preferentially pushed to client.
Further, in step 103 ' after, before step 103, can also determine according to the property value selected in advance the public attribute selected in advance further, concrete, the public attribute selected in advance can be determined by the following method:
Step 103 ", determine the public attribute selected in advance.
For each property value selected in advance, according to for described query word, this property value selected in advance belongs to the forecast power of each public attribute, and (described forecast power can be obtained by existing method, such as, can according to comprising in the data of " N97 " 100 data headers, the brand generic of 90 data is had to comprise " Nokia ", the forecast power of the brand generic of property value " Nokia " corresponding for query word " N97 " is set as 90%), determine the public attribute that property value that this is selected in advance is corresponding, concrete, can determine that the property value that this is selected in advance belongs to a forecast power the highest in the forecast power of each public attribute, and the public attribute that this is corresponding with forecast power is as the public attribute determined, and can using the public attribute determined as the public attribute selected in advance corresponding to this property value selected in advance.
Then now, in step 103, property value corresponding to the public attribute extracted and this public attribute is pushed to client by navigation server, specifically comprises: the public attribute selected in advance corresponding with the property value that this is selected in advance for each property value selected in advance is preferentially pushed to client.
Certainly, the present embodiment can further include step 101 ', step 101 ' after step 101, can perform before step 103, be not limited to as shown in Figure 1 after step 101, perform before step 102:
Step 101 ', navigation server determines and the category information of this query word degree of correlation higher than setting value.
Determine with this query word degree of correlation same as the prior art higher than the method for the category information of setting value, such as, for in the number of clicks of this query word, classification number of clicks being exceeded set point number is defined as and the classification of this query word degree of correlation higher than setting value, and can determine the category information such as category name, classification mark that this classification is corresponding.
Then now, in step 103, the public attribute extracted and property value corresponding to this public attribute are pushed to client by navigation server, can specifically comprise: property value corresponding to the public attribute extracted, this public attribute and the described category information with this query word degree of correlation higher than setting value are pushed to client by navigation server.Thus not only provide the public attribute information relevant to query word to user, also provide the category information relevant to this query word further, make subsequent user not only can carry out content (data) screening according to attribute information, Content Selection can also be carried out according to category information, improve the precision that user carries out Content Selection further.
If when user needs to carry out content search, not by means of only the query word that client input is relevant to query contents, also provide at least one non-leaf classification corresponding with this query word further by client, then after step 101, before step 102, also comprise step 102 further ', step 102 ' be not limited to as shown in Figure 1, in step 101 ' after, perform before step 102:
Step 102 ', at least one non-leaf classification of providing of navigation server determination client.
Certainly at least one non-leaf classification described also can not be provided by client, but at least one non-leaf classification that this query word doped according to described query word by navigation server is corresponding.
Then now, in a step 102, navigation server extracts at least one public attribute corresponding to this query word, and the property value that each described public attribute is corresponding, specifically comprises:
For each non-leaf classification, determine each attribute in the leaf classification that this non-leaf classification is corresponding, concrete, the each legal attribute (effective attribute) in the leaf classification that this non-leaf classification is corresponding can be determined, and the second attribute that this non-leaf classification is corresponding can be determined, described second attribute can be obtain by merging the attribute with same alike result mark, and the property value of each second attribute is the property value of the attribute for merging into this second attribute.For each second attribute, can determine the data number with this second attribute, this data number is the data number sum that the attribute for merging into this second attribute is corresponding.
Further, the first ratio of classification total quantity under the quantity of the leaf classification with this second attribute and this non-leaf classification can be determined, and determine to have the second ratio of the data number of this second attribute and this non-leaf class data total number now, when described first ratio is not less than the first setting value and described second ratio is not less than the second setting value, this second attribute is defined as the public attribute extracted, and property value corresponding for this second attribute is defined as property value corresponding to this public attribute, or, determine to have the second ratio of the data number of this second attribute and this non-leaf class data total number now, when described second ratio is not less than the 3rd setting value, this second attribute is defined as the public attribute extracted, and property value corresponding for this second attribute is defined as property value corresponding to this public attribute, concrete, also legal property value corresponding for this second attribute can be defined as property value corresponding to this public attribute.
