CN107391584A - Facet searching method and system based on formal notion lattice - Google Patents
Facet searching method and system based on formal notion lattice Download PDFInfo
- Publication number
- CN107391584A CN107391584A CN201710483747.0A CN201710483747A CN107391584A CN 107391584 A CN107391584 A CN 107391584A CN 201710483747 A CN201710483747 A CN 201710483747A CN 107391584 A CN107391584 A CN 107391584A
- Authority
- CN
- China
- Prior art keywords
- concept
- lattice
- facet
- formal notion
- intension
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2457—Query processing with adaptation to user needs
- G06F16/24575—Query processing with adaptation to user needs using context
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention relates to computer search technical field, a kind of facet searching method and system based on formal notion lattice are disclosed, to be directed to raw information pre-structured formal notion lattice based on form concept analysis theory, and index is established on the basis of formal notion lattice, so as to establish facet search system.The inventive method includes：Structural form concept lattice, in construction process, formal notion lattice initialize minimum concept Bottom using attribute fake_attribute is forged；The index of leaf concept is established in formal notion lattice；In query process, the intension of concept corresponds to query statement, and the extension of concept corresponds to result set；After the facet value collection of user's inquiry is obtained, using bottom-up minimum intension of the matching comprising property set of leaf concept to find target concept corresponding with facet value collection in the formal notion lattice constructed, and the extension corresponding to the target concept is returned to.
Description
Technical field
The present invention relates to computer search technical field, more particularly to a kind of facet searching method based on formal notion lattice
And system.
Background technology
Facet search (Faceted Search) is a kind of on the basis of keyword search, is carried according to current search result
For context-sensitive facet information (Facet Information) information retrieval technique.User can be designed with detachment system
The set classification tree of personnel, from the self-defined classification interested of various angles.Some facet value is specified in user
After (Facet Value), dynamic access more refines in system result set according to corresponding to the facet value information, new refinement knot
Fruit can divide from multiple facets to result set, help user to further appreciate that their data messages interested.Whole
In individual search procedure, user can neatly switch facet value, so as to quick obtaining related content.
Form concept analysis (Formal Concept Analysis) theory is a kind of for structural data progress knowledge
The method with analysis is excavated, is widely used in the fields such as Knowledge Discovery, soft project.The core data knot of form concept analysis
Structure is formal notion lattice, and concept lattice represents the level knot between concept and concept by Hasse diagram (Hasse Diagram)
Structure.
Currently a popular facet search technique is primarily rested on traditional relevant database, quick-searching these
Content and to provide corresponding facet information be a problem urgently to be resolved hurrily.
The content of the invention
Present invention aims at a kind of facet searching method and system based on formal notion lattice is disclosed, by general based in the form of
Read analysis theories and be directed to raw information pre-structured formal notion lattice, and index is established on the basis of formal notion lattice, so as to build
Vertical facet search system.
To achieve the above object, the invention discloses a kind of facet searching method based on formal notion lattice, including：
Structural form concept lattice, the formal notion lattice initialize minimum concept using attribute fake_attribute is forged
Bottom, and in renewal concept lattice structure every time, input object Obj property set is added in Bottom intension, most
Fake_attribute rejected again afterwards to obtain complete and correct formal notion lattice；Meanwhile increase in formal notion lattice new
During object, after property set Y is inputted, the unique corresponding criterion generator of Y institutes is obtained first, if concept lattice is general
The intension of thought is equal with Y, then Y does not produce new ideas, and the new object is added to the extension of concept associated by its equal intension
In, and the new object is added to the outer Yanzhong of all father's concepts of this concept；If the interior of existing concept is not present in concept lattice
Culvert is equal with Y, then creates new ideas, and the direct father of all candidates of new ideas is calculated according to direct father's concept of criterion generator
Concept simultaneously filters out really direct father's concept, the then set membership more between new ideas, while the new object is added to
The outer Yanzhong of all father's concepts of new ideas；
The index of leaf concept is established in the formal notion lattice, the leaf concept refers to minimum in formal notion lattice
Direct father's concept of concept；
In query process, the intension of concept corresponds to query statement, and the extension of concept corresponds to result set；Looked into obtaining user
After the facet value collection of inquiry, the minimum of property set is included using the bottom-up matching of leaf concept in the formal notion lattice constructed
Intension returns to the extension corresponding to the target concept to find target concept corresponding with the facet value collection.
It is corresponding with the above method, invention additionally discloses a kind of facet search system based on formal notion lattice, including：
Subsystem one, for structural form concept lattice, the formal notion lattice use forgery attribute fake_attribute
Minimum concept Bottom is initialized, and in renewal concept lattice structure every time, input object Obj property set is added to
In Bottom intension, finally fake_attribute rejected again to obtain complete and correct formal notion lattice；Meanwhile in shape
During increasing new object in formula concept lattice, after property set Y is inputted, the unique corresponding criterion generator of Y institutes is obtained first,
If the intension of the existing concept of concept lattice is equal with Y, Y does not produce new ideas, and the new object is added into its equal intension
The outer Yanzhong of associated concept, and the new object is added to the outer Yanzhong of all father's concepts of this concept；If concept lattice is not
It is equal with Y in the presence of the intension of existing concept, then new ideas are created, and new ideas are calculated according to direct father's concept of criterion generator
The direct father's concept of all candidates and filter out really direct father concept, the then set membership more between new ideas, simultaneously
The new object is added to the outer Yanzhong of all father's concepts of new ideas；
Wherein, the subsystem one is additionally operable to establish the index of leaf concept, the leaf in the formal notion lattice
Concept refers to direct father's concept of minimum concept in formal notion lattice；
Subsystem two, in query process, the intension of concept to correspond to query statement, and the extension of concept corresponds to result
Collection；After the facet value collection of user's inquiry is obtained, the bottom-up matching of leaf concept is utilized in the formal notion lattice constructed
Minimum intension comprising property set is returned corresponding to the target concept with finding target concept corresponding with the facet value collection
Extension.
The invention has the advantages that：
On the one hand, during structural form concept lattice, minimum concept is dynamically updated, and avoids being loaded into advance completely
Formal Context, improve the flexibility of algorithm.
Another aspect, often adding an object can be all updated on the basis of original concept lattice based on criterion generator
To obtain new concept lattice, realization is simple, quick, reliable for operation, further, can also be by being carried out to the intension of concept
Cryptographic Hash is calculated as index, and all concepts are stored in Hash table, for it is determined that during criterion generator, with reference to
The property set Y calculated cryptographic Hash bottom-up search criterion generator to effectively improve the inquiry velocity of criterion generator,
Avoid redundant computation.
Querying condition is also parsed into attribute set by another further aspect, the present invention, is wrapped using the bottom-up matching of leaf concept
Minimum intension containing property set, impossible circuit can be quickly excluded, finds target concept.
Counted in advance before user is retrieved in addition, the present invention can realize in the facet search technique of formal notion lattice
The context relation and facet information of result set are calculated, or facet information corresponding to result is quickly calculated while retrieval,
System response time is shortened, Consumer's Experience is improved while reduces user and browse cost.
Below with reference to accompanying drawings, the present invention is further detailed explanation.
Brief description of the drawings
The accompanying drawing for forming the part of the application is used for providing a further understanding of the present invention, schematic reality of the invention
Apply example and its illustrate to be used to explain the present invention, do not form inappropriate limitation of the present invention.In the accompanying drawings：
A kind of exemplary plot of formal notion lattice disclosed in Fig. 1 embodiment of the present invention；
Facet searching method flow chart based on formal notion lattice disclosed in Fig. 2 embodiment of the present invention；
Fig. 3 is that a kind of formal notion lattice disclosed in the embodiment of the present invention makes process schematic.
Embodiment
The present invention is fully understood for ease of those skilled in the art, it is detailed to relational language of the present invention, theorem and inference
State as follows：
Define 1.1：If U is the set of object, M is the set of attribute, and I is the relation between two set U and M, then claims three
TupleFor a Formal Context (abbreviation background).(u, m) ∈ I (or writing uIm) represent that object u has attribute m.
Background can represent that its every a line is an object with the table of a rectangle, and each row are an attributes.If u rows m row
Infall is ×, then it represents that object u has attribute m.As table 1 provides example.
The exemplary context of table 1
Define 1.2：IfIt is a background, ifOrder
And
If A, B meet that f (A)=B, g (B)=A then claims two tuples (A, B) to be a concept (Concept).A is concept
The extension (Extent) of (A, B), B are concept (A, B) intensions (Intent).WithOrRepresent backgroundOn all concepts set.
Inference 1.1：For anyA is an extension and if only if A=g (f (A))；For anyB is
One intension and if only if B=f (g (B)).
Prove：(1) adequacy：For anyIn the presence of two tuples (A, f (A)), hence it is evident that there is f (A)=f (A), and because
For g (f (A))=A, so two tuples (A, f (A)) are a concept (defining 2.2), then A is an extension.
(2) necessity：Because A is an extension, two tuples (A, f (A)) are a concepts, then A=g (f (A)).
Here epitaxial part is proved, for intension proving by the same methods.
Theorem 2.1：(if U, M, I) is a background, A, A1, A2∈ U, B, B1, B2∈ M, then have：
First,
(1)
(2)
(3) f (A)=f (g (f (A)))
(4)f(A1)∩f(A2)=f (A1∪A2)
(5)
(6) g (f (A))=g (f (g (f (A))))
(7)
2nd,
(1`)
(2`)
(3`) g (B)=g (f (g (B)))
(4`)g(B1)∩g(B2)=g (B1UB2)
(5`)
(6`) f (g (B))=f (g (f (g (B))))
Theorem 2.2：Subset A, f (A) to any U must be intensions, because must be general shaped like (g (f (A)), f (A))
Read；It must be extension equally to have to any M subset B, g (B), because (g (B), f (g (B)) must be concepts.
Inference 2.2：The friendship of intension is still intension, and the friendship of extension is still extension.
Prove：For any two concept (A1, B1), (A2, B2), obtain B according to theorem 2.1 (4), 2.1 (4`)1∩B2=f
(A1)∩f(A2)=f (A1∪A2), according to theorem 2.2 forf(A1∪A2) must be intension, similarly there is A1∩
A2=g (B1)∩g(B2)=g (B1∪B2), easily demonstrate,prove g (B1∪B2) must be extension.Card is finished.