In step 103, property value corresponding to the public attribute extracted and this public attribute is pushed to client by navigation server, can specifically comprise:
For each public attribute extracted, determine to set in the number of clicks for described query word in duration, to the number of clicks of this public attribute, determine the number of clicks to this public attribute and the ratio for the number of clicks of described query word, and when this ratio is not less than setting threshold value, property value corresponding to this public attribute and this public attribute is pushed to client.
Below by two concrete examples, the scheme of the extraction public attribute related in the embodiment of the present application one is described in detail.
Embodiment two,
The embodiment of the present application two provides a kind of method of information navigation, determines the situation of the query word that client provides for navigation server, is mainly described the process extracting public attribute.The steps flow chart of the method as shown in Figure 2, specifically comprises the following steps:
Step 201, determine backstage leaf classification click distribution.
Concrete, in this step, can obtain the number of clicks of each backstage leaf classification according to clickstream data daily record (for recording the click situation of each data), thus the point that can obtain for query word hits, the click distribution in the leaf classification of backstage.Such as, the mark of each backstage leaf classification represents with cat, and the number of clicks of each backstage leaf classification represents with n, and query word query represents, then can obtain following expression:
querycat1：n1；cat2：n2；......
Step 202, determine attribute click distribution.
Concrete, in this step, can obtain the number of clicks of each backstage leaf class each attribute now according to navigation click logs (for recording the click situation of navigation information), thus the point that can obtain for query word hits, the click distribution in each attribute.Such as, the mark of each attribute represents with pid, and the number of clicks of each backstage leaf classification represents with m, and query word query represents, then can obtain following expression:
querypid1：m1；pid2：m2；......
Step 203, extraction property value belong to the forecast power of each attribute.
In this step, the forecast power that the property value obtained in advance belongs to each attribute can be extracted, determine for follow-up the forecast power that property value belongs to each public attribute.
It should be noted that, in the present embodiment, the execution of step 201, step 202 and step 203 in no particular order.
The property value that step 204, extraction public attribute and public attribute are corresponding.
In this step, concrete, can on the basis of step 201, the backstage leaf classification of threshold value is exceeded for each number of clicks and the ratio for the number of clicks of described query word, determine the attribute of backstage leaf classification, and can according to the mode one in step 102 in embodiment one, according to the mark of attribute, get the common factor of the attribute determined, and can using each attribute in described common factor as first attribute determined.
Now, described first attribute can as the public attribute extracted, but conveniently the follow-up of user is checked, further, the first attribute that Land use systems one is determined is screened, concrete, can determine whether the first attribute that pass-through mode one is determined belongs to the attribute of foreground leaf classification, first attribute of the attribute belonging to foreground leaf classification is defined as the 3rd attribute, and the property value of this first attribute can be defined as the property value of the 3rd attribute.
Now, described 3rd attribute can as the public attribute extracted, but comprehensive in order to the public attribute that ensures to determine, in the present embodiment, further, the mode two in embodiment one in step 102 can also be utilized to continue to determine public attribute.Concrete, the attribute of leaf classification corresponding to the backstage leaf classification of threshold value is exceeded for each number of clicks and the ratio of the number of clicks for described query word, the attribute with same alike result name can be merged into a public attribute, and can using the described property value of union as this public attribute after merging with the property value of the attribute of same alike result name.
Step 205, determine the property value selected in advance.
Concrete, can will can realize with described query word the property value that property value that text matches or synonym mate is defined as selecting in advance.
Step 206, determine the public attribute selected in advance.
In this step, can on the basis of step 203, for each property value selected in advance, the forecast power of each public attribute is belonged to according to this property value selected in advance, concrete, if public attribute is determined by the mode one in step 102, then this property value selected in advance belongs to the forecast power of each public attribute and is, number of clicks exceedes each backstage leaf class of threshold value now with the ratio for the number of clicks of described query word, this property value selected in advance belongs to the forecast power sum of the mark attribute identical with the mark of this public attribute, if public attribute is determined by the mode two in step 102, then this property value selected in advance belongs to the forecast power of each public attribute and is, number of clicks exceedes each backstage leaf class of threshold value now with the ratio for the number of clicks of described query word, this property value selected in advance belongs to the forecast power sum that attribute is called the attribute of the attribute-name of this public attribute.Determine the public attribute that property value that this is selected in advance is corresponding, using the public attribute determined as the public attribute selected in advance corresponding to this property value selected in advance.
Step 207, navigation server push each public attribute and property value corresponding to this public attribute.
Concrete, in this step, the public attribute selected in advance corresponding with the property value that this is selected in advance for each property value selected in advance preferentially can be pushed to client.
And can on the basis of step 202, except the public attribute selected in advance that each property value selected in advance is corresponding with the property value that this is selected in advance, by the public attribute of rank top N selected with for each public attribute, the property value of M position before the rank selected, or the property value after descending or ascending order arrangement pushes to client.
Embodiment three,
The embodiment of the present application three provides a kind of method of information navigation, determine the situation of query word that client provides and at least one non-leaf classification (this non-leaf classification is foreground non-leaf classification) for navigation server, mainly the process extracting public attribute is described.The non-leaf classification provided for user is below described.The steps flow chart of the method as shown in Figure 3, specifically comprises the following steps:
Step 301, determine corresponding data number.
In this step, can according to foreground category list, determine each attribute in the foreground leaf classification that this non-leaf classification is corresponding, and foreground category path field that can be corresponding according to every bar data, for each attribute, determine the data number with this attribute, concrete, the data number that the attribute-bit of this attribute is corresponding can be determined, and can, for each classification, determine to belong to such destination data number.
Step 302, structure classification tree.
In this step, according to foreground category list, can build classification tree, the classification tree of structure can as shown in Figure 4, and wherein, women's dress classification represents described non-leaf classification, corresponding one-piece dress, T-shirt and the trousers of this non-leaf classification 3 leaf classifications.
And in the node can set at described classification, record each class data number now, and record the first attribute corresponding to each non-leaf classification, this first attribute can have same alike result mark now attribute by merging this non-leaf class obtains, and the property value of this first attribute is identical with the property value of the attribute for merging into this first attribute, and the data number that each first attribute is corresponding, this data number is the data number sum that the attribute with same alike result mark for merging into this first attribute is corresponding.
The property value that step 303, extraction public attribute and public attribute are corresponding.
In this step, according to the classification tree built, for this non-leaf class each first attribute now, determine whether this first attribute is public attribute.Can for this non-leaf class each first attribute now, determine that the leaf classification quantity with this first attribute is (concrete, can for determining the leaf classification quantity that the attribute-bit of this first attribute is corresponding) with the first ratio of classification total quantity under this non-leaf classification, such as, in classification tree as shown in Figure 4, for women's dress class " cotton " now first attribute, when the leaf classification quantity with this first attribute is 2, because classification total quantity under women's dress classification is 3, then the first ratio that " cotton " first attribute is corresponding is 0.5.
In this step, can also determine to have the second ratio of the data number of this first attribute and this non-leaf class data total number now, concrete, can determine to have " cotton " the first data number of attribute, and determine the second ratio of this data number and women's dress class data total number now.
The first setting value (classification accounting threshold value can be not less than at the first ratio, can set and represent with α) and the second ratio is not less than the second setting value (can set value is 0.1) time, or be not less than the 3rd setting value (data number accounting threshold value at the second ratio, can set and represent with β) time, this first attribute is defined as the public attribute extracted, and property value corresponding for this first attribute is defined as property value corresponding to this public attribute.
Step 304, determine the property value selected in advance.