Define 2.3：If (A1, B1), (A2, B2) it is some backgroundOn two concepts.And
(by theorem 2.1 (1), it is apparent from being equivalent to), then we claim (A1, B1) it is (A2, B2) sub- concept, (A2, B2) it is (A1,
B1) father's concept, and be denoted as (A1, B1)≤(A2, B2).IfAnd concept (A is not present3, B3) causeThen claim (A1, B1) it is (A2, B2) direct sub- concept, (A2, B2) it is (A1, B1) direct father's concept, be denoted as (A1,
B1) ＜ (A2, B2).Thus a partially ordered set is obtainedWhereinRepresent all concepts of (U, M, I).
The partially ordered set obtained in above-mentioned definition 2.3It is referred to as formal notion lattice.Wherein make Then claim the Hasse diagram that figure (V, E) is formal notion lattice.According to Formal Context example
Table 1 can draw its corresponding form concept lattice such as Fig. 1.
Define 2.4：Direct sub- concept is that the concept of minimum concept in formal notion lattice is referred to as leaf in formal notion lattice
Concept.Leaf concept has and only minimum concept is as sub- concept.Such as in Fig. 1, concept (1, abef) and concept (4, abcdf)
It is leaf concept.
Define 2.5：In formal notion lattice, a certain attribute m support number is equal to the maximum concept that intension contains the attribute
Extension cardinal of the set；The support number of a certain concept is equal to the cardinal of the set of its extension.In Fig. 1, attribute f support number is
4, the support number of concept (1, abef) is 1.
Define 3.1：If L is corresponding form backgroundFormal notion lattice, in Formal ContextIn
New Formal Context is obtained after adding a new object oCorresponding formal notion lattice are L, wherein using f`
(o) attribute that o is possessed is represented.It is any one concept in L` to make (A, B), then
(1) (A, B) is the intension of a new ideas and if only if B is not any concept in L.
(2) (A, B) is that and if only if for improvement (Modified) concept
(3) (A, B) if existing in L, referred to as old concept.
(4) for any one new ideas (X, Y), if B ∩ f` (o)=Y ≠ B, it is (X, Y) to claim concept (A, B)
One maker.Minimum concept is the maker of other any concepts.
(5) generally, a new ideas (X, Y) has multiple makers (at least one maker), and it is got the bid
Quasi- maker is father's concept of other all makers, the supremum of as all makers.
Inference 3.1：For any new ideas (X, Y) one and only one criterion generator.Prove：
(1) if (X, Y) only has a maker, then it is criterion generator.
(2) if (X, Y) has multiple criterion generators, this is disagreed with definition, it is not necessary to is discussed.
(3) if there is no maker, then there is counterevidence.Assuming that there is concept (A two very big in these makers1, B1),
(A2, B2), then haveSoAccording to inference 2.2, B1∩B2Must be an intension, then one
Surely concept (g (B be present1∩B2), B1∩B2) and (X, Y) a maker because
This and (A1, B1), (A2, B2) it is that very big concept contradicts.Therefore the hypothesis is invalid.
To sum up, (X, Y) one and only one criterion generator.
Criterion generator is a critically important concept, during algorithm construction concept lattice, it is necessary first to is found new
The criterion generator of concept (X, Y).Actually criterion generator is the direct sub- concept for being currently generated new ideas in concept lattice.That
New ideas are come from what againIt is understood that a new object o is added in a form known concept lattice L to be obtained
To new formal notion lattice L`, it is the factor for causing L to update to illustrate o.To this just like drawing a conclusion：
Inference 3.2：New object o is added in concept lattice L according to defining 3.1, then is had：
(1) the property set f` (o) corresponding to object o must be an intension in new ideas lattice L`.If f` (o) is not
The intension of any concept in L, then (g` (f` (o)), f` (o)) is a new ideas；Otherwise, there is no new ideas in L`.
(2) if generating new ideas (A, B), in L, all intensions and B non-NULL, which occur simultaneously, is also possible to produce new connotation
(inference 2.2), thus produce more new ideas.
(3) the direct sub- concept of new ideas is its criterion generator.
(4) direct father's concept of new ideas is probably direct father's concept of criterion generator, or new ideas are given birth to standard
The concept corresponding to direct father's concept connotation common factor grown up to be a useful person.
(5) object o addition is bound to cause to improve the institute of the generation, specifically (g` (f` (o)), f` (o)) of concept
There is father's concept to be required for object o adding its extension.
To sum up, Formal Context is represented by triple (U, M, I), and wherein U represents the set of object, and M represents attribute set, I
Represent the relation between object and attribute.Also there is the concept of object and attribute in facet search, and attribute is carried out by facet
Division, this is facet search and the common point of form concept analysis.If representing the set of facet with F, FV represents facet pair
All facet values answered, then two tuples (F, FV) can be used to represent property set M.The new table of Formal Context is thus obtained
Up to formula (U, (F, FV), I) or four-tuple (U, F, M, I).It is general that corresponding form can be obtained according to Formal Context (U, F, M, I)
Read latticeWhereinRepresent all concepts.Newly-increased facet enriches the semanteme of attribute, together
When do not influenceed for the original structure of formal notion lattice.Facet search has good with form concept analysis in logical concept
Compatible degree, all operations that form concept analysis can be searched for facet, including semantic related operation provide support.
According to defined above, theorem and inference, embodiments of the invention are described in detail below in conjunction with accompanying drawing, still
The multitude of different ways that the present invention can be defined by the claims and cover is implemented.