Concrete, can will can realize with described query word the property value that property value that text matches or synonym mate is defined as selecting in advance.
Step 305, determine the public attribute selected in advance.
For each property value selected in advance, the forecast power of each attribute can be belonged to according to the property value obtained in advance extracted, determine that property value that this is selected in advance belongs to the forecast power of each public attribute, concrete, the forecast power that this property value selected in advance belongs to each public attribute is, in this non-leaf class now, this property value selected in advance belongs to the forecast power sum of the mark attribute identical with the mark of this public attribute.
Determine the public attribute that property value that this is selected in advance is corresponding, using the public attribute determined as the public attribute selected in advance corresponding to this property value selected in advance.
The public attribute extracted and property value corresponding to this public attribute are pushed to client by step 306, navigation server.
For each public attribute extracted, can determine to set in the number of clicks for described query word in duration, to the number of clicks of this public attribute, and determine the number of clicks to this public attribute and the ratio for the number of clicks of described query word, when this ratio is not less than setting threshold value, property value corresponding to this public attribute and this public attribute is pushed to client.Thus optionally can push this public attribute property value corresponding with this public attribute, the load of mitigation system.
According to the scheme that the embodiment of the present application one ~ embodiment three provides, not only attribute (public attribute) information across classification can be provided to user for query word, can also while the attribute information across classification be provided, the category information relevant to this query word is provided, thus the variation of the information provided is provided.When providing attribute information, the attribute selected in advance and the property value selected in advance preferentially can also be pushed to user, the screening complexity of further minimizing user, simplify the operating process of user, and the attribute and property value that set quantity can be provided according to the degree of correlation with query word, thus while mitigation system load, the attribute information providing correlativity higher is to user.In addition, while user provides query word, when further providing category information, in conjunction with this category information determination attribute information, thus the precision of the attribute information determined can be improved, the attribute information that further refinement is determined.
With the embodiment of the present application one ~ embodiment three based on same inventive concept, provide following device and system.
Embodiment four,
The embodiment of the present application four provides a kind of device of information navigation, and the structure of this device as shown in Figure 5, comprising:
Determination module 11 is for determining the query word that client provides; First extraction module 12 is for extracting at least one public attribute corresponding to this query word, and the property value that each described public attribute is corresponding; Pushing module 13 is for pushing to client by property value corresponding to the public attribute extracted and this public attribute.
Described first extraction module 12 comprises:
First submodule 121 is for determining that in setting duration, number of clicks exceedes the backstage leaf classification of threshold value with the ratio for the number of clicks of described query word, determines the attribute of each backstage leaf classification determined;
Second submodule 122 is for the mark according to attribute, get the common factor of the attribute determined, using each attribute in described common factor as first attribute determined, using the property value of the property value of this attribute as this first attribute, and/or, determine the attribute with same alike result name, and the attribute with same alike result name is merged into first attribute, using the described property value of union as this first attribute after merging with the property value of the attribute of same alike result name;
3rd submodule 123 is for being defined as a public attribute by each the first attribute determined, or, for each the first attribute determined, when determining that this first attribute belongs to the attribute of foreground leaf classification, this first attribute is defined as a public attribute.
Described determination module 11 is also for determining at least one non-leaf classification that client provides;
Described first extraction module 12 can also comprise the 4th submodule 124 and the 5th submodule 125, wherein:
4th submodule 124 is for determining the second attribute that each non-leaf classification is corresponding, described second attribute is that the attribute having same alike result mark now by the leaf class corresponding to this non-leaf classification merges and obtains, and the property value of each second attribute is the property value of the attribute for merging into this second attribute;
5th submodule 125 is for for each second attribute, determine the first ratio of classification total quantity under the quantity of the leaf classification with this second attribute and this non-leaf classification, and determine to have the second ratio of the data number of this second attribute and this non-leaf class data total number now, when described first ratio is not less than the first setting value and described second ratio is not less than the second setting value, this second attribute is defined as the public attribute extracted, and property value corresponding for this second attribute is defined as property value corresponding to this public attribute, or, when described second ratio is not less than the 3rd setting value, this second attribute is defined as the public attribute extracted, and property value corresponding for this second attribute is defined as property value corresponding to this public attribute,
Described pushing module 13 is specifically for for each public attribute extracted, determine to set in the number of clicks for described query word in duration, to the number of clicks of this public attribute, and determine the number of clicks to this public attribute and the ratio for the number of clicks of described query word, when this ratio is not less than setting threshold value, property value corresponding to this public attribute and this public attribute is pushed to client.