Embodiment 1
The present embodiment discloses a kind of facet searching method based on formal notion lattice, as shown in Fig. 2 including：
Step S1, structural form concept lattice, the formal notion lattice are initialized most using attribute fake_attribute is forged
Small concept Bottom, and in renewal concept lattice structure every time, input object Obj property set is added to Bottom intension
In, finally fake_attribute rejected again to obtain complete and correct formal notion lattice；Meanwhile increase in formal notion lattice
During adding new object, after property set Y is inputted, the unique corresponding criterion generator of Y institutes is obtained first, if concept lattice is
Have that the intension of concept is equal with Y, then Y does not produce new ideas, and the new object is added into concept associated by its equal intension
Outer Yanzhong, and the new object is added to the outer Yanzhong of all father's concepts of this concept；If existing concept is not present in concept lattice
Intension it is equal with Y, then create new ideas, and all candidates for calculating according to direct father's concept of criterion generator new ideas are straight
Connect father's concept and filter out really direct father's concept, the then set membership more between new ideas, while the new object is added
It is added to the outer Yanzhong of all father's concepts of new ideas.
In this step, formal notion lattice can use static constructions pattern, dynamic construction pattern or pre- based on build time
The hybrid structural model surveyed；Wherein, hybrid structural model is to predict that conceptual construct takes according to facet value collection, if less than
Default time threshold, then using dynamic mode structural form concept lattice, if being more than default time threshold, using static mould
Formula structural form concept lattice.
Relative with above-mentioned steps, existing minimum concept needs to know set of properties all in Formal Context (U, M, I)
Into set M, it is necessary to be pre-loaded into whole background, cause idle background data to waste limited memory source, data volume
The excessively huge algorithm that results even in can not normal operation.
In order to facilitate expression construction process, the present embodiment use [object number, object properties collection] represent object entity, two
Tuple (extension, intension) represents concept, and shorthand represents set, such as abc={ a, b, c }, 123={ 1,2,3 }.Such as Fig. 3 institutes
Show, illustrate the process that object [3, cdf] is added into the concept lattice for having been added to object [1, abef] and object [2, bcf],
Eliminate the process of extension renewal.Filled circles represent the minimum concept (forging attribute to be omitted) of pre-structured, broken circle table in figure
Show the concept of neotectonics, solid line circle represents old concept in concept lattice.Fig. 3 (a) represents not adding the concept lattice knot of object [3, cdf]
Structure.Fig. 3 (b) represents that adding object [3, cdf] generates new ideas (3, cdf), and its corresponding criterion generator is in renewal
Minimum concept after culvertDirect father's concept of new ideas is not calculated also now.Fig. 3 (c) represents to be generated according to standard
The direct father's concept of a candidate that direct father's concept (1, abef) of device calculates concept (3, cdf) is new ideas (123, f), and
The direct sub- concept of (123, f) is its criterion generator (12, bf).Fig. 3 (d) represents basisDirect father's concept
The direct father's concept of another candidate that (2, bcf) calculate (3, cdf) is new ideas (23, cf).To note here is that (23,
Cf criterion generator (2, bcf)) also has other direct father's concepts (12, bf), so it is new ideas to also need to according to (12, cf)
(23, cf) calculate its direct father's concept, precisely already present (123, f), therefore the step does not have new ideas generation.Finally exist
Select really direct father's concept (23, cf) in the direct father's concept of candidate of (3, cdf), set membership between more new ideas, generally
Read lattice construction complete.
Optionally, the present embodiment realizes the false code such as table 2 below of above-mentioned object addition with improved AddIntent algorithms：
Table 2：
As shown in table 2, function AddIntent has three parameters：A maker corresponding to attribute set Y, Y
Canonical_generator, formal notion lattice L.After property set Y is inputted, it is right that the row of algorithm the 1st to the 4th row obtains Y institutes first
The criterion generator answered, if the intension of the existing concepts of concept lattice L is equal to Y, Y will not produce new ideas, directly return and find
Concept, algorithm terminates；Otherwise, the row of algorithm the 5th to the 14th further calculates new general according to direct father's concept of criterion generator
The direct father's concept of all candidates read；15th row to the 16th row generates a new ideas；It is straight that 17th row to the 30th row excludes candidate
Undesirable concept in father's concept is connect, then deletes old association that may be present, new ideas and direct father's concept are set
Association；31st row to the 32nd row updates the set membership between new ideas and criterion generator and returns to new ideas, and algorithm terminates.
Wherein, the function GetCanonicalGenerator of calling is used for obtaining the corresponding criterion generators of Y.
Preferably, in step S1, it is used as index by carrying out cryptographic Hash calculating to the intension of concept, by all concepts
It is stored in Hash table；It is determined that during criterion generator, bottom-up searched with reference to the property set Y calculated cryptographic Hash
Rope criterion generator；Thereby, to effectively improve the inquiry velocity of criterion generator, redundant computation is avoided.Deteriorated as a kind of
Realize：Can also by bottom-up mode by the intension of property set Y and concept one by one compared with, until finding a certain concept
All father's concepts are not the makers that attribute Y corresponds to concept, then the concept is criterion generator.
Step S2, the index of leaf concept is established in formal notion lattice, leaf concept refers to minimum in formal notion lattice
Direct father's concept of concept.