Described device also comprises the second extraction module 14:
Second extraction module 14 is for determining and the category information of this query word degree of correlation higher than setting value;
What described pushing module 13 was determined specifically for property value corresponding to public attribute, this public attribute of being extracted by described first extraction unit and the second extraction unit pushes to client with this query word degree of correlation higher than the category information of setting value.
Described pushing module 13, specifically for according in setting duration, in the number of clicks for described query word, to the number of clicks order from high to low of each public attribute, selects the public attribute of rank top N; For each public attribute selected, when the property value of this public attribute is discrete values type, determine the number of clicks of each property value in the number of clicks to this public attribute, according to the number of clicks order from high to low of each property value, the property value of M position before selection rank, when the property value of this public attribute is serial number type, the order descending or ascending according to property value arranges; By the public attribute of rank top N selected with for each public attribute, the property value of M position before the rank selected, or the property value after descending or ascending order arrangement pushes to client, and wherein, M, N are positive integer.
Described device also comprises matching module 15 and the first preliminary election module 16, wherein:
Described matching module 15 is for carrying out text matches by property value corresponding for each public attribute extracted with described query word or synonym mates;
Described first preliminary election module 16 for property value that described matching module and described query word are matched as the property value selected in advance;
Described pushing module 13 preferentially pushes to client specifically for the property value described preliminary election module selected in advance.
Described device also comprises the second preliminary election module 17:
The property value of described second preliminary election module 17 for selecting in advance for each described first preliminary election module, according to for described query word, this property value selected in advance belongs to the forecast power of each public attribute, determine the public attribute that property value that this is selected in advance is corresponding, using the public attribute determined as the public attribute selected in advance corresponding to this property value selected in advance;
Described pushing module 13 is specifically for preferentially pushing to client by the public attribute selected in advance corresponding with the property value that this is selected in advance for each property value selected in advance.
Embodiment five,
The embodiment of the present application five provides a kind of system of information navigation, and the structure of this system as shown in Figure 6, comprises client 21 and navigation server 22, wherein:
Client 21 provides query word for navigation server;
Navigation server 22 is for extracting at least one public attribute corresponding to this query word, and the property value that each described public attribute is corresponding, and property value corresponding to the public attribute extracted and this public attribute is pushed to client.
Wherein, described navigation server 22 is the device of the information navigation in the embodiment of the present application two, can have identical module, and have corresponding function, do not repeat them here with this device.
In the scheme that each embodiment of the application provides, because daily record data amount is huge, whole scheme can utilize cloud computing platform to realize, and can provide real-time query service by Apache (apache) framework.
Obviously, those skilled in the art can carry out various change and modification to the application and not depart from the spirit and scope of the application.Like this, if these amendments of the application and modification belong within the scope of the application's claim and equivalent technologies thereof, then the application is also intended to comprise these change and modification.
Claims (8)
1. a method for information navigation, is characterized in that, described method comprises:
The query word that navigation server determination client provides;
Navigation server extracts at least one public attribute corresponding to this query word, and the property value that each described public attribute is corresponding, wherein, when navigation server extracts public attribute corresponding to this query word, specifically comprise, navigation server determines that in setting duration, number of clicks exceedes the backstage leaf classification of threshold value with the ratio for the number of clicks of described query word; Determine the attribute of each backstage leaf classification determined; According to the mark of attribute, get the common factor of the attribute determined, using each attribute in described common factor as first attribute determined, using the property value of the property value of this attribute as this first attribute; And/or, determine the attribute with same alike result name, and the attribute with same alike result name is merged into first attribute, using the described property value of union as this first attribute after merging with the property value of the attribute of same alike result name; Each the first attribute determined is defined as a public attribute, or, for each the first attribute determined, when determining that this first attribute belongs to the attribute of foreground leaf classification, this first attribute is defined as a public attribute;
Property value corresponding to the public attribute extracted and this public attribute is pushed to client by navigation server.
2. the method for claim 1, it is characterized in that, after the query word that navigation server determination client provides, navigation server extracts at least one public attribute corresponding to this query word, and before property value corresponding to each described public attribute, described method also comprises:
At least one non-leaf classification that navigation server determination client provides;
Navigation server extracts at least one public attribute corresponding to this query word, and the property value that each described public attribute is corresponding, specifically comprises:
Determine the second attribute that each non-leaf classification is corresponding, described second attribute is that the attribute having same alike result mark now by the leaf class corresponding to this non-leaf classification merges and obtains, and the property value of each second attribute is the property value of the attribute for merging into this second attribute;
For each second attribute, determine the first ratio of classification total quantity under the quantity of the leaf classification with this second attribute and this non-leaf classification, and determine to have the second ratio of the data number of this second attribute and this non-leaf class data total number now, when described first ratio is not less than the first setting value and described second ratio is not less than the second setting value, this second attribute is defined as the public attribute extracted, and property value corresponding for this second attribute is defined as property value corresponding to this public attribute, or, when described second ratio is not less than the 3rd setting value, this second attribute is defined as the public attribute extracted, and property value corresponding for this second attribute is defined as property value corresponding to this public attribute,
Property value corresponding to the public attribute extracted and this public attribute is pushed to client by navigation server, specifically comprises:
For each public attribute extracted, determine to set in the number of clicks for described query word in duration, to the number of clicks of this public attribute;
Determine the number of clicks to this public attribute and the ratio for the number of clicks of described query word;
When this ratio is not less than setting threshold value, property value corresponding to this public attribute and this public attribute is pushed to client.
3. the method for claim 1, is characterized in that, after the query word that navigation server determination client provides, before property value corresponding to the public attribute extracted and this public attribute is pushed to client by navigation server, described method also comprises:
Navigation server is determined and the category information of this query word degree of correlation higher than setting value;
Property value corresponding to the public attribute extracted and this public attribute is pushed to client by navigation server, specifically comprises:
Property value corresponding to the public attribute extracted, this public attribute and the described category information with this query word degree of correlation higher than setting value are pushed to client by navigation server.
4. the method for claim 1, is characterized in that, property value corresponding to the public attribute extracted and this public attribute is pushed to client by navigation server, specifically comprises:
According in setting duration, in the number of clicks for described query word, to the number of clicks order from high to low of each public attribute, select the public attribute of rank top N;
For each public attribute selected, when the property value of this public attribute is discrete values type, determine the number of clicks of each property value in the number of clicks to this public attribute, according to the number of clicks order from high to low of each property value, the property value of M position before selection rank; When the property value of this public attribute is serial number type, the order descending or ascending according to property value arranges;
By the public attribute of rank top N selected with for each public attribute, the property value of M position before the rank selected, or the property value after descending or ascending order arrangement pushes to client;
Wherein, M, N are positive integer.
5. the method as described in as arbitrary in Claims 1 to 4, it is characterized in that, navigation server extracts at least one public attribute corresponding to this query word, and after property value corresponding to each described public attribute, before property value corresponding to the public attribute extracted and this public attribute is pushed to client by navigation server, described method also comprises:
Property value corresponding for each public attribute extracted is carried out text matches with described query word or synonym mates;
Using the property value that matches with described query word as the property value selected in advance;
Property value corresponding to the public attribute extracted and this public attribute is pushed to client by navigation server, specifically comprises:
The described property value selected in advance is preferentially pushed to client.