Step S3, in query process, the intension of concept corresponds to query statement, and the extension of concept corresponds to result set；Obtaining
After the facet value collection for taking family inquiry, attribute is included using the bottom-up matching of leaf concept in the formal notion lattice constructed
The minimum intension of collection returns to the extension corresponding to the target concept to find target concept corresponding with facet value collection.
In this step, facet is searched through the mode of facet, carries out conclusion retrieval to things, is usually used in professional domain
Among vertical search engine.The working specification of facet search is not specified by, and typical facet search operation is divided into two parts, is closed
Search is searched for and be interactively refined to keyword.Keyword search is traditional information retrieval technique, and facet search technique is basic herein
On allow user to interact the exploration of formula, existing keyword search results are carried out with further refinement and is decomposed, is carried
The high accuracy rate of search result, eliminates most of unrelated results collection, it is hereby achieved that good Consumer's Experience.
Facet search specifically has following steps：
(1) result set on keyword is obtained after by certain keyword search, and calculates point of result set
Face information, shows simultaneously.
(2) user selects facet value interested further to refine result according to the result set and facet of displaying.
(3) correlated results collection is re-searched for.
So can interactive cyclic query obtain satisfied result or result set as untill empty until user.
Further, the present embodiment also includes：
After target concept is found, facet value corresponding with the target concept is calculated according to the context of formal notion lattice
Number is supported, the facet value supports number to refer to the number of object corresponding to facet value.Such as：There can be following point for NBA sportsman
Face：Name, age, height, body weight, sportsman's type and affiliated team, each facet include many attributes, such as facet ---
" sportsman's type " includes centre forward, power forward, small forward etc., it is assumed that has centre forward sportsman 150 in NBA alliances, then Ke Yiyong
(' sportsman's type ', ' centre forward ', 150) represent.
Further, the present embodiment also includes：
The inquiry record of user is recorded, maximum public father's concept and the minimum public son for returning to multiple historical query results are general
Read and recommend as inquiry；Or inquiry recommendation is carried out according to concept similarity.Wherein：Obtain maximum public father's concept is divided to two
Step：
(1) all historical querys, are asked to correspond to the common factor of the intension of concept；
(2), the result of (1) is used to be inquired about, the concept of matching is maximum public father's concept.
Similarly, obtaining the method for maximum public sub- concept includes：
(1) all historical querys, are asked to correspond to the union of the intension of concept；
(2), the result of (1) is used to be inquired about, the concept of matching is minimum public sub- concept.
Optionally, the present embodiment compares concept similarity and is defined as follows.
Define 4.1：For formal notion latticeMiddle any two concept A, B, represent outer using extent
Prolong, then Jaccard coefficients are：
Current concepts are set as C, Jaccard similarities are k, and extensive operation searches for similarity for current concepts and is not less than k
Maximum father's concept, Refinement operation be current concepts search for similarity be not less than k most boy concept.For extensive operation, calculate
Method accesses his father's concept from C in a manner of breadth first traversal, until all father's concepts of a certain father's concept are similar to C's
Degree is respectively less than k, and the concept is extensive target concept；For Refinement operation, algorithm is visited from C in a manner of breadth first traversal
Its sub- concept is asked, until all sub- concepts of a certain sub- concept and C similarity are respectively less than k, the concept is refinement target concept.
To sum up, the facet searching method based on formal notion lattice disclosed in the present embodiment, has the advantages that：
On the one hand, during structural form concept lattice, minimum concept is dynamically updated, and avoids being loaded into advance completely
Formal Context, improve the flexibility of algorithm.
Another aspect, often adding an object can be all updated on the basis of original concept lattice based on criterion generator
To obtain new concept lattice, realization is simple, quick, reliable for operation, further, can also be by being carried out to the intension of concept
Cryptographic Hash is calculated as index, and all concepts are stored in Hash table, for it is determined that during criterion generator, with reference to
The property set Y calculated cryptographic Hash bottom-up search criterion generator to effectively improve the inquiry velocity of criterion generator,
Avoid redundant computation.
Querying condition is also parsed into attribute set by another further aspect, the present invention, is wrapped using the bottom-up matching of leaf concept
Minimum intension containing property set, impossible circuit can be quickly excluded, finds target concept.
Counted in advance before user is retrieved in addition, the present invention can realize in the facet search technique of formal notion lattice
The context relation and facet information of result set are calculated, or facet information corresponding to result is quickly calculated while retrieval,
System response time is shortened, Consumer's Experience is improved while reduces user and browse cost.
Embodiment 2
Corresponding with above method embodiment, the present embodiment discloses a kind of facet search system based on formal notion lattice
System, including at least following subsystems one and subsystem two, it is preferable that can further include follow-up subsystem three and/or
Subsystem four.
Subsystem one, for structural form concept lattice, the formal notion lattice use forgery attribute fake_attribute
Minimum concept Bottom is initialized, and in renewal concept lattice structure every time, input object Obj property set is added to
In Bottom intension, finally fake_attribute rejected again to obtain complete and correct formal notion lattice；Meanwhile in shape
During increasing new object in formula concept lattice, after property set Y is inputted, the unique corresponding criterion generator of Y institutes is obtained first,
If the intension of the existing concept of concept lattice is equal with Y, Y does not produce new ideas, and the new object is added into its equal intension
The outer Yanzhong of associated concept, and the new object is added to the outer Yanzhong of all father's concepts of this concept；If concept lattice is not
It is equal with Y in the presence of the intension of existing concept, then new ideas are created, and new ideas are calculated according to direct father's concept of criterion generator
The direct father's concept of all candidates and filter out really direct father concept, the then set membership more between new ideas, simultaneously
The new object is added to the outer Yanzhong of all father's concepts of new ideas.
Wherein, above-mentioned subsystem one is additionally operable to establish the index of leaf concept, the leaf in the formal notion lattice
Concept refers to direct father's concept of minimum concept in formal notion lattice.Further, the subsystem one of the present embodiment is additionally operable to：It is logical
Cross and cryptographic Hash calculating is carried out to the intension of concept as index, all concepts are stored in Hash table；It is determined that standard generates
During device, with reference to the property set Y calculated cryptographic Hash bottom-up search criterion generator.
Subsystem two, in query process, the intension of concept to correspond to query statement, and the extension of concept corresponds to result
Collection；After the facet value collection of user's inquiry is obtained, the bottom-up matching of leaf concept is utilized in the formal notion lattice constructed
Minimum intension comprising property set is returned corresponding to the target concept with finding target concept corresponding with the facet value collection
Extension.
Subsystem three, for after target concept is found, being calculated according to the context of formal notion lattice general with the target
Facet value corresponding to thought supports number, and the facet value supports number to refer to the number of object corresponding to facet value.
Subsystem four, the inquiry for recording user record, and return to maximum public father's concept of multiple historical query results
Recommend with minimum public sub- concept as inquiry；Or inquiry recommendation is carried out according to concept similarity.
In the system, formal notion lattice are using static constructions pattern, dynamic construction pattern or based on build time prediction
Hybrid structural model；The hybrid structural model predicts that conceptual construct takes according to facet value collection, if less than default
Time threshold, then using dynamic mode structural form concept lattice, if being more than default time threshold, constructed using static schema
Formal notion lattice.
Similarly, the facet search system based on formal notion lattice disclosed in the present embodiment, has the advantages that：
On the one hand, during structural form concept lattice, minimum concept is dynamically updated, and avoids being loaded into advance completely
Formal Context, improve the flexibility of algorithm.
Another aspect, often adding an object can be all updated on the basis of original concept lattice based on criterion generator
To obtain new concept lattice, realization is simple, quick, reliable for operation, further, can also be by being carried out to the intension of concept
Cryptographic Hash is calculated as index, and all concepts are stored in Hash table, for it is determined that during criterion generator, with reference to
The property set Y calculated cryptographic Hash bottom-up search criterion generator to effectively improve the inquiry velocity of criterion generator,
Avoid redundant computation.
Querying condition is also parsed into attribute set by another further aspect, the present invention, is wrapped using the bottom-up matching of leaf concept
Minimum intension containing property set, impossible circuit can be quickly excluded, finds target concept.
Counted in advance before user is retrieved in addition, the present invention can realize in the facet search technique of formal notion lattice
The context relation and facet information of result set are calculated, or facet information corresponding to result is quickly calculated while retrieval,
System response time is shortened, Consumer's Experience is improved while reduces user and browse cost.
The preferred embodiments of the present invention are the foregoing is only, are not intended to limit the invention, for the skill of this area
For art personnel, the present invention can have various modifications and variations.Within the spirit and principles of the invention, that is made any repaiies
Change, equivalent substitution, improvement etc., should be included in the scope of the protection.
Claims (10)
- A kind of 1. facet searching method based on formal notion lattice, it is characterised in that including：Structural form concept lattice, the formal notion lattice initialize minimum concept using attribute fake_attribute is forged Bottom, and in renewal concept lattice structure every time, input object Obi property set is added in Bottom intension, most Fake_attribute rejected again afterwards to obtain complete and correct formal notion lattice；Meanwhile increase in formal notion lattice new During object, after property set Y is inputted, the unique corresponding criterion generator of Y institutes is obtained first, if concept lattice is general The intension of thought is equal with Y, then Y does not produce new ideas, and the new object is added to the extension of concept associated by its equal intension In, and the new object is added to the outer Yanzhong of all father's concepts of this concept；If the interior of existing concept is not present in concept lattice Culvert is equal with Y, then creates new ideas, and the direct father of all candidates of new ideas is calculated according to direct father's concept of criterion generator Concept simultaneously filters out really direct father's concept, the then set membership more between new ideas, while the new object is added to The outer Yanzhong of all father's concepts of new ideas；The index of leaf concept is established in the formal notion lattice, the leaf concept refers to minimum concept in formal notion lattice Direct father's concept；In query process, the intension of concept corresponds to query statement, and the extension of concept corresponds to result set；Obtaining user's inquiry After facet value collection, the minimum intension of property set is included using the bottom-up matching of leaf concept in the formal notion lattice constructed To find target concept corresponding with the facet value collection, and return to the extension corresponding to the target concept.
- 2. the facet searching method according to claim 1 based on formal notion lattice, it is characterised in that also include：It is used as index by carrying out cryptographic Hash calculating to the intension of concept, all concepts is stored in Hash table；It is determined that mark During quasi- maker, with reference to the property set Y calculated cryptographic Hash bottom-up search criterion generator.