6. method as claimed in claim 5, it is characterized in that, using the property value that matches with described query word as after the property value selected in advance, before the public attribute extracted and property value corresponding to this public attribute are pushed to client by navigation server, described method also comprises:
For each property value selected in advance, according to for described query word, this property value selected in advance belongs to the forecast power of each public attribute, determines the public attribute that property value that this is selected in advance is corresponding;
Using the public attribute determined as the public attribute selected in advance corresponding to this property value selected in advance;
Property value corresponding to the public attribute extracted and this public attribute is pushed to client by navigation server, specifically comprises:
The public attribute selected in advance corresponding with the property value that this is selected in advance for each property value selected in advance is preferentially pushed to client.
7. a device for information navigation, is characterized in that, described device comprises:
Determination module, for determining the query word that client provides;
First extraction module, for extracting at least one public attribute corresponding to this query word, and the property value that each described public attribute is corresponding, wherein, first extraction module specifically comprises, the first submodule, for determining in setting duration, number of clicks exceedes the backstage leaf classification of threshold value with the ratio for the number of clicks of described query word, determines the attribute of each backstage leaf classification determined; Second submodule, for the mark according to attribute, get the common factor of the attribute determined, using each attribute in described common factor as first attribute determined, using the property value of the property value of this attribute as this first attribute, and/or, determine the attribute with same alike result name, and the attribute with same alike result name is merged into first attribute, using the described property value of union as this first attribute after merging with the property value of the attribute of same alike result name; 3rd submodule, for each the first attribute determined is defined as a public attribute, or, for each the first attribute determined, when determining that this first attribute belongs to the attribute of foreground leaf classification, this first attribute is defined as a public attribute;
Pushing module, for pushing to client by property value corresponding to the public attribute extracted and this public attribute.
8. a system for information navigation, is characterized in that, described system comprises client and navigation server, wherein:
Client, provides query word for navigation server;
Navigation server, for extracting at least one public attribute corresponding to this query word, and the property value that each described public attribute is corresponding, and property value corresponding to the public attribute extracted and this public attribute is pushed to client, wherein, when the public attribute that this query word of extraction is corresponding, specifically for, navigation server determines that in setting duration, number of clicks exceedes the backstage leaf classification of threshold value with the ratio for the number of clicks of described query word; Determine the attribute of each backstage leaf classification determined; According to the mark of attribute, get the common factor of the attribute determined, using each attribute in described common factor as first attribute determined, using the property value of the property value of this attribute as this first attribute; And/or, determine the attribute with same alike result name, and the attribute with same alike result name is merged into first attribute, using the described property value of union as this first attribute after merging with the property value of the attribute of same alike result name; Each the first attribute determined is defined as a public attribute, or, for each the first attribute determined, when determining that this first attribute belongs to the attribute of foreground leaf classification, this first attribute is defined as a public attribute.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110432357.3A CN103176995B (en) | 2011-12-21 | 2011-12-21 | A kind of method of information navigation, equipment and system |
| HK13109938.1A HK1182793A1 (en) | 2011-12-21 | 2013-08-26 | Method, device and system for information navigation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110432357.3A CN103176995B (en) | 2011-12-21 | 2011-12-21 | A kind of method of information navigation, equipment and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103176995A CN103176995A (en) | 2013-06-26 |
| CN103176995B true CN103176995B (en) | 2016-04-06 |
Family
ID=48636876
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201110432357.3A Active CN103176995B (en) | 2011-12-21 | 2011-12-21 | A kind of method of information navigation, equipment and system |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN103176995B (en) |
| HK (1) | HK1182793A1 (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104391977B (en) * | 2014-12-05 | 2018-04-03 | 北京国双科技有限公司 | Web Page Key Words frequency of occurrence detection method and device |
| CN106202090B (en) * | 2015-05-04 | 2020-02-07 | 阿里巴巴集团控股有限公司 | Information processing method, information searching method, information processing device, information searching device and server |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101582909A (en) * | 2008-05-16 | 2009-11-18 | 上海神图信息科技有限公司 | System and method for providing information service for movable terminal user |
| CN101615277A (en) * | 2008-06-26 | 2009-12-30 | 阿里巴巴集团控股有限公司 | A kind of method and apparatus of statistics |
| CN102053983A (en) * | 2009-11-02 | 2011-05-11 | 阿里巴巴集团控股有限公司 | Method, system and device for querying vertical search |
| CN102253936A (en) * | 2010-05-18 | 2011-11-23 | 阿里巴巴集团控股有限公司 | Method for recording access of user to merchandise information, search method and server |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7840448B2 (en) * | 2003-05-07 | 2010-11-23 | Cbs Interactive Inc. | System and method for automatically generating a narrative product summary |
| CN101770498A (en) * | 2009-01-05 | 2010-07-07 | 李铭 | Step searching method |
-
2011
- 2011-12-21 CN CN201110432357.3A patent/CN103176995B/en active Active
-
2013
- 2013-08-26 HK HK13109938.1A patent/HK1182793A1/en unknown
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101582909A (en) * | 2008-05-16 | 2009-11-18 | 上海神图信息科技有限公司 | System and method for providing information service for movable terminal user |
| CN101615277A (en) * | 2008-06-26 | 2009-12-30 | 阿里巴巴集团控股有限公司 | A kind of method and apparatus of statistics |
| CN102053983A (en) * | 2009-11-02 | 2011-05-11 | 阿里巴巴集团控股有限公司 | Method, system and device for querying vertical search |
| CN102253936A (en) * | 2010-05-18 | 2011-11-23 | 阿里巴巴集团控股有限公司 | Method for recording access of user to merchandise information, search method and server |
Also Published As
| Publication number | Publication date |
|---|---|
| HK1182793A1 (en) | 2013-12-06 |
| CN103176995A (en) | 2013-06-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106156127B (en) | Method and device for selecting data content to push to terminal | |
| CN103218719B (en) | A kind of e-commerce website air navigation aid and system | |
| US8935197B2 (en) | Systems and methods for facilitating open source intelligence gathering | |
| CN105930469A (en) | Hadoop-based individualized tourism recommendation system and method | |
| CN104077357B (en) | Collaborative filtering combined recommendation method based on user | |
| CN101593200A (en) | Chinese Web page classification method based on the keyword frequency analysis | |
| CN103390044B (en) | Method and device for identifying linkage type POI (Point Of Interest) data | |
| CN103559622A (en) | Characteristic-based collaborative filtering recommendation method | |
| CN102411754A (en) | Personalized recommendation method based on commodity property entropy | |
| CN106250513A (en) | A kind of event personalization sorting technique based on event modeling and system | |
| WO2011025696A1 (en) | Method and system of information matching in electronic commerce website | |
| CN103729359A (en) | Method and system for recommending search terms | |
| CN109165367B (en) | News recommendation method based on RSS subscription | |
| CN102375885A (en) | Method and device for providing search suggestions corresponding to query sequence | |
| CN103186550A (en) | Method and system for generating video-related video list | |
| CN103294815A (en) | Search engine device with various presentation modes based on classification of key words and searching method | |
| JP2013531289A (en) | Use of model information group in search | |
| CN104133868B (en) | A kind of strategy integrated for the classification of vertical reptile data | |
| CN105138508A (en) | Preference diffusion based context recommendation system | |
| CN103699603A (en) | Information recommendation method and system based on user behaviors | |
| CN104077392B (en) | Reminding method and device are suggested in a kind of search | |
| CN104516980B (en) | The output method and server system of search result | |
| CN103995905A (en) | Electronic commerce content multi-dimensional classification, navigation and skipping method | |
| CN103838754A (en) | Information searching device and method | |
| CN103778206A (en) | Method for providing network service resources |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: DE Ref document number: 1182793 Country of ref document: HK |
|
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: GR Ref document number: 1182793 Country of ref document: HK |