- 3. the facet searching method according to claim 1 or 2 based on formal notion lattice, it is characterised in that the form Concept lattice is using static constructions pattern, dynamic construction pattern or the hybrid structural model predicted based on build time；It is described mixed Box-like structural model predicts that conceptual construct takes according to facet value collection, if less than default time threshold, then using dynamic analog Formula structural form concept lattice, if being more than default time threshold, using static schema structural form concept lattice.
- 4. the facet searching method according to claim 3 based on formal notion lattice, it is characterised in that also include：After target concept is found, facet value corresponding with the target concept is calculated according to the context of formal notion lattice and supported Number, the facet value support number to refer to the number of object corresponding to facet value.
- 5. the facet searching method according to claim 4 based on formal notion lattice, it is characterised in that also include：The inquiry record of user is recorded, the maximum public father's concept and minimum public sub- concept for returning to multiple historical query results are made Recommend for inquiry；Or inquiry recommendation is carried out according to concept similarity.
- A kind of 6. facet search system based on formal notion lattice, it is characterised in that including：Subsystem one, for structural form concept lattice, the formal notion lattice are initial using attribute fake_attribute is forged Change minimum concept Bottom, and in renewal concept lattice structure every time, input object Obi property set is added to Bottom's In intension, finally fake_attribute rejected again to obtain complete and correct formal notion lattice；Meanwhile in formal notion lattice During middle increase new object, after property set Y is inputted, the unique corresponding criterion generator of Y institutes is obtained first, if concept The intension of the existing concept of lattice is equal with Y, then Y does not produce new ideas, and the new object is added to associated by its equal intension generally The outer Yanzhong read, and the new object is added to the outer Yanzhong of all father's concepts of this concept；If concept lattice is not present existing The intension of concept is equal with Y, then creates new ideas, and all times of new ideas are calculated according to direct father's concept of criterion generator Select direct father's concept and filter out really direct father's concept, the then set membership more between new ideas, while this is new right Outer Yanzhong as being added to all father's concepts of new ideas；Wherein, the subsystem one is additionally operable to establish the index of leaf concept, the leaf concept in the formal notion lattice Refer to direct father's concept of minimum concept in formal notion lattice；Subsystem two, in query process, the intension of concept to correspond to query statement, and the extension of concept corresponds to result set； After the facet value collection for obtaining user's inquiry, category is included using the bottom-up matching of leaf concept in the formal notion lattice constructed Property collection minimum intension to find target concept corresponding with the facet value collection, and return outer corresponding to the target concept Prolong.
- 7. the facet search system according to claim 6 based on formal notion lattice, it is characterised in that the subsystem one It is additionally operable to：It is used as index by carrying out cryptographic Hash calculating to the intension of concept, all concepts is stored in Hash table；It is determined that During criterion generator, with reference to the property set Y calculated cryptographic Hash bottom-up search criterion generator.
- 8. the facet search system based on formal notion lattice according to claim 6 or 7, it is characterised in that the form Concept lattice is using static constructions pattern, dynamic construction pattern or the hybrid structural model predicted based on build time；It is described mixed Box-like structural model predicts that conceptual construct takes according to facet value collection, if less than default time threshold, then using dynamic analog Formula structural form concept lattice, if being more than default time threshold, using static schema structural form concept lattice.
- 9. the facet search system according to claim 8 based on formal notion lattice, it is characterised in that also include：Subsystem three, for after target concept is found, being calculated and the target concept pair according to the context of formal notion lattice The facet value answered supports number, and the facet value supports number to refer to the number of object corresponding to facet value.
- 10. the facet search system according to claim 9 based on formal notion lattice, it is characterised in that also include：Subsystem four, the inquiry for recording user record, and return to maximum public father's concept and most of multiple historical query results The sub- concept of mini-bus is recommended as inquiry；Or inquiry recommendation is carried out according to concept similarity.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710483747.0A CN107391584B (en) | 2017-06-22 | 2017-06-22 | Formal concept lattice-based facet search method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710483747.0A CN107391584B (en) | 2017-06-22 | 2017-06-22 | Formal concept lattice-based facet search method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107391584A true CN107391584A (en) | 2017-11-24 |
| CN107391584B CN107391584B (en) | 2020-12-11 |
Family
ID=60332700
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710483747.0A Active CN107391584B (en) | 2017-06-22 | 2017-06-22 | Formal concept lattice-based facet search method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107391584B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113094456A (en) * | 2021-04-09 | 2021-07-09 | 郑州大学 | Robot walking path generation method |
| CN114358062A (en) * | 2021-12-23 | 2022-04-15 | 河南大学 | Yellow river bank dam dangerous case identification method based on formal concept analysis |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101504746A (en) * | 2008-02-04 | 2009-08-12 | 株式会社普罗菲德 | Information processing apparatus, information processing method, and program |
| CN101595476A (en) * | 2006-08-31 | 2009-12-02 | 启创互联公司 | The system, the method and computer program that are used for the client definition information architecture |
| CN102495844A (en) * | 2011-11-15 | 2012-06-13 | 河海大学 | Improved GuTao method for creating user models |
| US20130198195A1 (en) * | 2012-01-30 | 2013-08-01 | Formcept Technologies and Solutions Pvt Ltd | System and method for identifying one or more resumes based on a search query using weighted formal concept analysis |
| CN103782250A (en) * | 2011-05-26 | 2014-05-07 | 电气银普股份有限公司 | Modularized control system to enable networked control and sensing of other devices |
| CN104036046A (en) * | 2014-07-02 | 2014-09-10 | 重庆大学 | Deep Web query interface pattern matching method based on attribute co-occurrence mode |
| JP5725623B2 (en) * | 2012-05-08 | 2015-05-27 | 日本電信電話株式会社 | Program analysis apparatus and method, and program |
| US20150156230A1 (en) * | 2013-11-29 | 2015-06-04 | Raghava Kalyanaraman | Method and system for social collaborative curation |
| CN105474166A (en) * | 2013-03-15 | 2016-04-06 | 先进元素科技公司 | Methods and systems for purposeful computing |
| CN106227805A (en) * | 2016-07-17 | 2016-12-14 | 河南理工大学 | A kind of term definition method and system theoretical based on form concept analysis |
-
2017
- 2017-06-22 CN CN201710483747.0A patent/CN107391584B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101595476A (en) * | 2006-08-31 | 2009-12-02 | 启创互联公司 | The system, the method and computer program that are used for the client definition information architecture |
| CN101504746A (en) * | 2008-02-04 | 2009-08-12 | 株式会社普罗菲德 | Information processing apparatus, information processing method, and program |
| CN103782250A (en) * | 2011-05-26 | 2014-05-07 | 电气银普股份有限公司 | Modularized control system to enable networked control and sensing of other devices |
| CN102495844A (en) * | 2011-11-15 | 2012-06-13 | 河海大学 | Improved GuTao method for creating user models |
| US20130198195A1 (en) * | 2012-01-30 | 2013-08-01 | Formcept Technologies and Solutions Pvt Ltd | System and method for identifying one or more resumes based on a search query using weighted formal concept analysis |
| JP5725623B2 (en) * | 2012-05-08 | 2015-05-27 | 日本電信電話株式会社 | Program analysis apparatus and method, and program |
| CN105474166A (en) * | 2013-03-15 | 2016-04-06 | 先进元素科技公司 | Methods and systems for purposeful computing |
| US20150156230A1 (en) * | 2013-11-29 | 2015-06-04 | Raghava Kalyanaraman | Method and system for social collaborative curation |
| CN104036046A (en) * | 2014-07-02 | 2014-09-10 | 重庆大学 | Deep Web query interface pattern matching method based on attribute co-occurrence mode |
| CN106227805A (en) * | 2016-07-17 | 2016-12-14 | 河南理工大学 | A kind of term definition method and system theoretical based on form concept analysis |
Non-Patent Citations (2)
| Title |
|---|
| LIGENG ZOU等: "An efficient algorithm for increasing the granularity levels of attributes in formal concept analysis", 《EXPERT SYSTEMS WITH APPLICATIONS》 * |
| 林春杰等: "近似概念格及其增量构造算法研究", 《计算机应用研究》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113094456A (en) * | 2021-04-09 | 2021-07-09 | 郑州大学 | Robot walking path generation method |
| CN113094456B (en) * | 2021-04-09 | 2022-09-13 | 郑州大学 | Robot walking path generation method |
| CN114358062A (en) * | 2021-12-23 | 2022-04-15 | 河南大学 | Yellow river bank dam dangerous case identification method based on formal concept analysis |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107391584B (en) | 2020-12-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105335519B (en) | Model generation method and device and recommendation method and device | |
| US7593932B2 (en) | Information data retrieval, where the data is organized in terms, documents and document corpora | |
| CN103927358B (en) | text search method and system | |
| CN105808590B (en) | Search engine implementation method, searching method and device | |
| JP5989665B2 (en) | System and method for analyzing and synthesizing complex knowledge expressions | |
| KR20150031234A (en) | Updating a search index used to facilitate application searches | |
| Ayouni et al. | Extracting compact and information lossless sets of fuzzy association rules | |
| US20170024659A1 (en) | Method for data searching by learning and generalizing relational concepts from a few positive examples | |
| CN107391584A (en) | Facet searching method and system based on formal notion lattice | |
| Getoor et al. | Learning probabilistic relational models with structural uncertainty | |
| CN115422432A (en) | Dynamic searchable encryption method for massive high-dimensional medical data | |
| CN113392332B (en) | Simplified visual analysis method for large-scale multi-element network data | |
| An et al. | Automatic generation of ontology from the deep web | |
| Liu et al. | Intuitive representation of decision trees using general rules and exceptions | |
| Lao | Efficient random walk inference with knowledge bases | |
| CN107229704A (en) | A kind of resource description framework querying method and system based on KSP algorithms | |
| Leite et al. | Relating ontologies with a fuzzy information model | |
| Zeng et al. | Construction of scenic spot knowledge graph based on ontology | |
| Strobin et al. | Linguistic summaries of graph datasets using ontologies: An application to Semantic Web | |
| Ouaftouh et al. | UPCAR: user profile clustering based approach for recommendation | |
| CN111506797A (en) | Method and system for solving why-not problem in direction perception SKQ | |
| Shi et al. | Transfer significant subgraphs across graph databases | |
| CN110874348A (en) | Privacy differential data retrieval method in mixed cloud environment | |
| Mukherjee et al. | Graph node embeddings using domain-aware biased random walks | |
| CN110147393A (en) | The entity resolution method in data-oriented space |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |