CN108021640B - Keyword expanding method and device based on associated application - Google Patents
Keyword expanding method and device based on associated application Download PDFInfo
- Publication number
- CN108021640B CN108021640B CN201711227933.4A CN201711227933A CN108021640B CN 108021640 B CN108021640 B CN 108021640B CN 201711227933 A CN201711227933 A CN 201711227933A CN 108021640 B CN108021640 B CN 108021640B
- Authority
- CN
- China
- Prior art keywords
- app
- keyword
- level
- words
- covering
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention relates to keyword expanding methods and device based on associated application.The described method includes: obtain the first order keyword of APP to be expanded covering, according to each first order keyword search to APP obtain associated second level APP；The second level keyword for obtaining each second level APP covering, according to each second level keyword search to APP obtain the associated third level APP of APP to be expanded；A candidate key set of words is obtained according to the keyword that each third level APP is covered；It determines similarity of each acquisition third level APP relative to second level APP, obtains specific gravity shared by each keyword in candidate key set of words；Calculate the similarity score of each keyword in candidate key set of words；Candidate key set of words is screened according to the similarity score, obtains the association keyword of APP to be expanded.The present invention can expand out the relevant keyword of APP automatically, both realize volume production, while ensure that expansion quality.
Description
Technical field
The present invention relates to data analysis technique fields, more particularly to keyword expanding method and dress based on associated application
It sets.
Background technique
With the rapid development of intelligent terminal, the development of mobile Internet Software Industry has been driven.More and more users exist
Application library platform (i.e. application shop) in intelligent terminal downloads various APP (application is also referred to as applied), according to dimension
Base encyclopaedia data show that 65% user passes through application needed for application shop search downloading.So APP developer is to improve itself
APP needs to carry out the Optimization Work of application shop in the search quality of application shop.Its key job first is that carrying out APP's
Key word analysis is to optimize itself APP.
Currently, the specific industry knowledge background based on intelligent terminal application shop, the keyword of APP is expanded more by people
Work carries out judgement expansion, for manually expanding, expands quality and is affected by human subjective's human-subject test, therefore exist and close
Keyword expands the defect of the unstable quality of result.
Summary of the invention
Based on this, the present invention provides keyword expanding methods and device based on associated application, can overcome existing answer
The defect of unstable quality is expanded with program keyword.
Scheme provided in an embodiment of the present invention includes:
A kind of keyword expanding method based on associated application, comprising:
The first order keyword for obtaining APP covering to be expanded is searched for according to each first order keyword in application library platform
The APP arrived obtains the associated second level APP of APP to be expanded；
The second level keyword for obtaining each second level APP covering is searched for according to each second level keyword in application library platform
The APP arrived obtains the associated third level APP of APP to be expanded；The keyword for obtaining each third level APP covering, according to each third level
The keyword of APP covering obtains a candidate key set of words；
It determines similarity of each third level APP relative to second level APP, obtains each keyword institute in candidate key set of words
The specific gravity accounted for；The similarity score of each keyword in candidate keywords set is calculated according to the similarity and the specific gravity；
Candidate key set of words is screened according to the similarity score, obtains the association keyword of APP to be expanded；
Wherein, the keyword of APP covering need to meet condition: include described in the corresponding search result of the keyword
APP。
A kind of keyword expanding device based on associated application, comprising:
Application extension module, for obtaining the first order keyword of APP covering to be expanded, according to each first order keyword
In the APP that application library platform searches, the associated second level APP of APP to be expanded is obtained；
Candidate word expands module, crucial according to each second level for obtaining the second level keyword of each second level APP covering
The APP that word is searched in application library platform obtains the associated third level APP of APP to be expanded；Obtain each third level APP covering
Keyword obtains a candidate key set of words according to the keyword that each third level APP is covered；
Similarity calculation module obtains candidate close for determining similarity of each third level APP relative to second level APP
Each keyword specific gravity shared in the candidate key set of words in keyword set；According to the similarity and the specific gravity
Calculate the similarity score of each keyword in candidate key set of words；
And key word screening module is obtained for being screened according to the similarity score to candidate key set of words
To the association keyword of APP to be expanded；
Wherein, the keyword of APP covering need to meet condition: include described in the corresponding search result of the keyword
APP。
A kind of computer readable storage medium, is stored thereon with computer program, realization when which is executed by processor
The step of method described above.
A kind of computer equipment can be run on a memory and on a processor including memory, processor and storage
The step of computer program, the processor realizes method described above when executing described program.
Implement above-described embodiment, after receiving wait expand APP, can determine the first order of APP covering to be expanded first
Keyword obtains second level APP (competing product APP) further according to the corresponding APP information of each first order keyword；Further, may be used
The second level keyword covered by each second level APP, the corresponding APP information of each second level keyword obtain APP to be expanded and close
The third level APP of connection；A candidate key set of words can be obtained further according to the keyword of each third level APP covering；Determine each third
Similarity of the grade APP relative to second level APP, in conjunction with each keyword specific gravity shared in the candidate key set of words, meter
Calculate the similarity score of each keyword in candidate key set of words；It finally can be according to the similarity score to candidate key word set
Conjunction is screened, and the association keyword above-mentioned technical proposal for obtaining APP to be expanded can be according to the APP for treating expansion, based on competing
Product APP realizes the expansion of keyword, can be improved the quality of keyword expansion.In addition, keyword through the foregoing embodiment
Expanding method, is also convenient for batch and exports the corresponding keyword of APP to be expanded opening up word scheme, realizes that efficiency also obtains larger mention
It rises；Both it realizes volume production, while ensuring that expansion quality.
Detailed description of the invention
Fig. 1 is the schematic flow chart of the keyword expanding method based on associated application of an embodiment；
Fig. 2 is that the APP level of the keyword expanding method based on associated application of an embodiment is schematic；
Fig. 3 is the schematic diagram of the keyword expanding device based on associated application of an embodiment.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.
The term " includes " of the embodiment of the present invention and " having " and their any deformations, it is intended that cover non-exclusive
Include.Such as contain series of steps or the process, method, system, product or equipment of (module) unit are not limited to
The step of listing or unit, but optionally further comprising the step of not listing or unit, or optionally further comprising for these
The intrinsic other step or units of process, method, product or equipment.
Referenced herein " multiple " refer to two or more."and/or", the association for describing affiliated partner are closed
System indicates may exist three kinds of relationships, for example, A and/or B, can indicate: individualism A exists simultaneously A and B, individualism
These three situations of B.Character "/" typicallys represent the relationship that forward-backward correlation object is a kind of "or".
Referenced herein " embodiment " is it is meant that a particular feature, structure, or characteristic described can wrap in conjunction with the embodiments
It is contained at least one embodiment of the application.Each position in the description occur the phrase might not each mean it is identical
Embodiment, nor the independent or alternative embodiment with other embodiments mutual exclusion.Those skilled in the art explicitly and
Implicitly understand, embodiment described herein can be combined with other embodiments.
Although the step in various embodiments of the present invention is arranged with label, it is not used to successive time that limits step
Sequence, based on the order of step or the execution of certain step need other steps unless expressly stated, the otherwise phase of step
Order is adjustable.
Fig. 1 is the schematic flow chart of the keyword expanding method based on associated application of an embodiment；As shown in Figure 1,
The keyword expanding method based on associated application in the present embodiment comprising steps of
S11 obtains the first order keyword of APP covering to be expanded, according to each first order keyword in application library platform
The APP searched obtains the associated second level APP of APP to be expanded.
Keyword in the embodiment of the present invention can be used for searching for the character of APP, such as the Chinese in application library platform including all
Word, English word or letter, number or other letter symbols, can also be the combining form of several characters.Described first
Grade keyword can be to be obtained by the historical search information for analyzing application library platform, comprising closing in the historical search information
The mapping relations of keyword and APP can also be preassigned based on experience value.
Wherein, the keyword of APP covering need to meet condition: include described in the corresponding search result of the keyword
APP.I.e. each first order keyword includes the APP to be expanded in the search result of application library platform.
S12 obtains the second level keyword of each second level APP covering, according to each second level keyword in application library platform
The APP searched obtains the associated third level APP of APP to be expanded.The keyword for obtaining each third level APP covering, according to each the
The keyword of three-level APP covering obtains a candidate key set of words.
Wherein, the second level keyword can be is obtained by the historical search information for analyzing application library platform,
But it is preassigned based on experience value.The second level keyword of one second level APP covering, need to meet condition: close the second level
Keyword includes second level APP in the search result of application library platform.
Wherein, the keyword of the third level APP covering can be the historical search information by analyzing application library platform
It obtains, can also be preassigned based on experience value.The keyword of one third level APP covering, need to meet condition: the pass
Keyword includes the third level APP in the search result of application library platform.
S13 determines similarity of each third level APP relative to second level APP, obtains each key in candidate key set of words
Word specific gravity shared in the candidate key set of words；Candidate keywords collection is calculated according to the similarity and the specific gravity
The similarity score of each keyword in conjunction.
Wherein, third level APP indicates third level APP with corresponding second level APP's relative to the similarity of second level APP
Synthesis Relational Grade.In one embodiment, if the corresponding second level APP of third level APP be one, obtain third level APP with it is right
The similarity of the second level APP answered, the similarity as the third level APP relative to second level APP；If APP pairs of the third level
The second level APP answered is two or more, then obtains the similarity of third level APP and each corresponding second level APP, respectively with this
Calculate similarity mean value, the similarity using the similarity mean value as the third level APP relative to second level APP.Wherein,
The similarity of the third level APP and single second level APP can be predetermined, and be also possible to searching based on application platform
What Suo Jilu was calculated in real time.The calculating similarity mean value both includes calculating absolute average, also includes calculating weighted average
Value.
Wherein, each keyword specific gravity shared in the candidate key set of words is based on keyword for the third level
What the different degree of APP determined, keyword characterizes the APP in the search result of the keyword for the different degree of an APP
Ranking information.Keyword can be the number of the pre- historical search record data for first passing through application library platform for the different degree of APP
According to the different degree that analysis obtains, it is also possible to preset different degree.It further include basis in one embodiment if the former
The historical search of application library platform records information, predefines the step of each keyword searches it different degree of APP.
S14 screens candidate key set of words according to the similarity score, and the association for obtaining APP to be expanded is closed
Keyword.
In one embodiment, the similarity score can be chosen from candidate key set of words ranking is preceding from high to low
The keyword for setting quantity, obtains the association keyword of APP to be expanded；Thus the association that APP to be expanded can be obtained in batches is crucial
Word.
It in another embodiment, can also be according to the sequence of the similarity score from high to low, from candidate key set of words
It is middle to choose the keyword phrase for setting number, it include multiple keywords in each keyword phrase, the association for obtaining APP to be expanded is closed
Keyword.The corresponding multiple keyword phrases of APP to be expanded can be obtained, convenient for exporting the association keyword of APP to be expanded in batches.
Keyword expanding method through the foregoing embodiment is can determine first after receiving wait expand APP wait expand
The first order keyword of APP covering, obtains second level APP (competing product further according to the corresponding APP information of each first order keyword
APP)；Further, the second level keyword that can be covered by each second level APP, the corresponding APP letter of each second level keyword
Breath, obtains the associated third level APP of APP to be expanded；A candidate key can be obtained further according to the keyword of each third level APP covering
Set of words；It determines similarity of each third level APP relative to second level APP, calculates candidate in conjunction with the specific gravity of each candidate keywords
The similarity score of each keyword in keyword set；Candidate key set of words can finally be carried out according to the similarity score
Screening, the association keyword above-mentioned technical proposal for obtaining APP to be expanded can be based on competing product APP according to the APP for treating expansion
The expansion for realizing keyword can be improved the quality of keyword expansion.
In one embodiment, the process of the first order keyword of APP covering to be expanded is obtained can include: flat according to application library
The historical search record of platform obtains whole keywords of APP covering to be expanded；Treat the whole keywords progress for expanding APP covering
Screening anomaly obtains the first order keyword of APP covering to be expanded to delete abnormal keyword therein.Wherein, the exception
Keyword includes: searchable index exception, keyword search results data exception, APP ranking exception, number of characters in search result
The keyword of at least one of abnormal feature.
Searchable index is to carry out the accumulative of APP search in application library platform using the keyword according in setting statistical time
Number (volumes of searches), while considering to search for what the factors such as magnitude were calculated, both searchable index and volumes of searches are that forward direction is presented
Relationship, from empirically substantially estimating, the corresponding volumes of searches of searchable index is as follows:
Wherein, P is searchable index, and f (x) represents searchable index and the non-simple linear increase of volumes of searches both sides relation is closed
System.
Searchable index refers to that searchable index is less than setting numerical value extremely；Search result refers to keyword search extremely
The APP quantity arrived is less than setting quantity；Different degree refers to that APP ranking is more rearward in the search result of keyword extremely；
Number of words refers to that number of words is too short or too long extremely.
Correspondingly, the process of the second level keyword for obtaining each second level APP covering can include: flat according to application library
The historical search record of platform obtains whole keywords of each second level APP covering；To whole keywords of each second level APP covering
Screening anomaly is carried out, to delete abnormal keyword therein, obtains the second level keyword of the second level APP covering.
The process of the keyword for obtaining each third level APP covering can include: according to the historical search of application library platform
Record obtains whole keywords of each third level APP covering；Abnormal sieve is carried out to whole keywords of each third level APP covering
Choosing obtains the keyword of the third level APP covering to delete abnormal keyword therein.
The purpose of above-mentioned keyword filtration treatment is to carry out screening anomaly to keyword, such as keyword search results are too
Less, searchable index is too low, search rank rearward, number of words it is too short or it is too long etc. belong to keyword abnormal conditions, rejected, with
Interference of the abnormal data to subsequent expansion is prevented, the accuracy that keyword is expanded is improved.
In one embodiment, further include the steps that predefining keyword for the different degree of corresponding APP, specifically include:
According to the ranking information of APP in keyword search results, to keyword for the different degree assignment of APP:
V_2 (w)=(15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0.5)
V_3 (r)=(0,1,3,6,10,16,22,30,40,50,65,80,100,120,150,200, ∞)
wi=V_2 (w)t；V_3(r)t< rank≤V_3 (r)t+1
Wherein, [1,16] i ∈；V_2 (w) is different degree weight vectors；V_3 (r) is ranking interval vector；∞ indicates ranking
Positive infinity；Rank indicates the ranking of APP in search result；wiIndicate keyword kiTo the different degree of APP.For example, APP is being closed
Keyword kiSearch result in ranking be the 2nd, then keyword kiDifferent degree to the APP is wi=V_2 (w)2=14；V_3
(r)2< rank≤V_3 (r)3.Wherein, V_2 (w), V_3 (r) can be preset according to different application library platforms.
In an alternative embodiment, acquisition of information APP covering to be expanded is recorded according to the historical search of application library platform
It further include that the record information of the historical search to application library platform carries out pretreated step before keyword.Such as based on nearest
Information is recorded in the search that the application library platform occurs within one week, historical search record information includes the keyword letter for search
Breath and the corresponding search result information of each keyword.Such as nearest one week keyword search results, APP information (may include
The dimensions such as APPID, APP title, affiliated list), key word information (including keyword ID, keyword, searchable index, search knot
The dimensions such as fruit).
In an alternative embodiment, carrying out pretreated step to the historical search record information of application library platform can be wrapped
It includes:
Firstly, the historical search for obtaining application library platform in set period of time records information, remembered according to the historical search
Record information determines corresponding first mapping relations of each keyword；It include the corresponding APP letter of keyword in first mapping relations
The ranking information of breath and APP in the multiple search result of the keyword.Then, it is recorded and is believed according to the historical search
First mapping relations of multiple keywords in breath, determine corresponding second mapping relations of each APP；Second mapping relations
In include the corresponding keyword of APP, further include different degree of each keyword for the APP, the different degree is for indicating
Ranking information of the APP in the search result of the keyword, APP in the search result of keyword ranking more before, the key
Word is bigger for the different degree of the APP.Further, it is answered according to first mapping relations and the foundation of the second mapping relations
With the corresponding data mapping library of library platform.
Mapping library based on the data, it is described that acquisition of information APP to be expanded is recorded according to the historical search of application library platform
Corresponding first order keyword can include: the data mapping library is inquired, corresponding second mapping relations of APP to be expanded are obtained,
The weight of the corresponding first order keyword of APP to be expanded and the first order keyword is obtained according to second mapping relations
It spends.
Wherein, the APP information that each first order emphasis keyword is covered in application library platform is obtained can include: inquiry institute
Data mapping library is stated, corresponding first mapping relations of each first order emphasis keyword are obtained, according to first mapping relations
Obtain the APP information of each first order emphasis keyword covering.
In one embodiment, the APP searched according to each first order keyword in application library platform, obtains wait expand
The associated second level APP of APP, comprising:
According to historical search record in each first order keyword setting historical period in multiple search result, be somebody's turn to do
The frequency sequencing information of APP in the corresponding multiple search result of first order keyword；It obtains frequency sequence and arranges preceding setting number
The APP of amount, the APP information arrived as each first order keyword search.It is crucial according to whole first order keywords, each first order
The APP information that word searches obtains an APP matrix；The frequency of occurrence for counting each APP in the APP matrix, chooses the APP
The APP that frequency of occurrence is greater than or equal to the first setting frequency in matrix is used as the associated second level APP of APP to be expanded.
Refering to what is shown in Fig. 2, APP to be expanded is first order APP (i.e. APP(1)), the first order keyword of APP covering to be expanded
It is expressed as KW(1), first order keyword search to second level APP be expressed as APP(2), the keyword expression of second level APP covering
For KW(2), and so on.
In one embodiment, the APP searched according to each second level keyword in application library platform, obtains wait expand
The associated third level APP of APP, comprising:
According to historical search record in each second level keyword setting historical period in multiple search result, be somebody's turn to do
The frequency sequencing information of APP in the corresponding multiple search result of second level keyword；It obtains frequency sequence and arranges preceding setting number
The APP of amount, the APP information arrived as each second level keyword search；It is crucial according to whole second level keywords, each second level
The APP information that word searches obtains an APP matrix；The frequency of occurrence for counting each APP in the APP matrix, chooses the APP
Frequency of occurrence is greater than or equal to the APP of the second setting frequency as the associated third level APP of the second level APP in matrix.
Since the same keyword may be searched for repeatedly in (such as in one week) setting historical period, and search plain knot
Fruit changes with the variation of search time.Statistics is carried out to search result to summarize, and finally obtains keyword k0Corresponding APP collection
Close A (k0) and frequency ordering vector V (k0),
A(k0)=(appid1,appid2,…,appidn)
V(k0)=(count1,count2,…,countn)
Wherein k0Indicate keyword, countnIt indicates to use keyword k in setting historical period0There is appid in searchnIt is right
Cope with the frequency of app.Wherein, the frequency sequencing information of APP refers to that the frequency sorts in the corresponding multiple search result of keyword
Vector V (k0) described in the corresponding frequency of APP.
In one embodiment, after obtaining second level APP, obtain each second level APP covering second level keyword it
Before, further includes: APP to be expanded application list affiliated in application library platform is obtained, deletion belongs to difference with APP to be expanded and answers
With the second level APP of list.
Optionally, after obtaining third level APP, before the keyword for obtaining each third level APP covering, further includes: obtain
The application list for taking APP to be expanded affiliated in application library platform, deletes the third for belonging to different application list with APP to be expanded
Grade APP.Thus the accuracy of the associated AP P of APP can be improved, improve the accuracy that subsequent key word is expanded.
In one embodiment, each third level APP and the similarity of single second level APP be calculate in real time it is similar
Degree, specific calculating process include:
The feature vector of second level APP is obtained according to the second level keyword that second level APP is covered, according to each third level
The keyword of APP covering obtains the feature vector of each third level APP；By One-Hot coding to the feature vector of second level APP
And the feature vector of third level APP is handled, obtain second level APP sparse features vector and third level APP it is dilute
Dredge feature vector；According to the sparse features vector of second level APP and the sparse features vector of third level APP, each third is calculated
The similarity of grade APP and corresponding second level APP.Wherein, the sparse features vector of second level APP and third level APP's is sparse
The dimension of feature vector is equal, and meets condition: dV≤m+n；M indicates the dimension of the feature vector of second level APP, and n indicates the
The dimension of the feature vector of three-level APP, dVIndicate the dimension of the sparse features vector.
Such as: APP such as to be expanded is APP(1), it is assumed that its corresponding second level APP includes (APP(2) 1、APP(2) 2),
Middle second level APPAPP(2) 1The keyword of covering is (KW(2) 1, KW(2) 2, KW(2) 3), in this, as second level APPAPP(2) 1Spy
Vector is levied, feature vector dimension is 3；Second level APPAPP(2) 2The keyword of covering is (KW(2) 2, KW(2) 3, KW(2) 4,KW(2) 5),
In this, as second level APPAPP(2) 2Feature vector, feature vector dimension be 4.
Further, second level APPAPP(2) 1Corresponding third level APP includes (APP(3) 1, APP(3) 2, APP(3) 3)；Second
Grade APPAPP(2) 2Corresponding third level APP includes (APP(3) 3,APP(3) 4, APP(3) 5)；This makes it possible to obtain third level APP set
(APP(3) 1, APP(3) 2, APP(3) 3,APP(3) 4, APP(3) 5).In third level APP set, APP(3) 1Corresponding second level APP is only
There is APP(2) 1, therefore, APP(3) 1With similarity, that is, APP of second level APP(3) 1With APP(2) 1Similarity；APP(3) 3Corresponding
Second level APP has APP(2) 1And APP(2) 2, therefore, APP is obtained respectively(3) 3With APP(2) 1Similarity, APP(3) 3With APP(2) 2Phase
Like degree, similarity mean value is calculated with this, using the similarity mean value as APP(3) 3With the similarity of second level APP.
Further, third level APPAPP(3) 1The keyword of covering is (KW(3) 1, KW(3) 2, KW(3) 3), in this, as third
Grade APPAPP(3) 1Feature vector, feature vector dimension be 3；Third level APPAPP(3) 2The keyword of covering is (KW(3) 4, KW(3) 2, KW(3) 3, KW(3) 5), in this, as third level APPAPP(3) 2Feature vector, feature vector dimension be 4.Wherein, KW(3) 2With
KW(2) 2For same keyword.
Second level APPAPP as a result,(2) 1Feature vector (KW(2) 1, KW(2) 2, KW(2) 3), third level APPAPP(3) 1Feature
Vector (KW(3) 1, KW(3) 2, KW(3) 3), KW(3) 2With KW(2) 2For same keyword, thus the feature that constitutes in real number space of the two to
Amount is (KW(2) 1, KW(2) 2, KW(2) 3, KW(3) 1, KW(3) 3), dimension is 5≤3+3, and the sparse features vector for obtaining the two is respectively as follows:
Second level APPAPP(2) 1Sparse features vector: (1,1,1,0,0), third level APPAPP(3) 1Sparse features vector: (0,1,
0,1,1)。
Based on the above embodiment, optionally, the phase of each third level APP with single second level APP are calculated by the following formula
Like degree:
In formula, APP(2) tIndicate t-th of second level APP；S(3) iIndicate i-th of third level APP；V(APP(2) t)·V(S(3) i) indicate APP(2) tSparse features vector and S(3) iSparse features vector inner product；||V(APP(2) t)||2||V(S(3) i)|
|2Indicate APP(2) tSparse features vector and S(3) iSparse features vector 2- norm product.
It should be understood that between two APP similarity calculation method, it is including but not limited to above-mentioned similar based on cosine
Degree calculates the algorithm of similarity, can also be used to calculate the algorithm of similarity using other.
In one embodiment, a candidate key set of words is obtained according to the keyword that each third level APP is covered, comprising: root
According to the keyword that second level APP associated third level APP and each third level APP are covered, the associated pass second level APP is obtained
Keyword matrix.To the crucial conflation of words statistics in the keyword matrix, candidate key set of words KW is obtained(3)=(kw(3) 1, kw(3) 2..., kw(3) n) and corresponding keyword frequency vector be C(3)=(c1,c2,…,cn)。
Further, the specific gravity that each keyword is shared in the candidate key set of words in candidate key set of words is obtained
It can are as follows: determine candidate key set of words KW(3)In i-th of keyword shared by specific gravity are as follows:
In formula, i=1,2 ..., n, n indicate candidate key set of words KW(3)In include keyword sum.
In one embodiment, described that each key in candidate keywords set is calculated according to the similarity and the specific gravity
The similarity score of word, comprising: according to the corresponding third level APP phase of the specific gravity of keyword, keyword in candidate key set of words
For the product of the similarity of second level APP, the similarity score of keyword described in candidate key set of words is obtained.Concrete example
Such as: the similarity score of each keyword in candidate key set of words is calculated by following formula:
score(kw(3) i)=V(1) iV(2) i
Wherein, kw(3) iIndicate candidate key set of words KW(3)In i-th of keyword, V(1) iIndicate kw(3) iCorresponding
Similarity of the three-level APP relative to second level APP, V(2) iIndicate kw(3) iShared specific gravity；I=1,2 ..., n, n indicate candidate and close
Keyword set KW(3)In include keyword sum.
It should be understood that above-mentioned according to the specific gravity of keyword, the corresponding third of the keyword in candidate key set of words
Product of the grade APP relative to the similarity of second level APP, the similarity for obtaining keyword described in candidate key set of words obtain
Point, it can it is direct product, can also be multiplied by the product after proportionality coefficient.
Finally candidate key set of words is screened according to the similarity score, the association for obtaining APP to be expanded is closed
Keyword.Above-mentioned technical proposal can be realized the expansion of keyword based on competing product APP, can be improved according to the APP for treating expansion
The quality that keyword is expanded.In addition, keyword expanding method through the foregoing embodiment, is also convenient for exporting in batches to be expanded
The corresponding keyword of APP opens up word scheme, realizes that efficiency is also highly improved；Both it realizes volume production, while ensuring that expansion matter
Amount.
Below by taking apple application shop as an example, the keyword expanding method based on associated application of the embodiment of the present invention is done
Further instruction.In following embodiment, by taking apple application shop as an example, principle is identical therewith for other application library platform.It is described
Keyword expanding method based on associated application includes the following steps.
1. key words content grabs
The nearest historical search in one week of apple application shop is obtained using apple developer API and records data, including but not
Be limited to Apply Names, keyword details, keyword search index, keyword search results, using list etc..
2. history keyword word search record data prediction
The Direct mapping relationship of 2.1 keywords and APP, is denoted as A (k), indicates the search result of keyword k, under appid
The practical ranking of APP is searched in mark index expression with keyword k,
A (k)=(appid1,appid2,…,appidn) (1-1)
N is positive integer in formula.
It should be noted that APP can be identified by appid, and appid is unified by application library platform in the embodiment of the present invention
Distribution, for identifying different APP.
The reverse Mapping relationship of 2.2 app and keyword are denoted as K (a), indicate all keywords covered using a:
K (a)=(keyword1,...,keywordn) (1-2)
N is positive integer in formula.
3. obtaining competing product APP (i.e. associated second level APP)
3.1 remember that the appid of APP to be expanded is APP(1)；
3.2 obtain APP by K (a)(1)Keyword set K (the APP of covering(1)), the first order of APP covering to be expanded is closed
Keyword；
3.3 couples of keyword set K (APP(1)) carry out screening anomaly.Keyword search results are very little, searchable index is too low,
Search rank rearward, number of words it is too short or it is too long belong to data exception situation, rejected；
3.4 obtain keyword set K (APP by A (k)(1)) in each keyword correspond to appid, be denoted as A (K (APP(1)))；
3.5 couples of A (K (APP(1))) merger statistics is carried out, the appid of n before wherein frequency collating is taken, APP set S is denoted as(1)′；
3.6 reject APP set S(1)′In with APP(1)It is not belonging to the APP of same application list, finally only takes k conduct competing
Product APP is denoted as competing product APP set S(1), that is, the associated second level APP of APP to be expanded.
4.APP expands keyword
Remember S(1) iFor competing product APP set S(1)In i-th of APP, traverse competing product APP set S(1), steps are as follows:
4.1 obtain association appid
With preceding 5 steps in step 3, competing product are obtainedAssociated AP P, be denoted as third level APP, corresponding set
Use S(3)It indicates:
A(K(S(1) i))=(appid1,…,appidn) (4-1)
Further, the keyword matrix that (4-1) is covered can be obtained:
4.2 characteristic vector pickup.
By competing productFeature vector and S(3)In the feature vector that is covered of each APP (i.e. in (4-2) corresponding one
Row keyword) One-Hot coding is carried out, thus obtain competing productSparse features vector V (S(1) i) and S(3)In it is each
The sparse features vector V (S that APP is covered(3) i)。
Such as: there are vector A=(kw1,kw2,kw3) and B=(kw3,kw4,kw5), the two carries out One-Hot coding, then
Feature vector f=(kw on the two real number space R1,kw2,kw3,kw4,kw5) sparse features vector be respectively A '=(1,1,
1,0,0) with B '=(0,0,1,1,1).
4.3 calculate APP similarity.
Based on 4.2 as a result, calculating S(3)In each APP and S(1) iSimilarity, it is as follows:
In formula, S(3) jIndicate S(3)In j-th of third level APP；V(S(3) j)·V(S(1) i) indicate S(3) jSparse features
Vector and S(1) iSparse features vector inner product；||V(S(3) j)||2||V(S(1) i)||2Indicate S(3) jSparse features vector with
S(1) iSparse features vector 2- norm product.
To the crucial conflation of words statistics in (4-2), candidate key set of words KW is obtained(3)=(kw(3) 1,kw(3) 2,…,kw(3) n) and corresponding frequency vector be C(3)=(c1,c2,…,cn)；
The candidate key set of words KW(3)In i-th of keyword specific gravity are as follows:
In formula, i=1,2 ..., n, n indicate candidate key set of words KW(3)In include keyword sum.
4.4 calculate similarity of the candidate keywords relative to APP.
According to the specific gravity of the similarity of (4-3) and (4-4), candidate key set of words KW can be calculated(3)In each keyword phase
For the similarity score of APP to be expanded.
Finally, to candidate key set of words KW(3)Middle keyword carries out inverted order (from high to low) according to similarity score, takes
KW(3)In the first M expansion keyword as APP to be expanded, thus can keyword set W(1)。
In above-mentioned steps, 1~2 can be off-line calculation, regularly update, for example update one time again weekly.Step 3~4
It is corresponding appid to be obtained to each APP name query data mapping library of user's input, and then can in real time certainly in line computation
It is dynamic expand out the APP for keyword.
The technical application is expanded in Apple store APP association, 3 APP is tested and expands effect.It is manually first every
A APP has expanded 20 keywords, is then that each APP selects 100 keywords before similarity score automatically using the technology.
Comparing result discovery, the keyword 80% manually selected elected automatically, it was demonstrated that the validity of the technology.And compare people
Work is expanded, and the efficiency which obtains keyword gets a promotion.
It should be noted that for the various method embodiments described above, describing for simplicity, it is all expressed as a series of
Combination of actions, but those skilled in the art should understand that, the present invention is not limited by the sequence of acts described, because according to
According to the present invention, certain steps can use other sequences or carry out simultaneously.In addition, also any group can be carried out to above-described embodiment
It closes, obtains other embodiments.
Based on thought identical with the keyword expanding method based on associated application in above-described embodiment, the present invention is also mentioned
For the keyword expanding device based on associated application, which can be used for executing the above-mentioned keyword expansion side based on associated application
Method.For ease of description, in the structural schematic diagram of the keyword expanding device embodiment based on associated application, illustrate only with
The relevant part of the embodiment of the present invention, it will be understood by those skilled in the art that the restriction of schematic structure not structure twin installation, it can
To include perhaps combining certain components or different component layouts than illustrating more or fewer components.
Fig. 3 is the schematic diagram of the keyword expanding device based on associated application of one embodiment of the invention；Such as Fig. 3
Shown, the keyword expanding device based on associated application of the present embodiment includes:
Application extension module, for obtaining the first order keyword of APP covering to be expanded, according to each first order keyword
In the APP that application library platform searches, the associated second level APP of APP to be expanded is obtained；
Candidate word expands module, crucial according to each second level for obtaining the second level keyword of each second level APP covering
The APP that word is searched in application library platform obtains the associated third level APP of APP to be expanded；Obtain each third level APP covering
Keyword obtains a candidate key set of words according to the keyword that each third level APP is covered；
Similarity calculation module obtains candidate close for obtaining similarity of each third level APP relative to second level APP
Each keyword specific gravity shared in the candidate key set of words in keyword set；According to the similarity and the specific gravity
Calculate the similarity score of each keyword in candidate key set of words；
And key word screening module is obtained for being screened according to the similarity score to candidate key set of words
To the association keyword of APP to be expanded；
Wherein, the keyword of APP covering need to meet condition: include described in the corresponding search result of the keyword
APP。
It should be noted that in the embodiment of the keyword expanding device based on associated application of above-mentioned example, each mould
The contents such as information exchange, implementation procedure between block are brought due to being based on same design with preceding method embodiment of the present invention
Technical effect it is identical as preceding method embodiment of the present invention, for details, please refer to the description in the embodiment of the method for the present invention,
Details are not described herein again.
In addition, in the embodiment of the keyword expanding device based on associated application of above-mentioned example, each program module
Logical partitioning is merely illustrative of, and can according to need in practical application, such as the configuration requirement of corresponding hardware or soft
The convenient of the realization of part considers, above-mentioned function distribution is completed by different program modules, i.e., by described based on associated application
The internal structure of keyword expanding device is divided into different program modules, to complete all or part of function described above
Energy.
It will appreciated by the skilled person that realizing all or part of the process in above-described embodiment method, being can
It is completed with instructing relevant hardware by computer program, the program can be stored in a computer-readable storage and be situated between
In matter, sells or use as independent product.When being executed, the complete of the method such as the various embodiments described above can be performed in described program
Portion or part steps.Wherein, the storage medium can be magnetic disk, CD, read-only memory (Read-Only
Memory, ROM) or random access memory (Random Access Memory, RAM) etc..
Accordingly, a kind of storage medium is also provided in one embodiment, is stored thereon with computer program, wherein the journey
It realizes when sequence is executed by processor such as any one keyword expanding method based on associated application in the various embodiments described above.
In addition, the storage medium it is also settable with a kind of computer equipment in, further include place in the computer equipment
Manage device, when the processor executes the program in the storage medium, can be realized the method for the various embodiments described above whole or
Part steps.
Accordingly, a kind of computer equipment is also provided in one embodiment, which includes memory, processor
And store the computer program that can be run on a memory and on a processor, wherein processor is realized when executing described program
Keyword expanding method such as any one in the various embodiments described above based on associated application.
In the above-described embodiments, it all emphasizes particularly on different fields to the description of each embodiment, there is no the portion being described in detail in some embodiment
Point, it may refer to the associated description of other embodiments.It is appreciated that term " first order ", " second level " used in wherein etc.
Herein for distinguishing object, but these objects should not be limited by these terms.
The embodiments described above only express several embodiments of the present invention, should not be understood as to the invention patent range
Limitation.It should be pointed out that for those of ordinary skill in the art, without departing from the inventive concept of the premise,
Various modifications and improvements can be made, and these are all within the scope of protection of the present invention.Therefore, the scope of protection of the patent of the present invention
It should be determined by the appended claims.
Claims (12)
1. a kind of keyword expanding method based on associated application characterized by comprising
The first order keyword for obtaining APP covering to be expanded, searches according to each first order keyword in application library platform
APP obtains the associated second level APP of APP to be expanded；
The second level keyword for obtaining each second level APP covering, searches according to each second level keyword in application library platform
APP obtains the associated third level APP of APP to be expanded；The keyword for obtaining each third level APP covering, according to each third level APP
The keyword of covering obtains a candidate key set of words；
It determines similarity of each third level APP relative to second level APP, obtains in candidate key set of words shared by each keyword
Specific gravity；According to the similarity and the product of the specific gravity, the similarity for calculating each keyword in candidate key set of words is obtained
Point；Wherein, according in candidate key set of words in the keyword frequency and the candidate key set of words of i-th of keyword it is complete
The ratio of the sum of the keyword frequency of portion's keyword calculates specific gravity shared by i-th of keyword；
Candidate key set of words is screened according to the similarity score, obtains the association keyword of APP to be expanded；
Wherein, the keyword of APP covering need to meet condition: include the APP in the corresponding search result of the keyword.
2. the keyword expanding method according to claim 1 based on associated application, which is characterized in that described to obtain wait open up
Opening up the first order keyword that APP is covered includes:
Whole keywords of APP covering to be expanded are obtained according to the historical search of application library platform record；It treats and expands APP covering
Whole keywords carry out screening anomaly to delete abnormal keyword therein, the first order for obtaining APP covering to be expanded is crucial
Word；
And/or
The second level keyword for obtaining each second level APP covering, comprising:
Whole keywords of each second level APP covering are obtained according to the historical search of application library platform record；To each second level APP
Whole keywords of covering carry out screening anomaly to delete abnormal keyword therein, obtain the of the second level APP covering
Second level keyword；
And/or
The keyword for obtaining each third level APP covering, comprising:
Whole keywords of each third level APP covering are obtained according to the historical search of application library platform record；To each third level APP
Whole keywords of covering carry out screening anomaly to delete abnormal keyword therein, obtain the pass of the third level APP covering
Keyword；
The exception keyword includes: that searchable index exception, keyword search results data exception, APP are arranged in search result
The keyword of at least one of name is abnormal, number of characters is abnormal feature.
3. the keyword expanding method according to claim 2 based on associated application, which is characterized in that described according to each
The APP that first order keyword is searched in application library platform obtains the associated second level APP of APP to be expanded, comprising:
According to historical search record in each first order keyword setting historical period in multiple search result, obtain this first
The frequency sequencing information of APP in the corresponding multiple search result of grade keyword；Obtain the preceding setting quantity of frequency sequence row
APP, the APP arrived as each first order keyword search；
According to the APP information that whole first order keywords, each first order keyword search are arrived, an APP matrix is obtained；Statistics institute
The frequency of occurrence for stating each APP in APP matrix chooses frequency of occurrence in the APP matrix and is greater than or equal to the first setting frequency
APP is used as the associated second level APP of APP to be expanded；
And/or
The APP searched according to each second level keyword in application library platform, obtains the associated third level of APP to be expanded
APP, comprising:
According to historical search record in each second level keyword setting historical period in multiple search result, obtain this second
The frequency sequencing information of APP in the corresponding multiple search result of grade keyword；Obtain the preceding setting quantity of frequency sequence row
APP, the APP arrived as each second level keyword search；
According to the APP that whole second level keywords, each second level keyword search are arrived, an APP matrix is obtained；Described in statistics
The frequency of occurrence of each APP in APP matrix chooses the APP that frequency of occurrence in the APP matrix is greater than or equal to the second setting frequency
As the associated third level APP of the second level APP.
4. the keyword expanding method according to claim 3 based on associated application, which is characterized in that obtaining the second level
After APP, before the second level keyword for obtaining each second level APP covering, further includes:
APP to be expanded application list affiliated in application library platform is obtained, deletes and belongs to different application list with APP to be expanded
Second level APP；
And/or
After obtaining third level APP, before the keyword for obtaining each third level APP covering, further includes:
APP to be expanded application list affiliated in application library platform is obtained, deletes and belongs to different application list with APP to be expanded
Third level APP.
5. the keyword expanding method according to claim 1 based on associated application, which is characterized in that determine each third level
Similarity of the APP relative to second level APP, comprising:
If the corresponding second level APP of third level APP is one, it is similar with corresponding second level APP's to obtain third level APP
Degree, the similarity as the third level APP relative to second level APP；
If the corresponding second level APP of third level APP is two or more, third level APP and each corresponding second is obtained respectively
The similarity of grade APP, calculates similarity mean value with this, using the similarity mean value as the third level APP relative to second
The similarity of grade APP.
6. the keyword expanding method according to claim 5 based on associated application, which is characterized in that the acquisition third
The similarity of grade APP and corresponding second level APP, comprising:
The feature vector of second level APP is obtained according to the second level keyword that second level APP is covered, is covered according to each third level APP
The keyword of lid obtains the feature vector of each third level APP；
The feature vector of second level APP and the feature vector of third level APP are handled by One-Hot coding, obtained
The sparse features vector of second level APP and the sparse features vector of third level APP；
According to the sparse features vector of second level APP and the sparse features vector of third level APP, calculate each third level APP with
The similarity of corresponding second level APP.
7. the keyword expanding method according to claim 6 based on associated application, which is characterized in that
It is calculated by the following formula the similarity of each third level APP with corresponding second level APP:
In formula, APP(2) tIndicate t-th of second level APP；S(3) iIndicate i-th of third level APP；V(APP(2) t)·V(S(3) i) indicate
APP(2) tSparse features vector and S(3) iSparse features vector inner product；-||V(APP(2) t)||2||V(S(3) i)||2It indicates
APP(2) tSparse features vector and S(3) iSparse features vector 2- norm product.
8. the keyword expanding method according to claim 7 based on associated application, which is characterized in that according to each third level
The keyword of APP covering obtains a candidate key set of words, comprising:
A keyword matrix is obtained according to the keyword that second level APP associated third level APP and each third level APP are covered；
To the crucial conflation of words statistics in the keyword matrix, candidate key set of words KW is obtained(3)=(kw(3) 1,kw(3) 2,…,kw(3) n) and corresponding keyword frequency vector be C(3)=(c1,c2,…,cn), each element of the keyword frequency vector
Respectively correspond the frequency of occurrence of each keyword in candidate key set of words；
The candidate key set of words KW(3)In i-th of keyword shared by specific gravity are as follows:
In formula, i=1,2 ..., n, n indicate candidate key set of words KW(3)In include keyword sum.
9. the keyword expanding method according to any one of claims 1 to 8 based on associated application, which is characterized in that described
The similarity score of each keyword in candidate keywords set is calculated according to the similarity and the specific gravity, comprising:
According to the corresponding third level APP of the specific gravity of keyword, the keyword in candidate key set of words relative to second level APP's
The product of similarity obtains the similarity score of keyword described in candidate key set of words；
And/or
Candidate key set of words is screened according to the similarity score, obtains the association keyword of APP to be expanded, is wrapped
It includes:
The similarity score preceding keyword for setting quantity of ranking from high to low is chosen from candidate key set of words, is obtained
To the association keyword of APP to be expanded；Alternatively,
According to the sequence of the similarity score from high to low, the keyword word of setting number is chosen from candidate key set of words
Group includes multiple keywords in each keyword phrase, obtains the association keyword of APP to be expanded.
10. a kind of keyword expanding device based on associated application characterized by comprising
Application extension module is being answered for obtaining the first order keyword of APP covering to be expanded according to each first order keyword
The APP searched with library platform obtains the associated second level APP of APP to be expanded；
Candidate word expands module, for obtaining the second level keyword of each second level APP covering, is existed according to each second level keyword
The APP that application library platform searches obtains the associated third level APP of APP to be expanded；Obtain the key of each third level APP covering
Word obtains a candidate key set of words according to the keyword that each third level APP is covered；
Similarity calculation module obtains candidate keywords for determining similarity of each third level APP relative to second level APP
Each keyword specific gravity shared in the candidate key set of words in set；According to multiplying for the similarity and the specific gravity
Product calculates the similarity score of each keyword in candidate key set of words；Wherein, according to i-th of pass in candidate key set of words
The ratio of the sum of keyword frequency of whole keywords, calculates in the keyword frequency of keyword and the candidate key set of words
Specific gravity shared by i-th of keyword；
And key word screening module, for being screened according to the similarity score to candidate key set of words, obtain to
Expand the association keyword of APP；
Wherein, the keyword of APP covering need to meet condition: include the APP in the corresponding search result of the keyword.
11. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the program is by processor
The step of claim 1 to 9 any the method is realized when execution.
12. a kind of computer equipment including memory, processor and stores the meter that can be run on a memory and on a processor
Calculation machine program, which is characterized in that the processor realizes the step of claim 1 to 9 any the method when executing described program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711227933.4A CN108021640B (en) | 2017-11-29 | 2017-11-29 | Keyword expanding method and device based on associated application |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711227933.4A CN108021640B (en) | 2017-11-29 | 2017-11-29 | Keyword expanding method and device based on associated application |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108021640A CN108021640A (en) | 2018-05-11 |
| CN108021640B true CN108021640B (en) | 2019-08-16 |
Family
ID=62077455
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711227933.4A Active CN108021640B (en) | 2017-11-29 | 2017-11-29 | Keyword expanding method and device based on associated application |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108021640B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101246484A (en) * | 2007-02-15 | 2008-08-20 | 刘二中 | Electric text similarity processing method and system convenient for query |
| CN103488780A (en) * | 2013-09-27 | 2014-01-01 | 中国联合网络通信集团有限公司 | Application program searching method and device |
| CN103902720A (en) * | 2014-04-10 | 2014-07-02 | 北京博雅立方科技有限公司 | Method and device for acquiring expansion words of keywords |
| CN105426360A (en) * | 2015-11-12 | 2016-03-23 | 中国建设银行股份有限公司 | Keyword extracting method and device |
| CN106326300A (en) * | 2015-07-02 | 2017-01-11 | 富士通株式会社 | Information processing method and information processing device |
| CN106503224A (en) * | 2016-11-04 | 2017-03-15 | 维沃移动通信有限公司 | A kind of method and device for recommending application according to keyword |
-
2017
- 2017-11-29 CN CN201711227933.4A patent/CN108021640B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101246484A (en) * | 2007-02-15 | 2008-08-20 | 刘二中 | Electric text similarity processing method and system convenient for query |
| CN103488780A (en) * | 2013-09-27 | 2014-01-01 | 中国联合网络通信集团有限公司 | Application program searching method and device |
| CN103902720A (en) * | 2014-04-10 | 2014-07-02 | 北京博雅立方科技有限公司 | Method and device for acquiring expansion words of keywords |
| CN106326300A (en) * | 2015-07-02 | 2017-01-11 | 富士通株式会社 | Information processing method and information processing device |
| CN105426360A (en) * | 2015-11-12 | 2016-03-23 | 中国建设银行股份有限公司 | Keyword extracting method and device |
| CN106503224A (en) * | 2016-11-04 | 2017-03-15 | 维沃移动通信有限公司 | A kind of method and device for recommending application according to keyword |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108021640A (en) | 2018-05-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7853599B2 (en) | Feature selection for ranking | |
| CN105808590B (en) | Search engine implementation method, searching method and device | |
| CN106874292A (en) | Topic processing method and processing device | |
| CN103064903B (en) | Picture retrieval method and device | |
| CN110019889A (en) | Training characteristics extract model and calculate the method and relevant apparatus of picture and query word relative coefficient | |
| US20030212663A1 (en) | Neural network feedback for enhancing text search | |
| CN107918657A (en) | The matching process and device of a kind of data source | |
| CN109933660A (en) | The API information search method based on handout and Stack Overflow towards natural language form | |
| CN111061954B (en) | Search result sorting method and device and storage medium | |
| CN116455861B (en) | Big data-based computer network security monitoring system and method | |
| CN111078859B (en) | Author recommendation method based on reference times | |
| CN108182200B (en) | Keyword expansion method and device based on semantic similarity | |
| CN114330329A (en) | Service content searching method and device, electronic equipment and storage medium | |
| CN111160699A (en) | Expert recommendation method and system | |
| CN109344400A (en) | A kind of judgment method and device of document storage | |
| JP4891638B2 (en) | How to classify target data into categories | |
| CN108170664A (en) | Keyword expanding method and device based on emphasis keyword | |
| CN108021640B (en) | Keyword expanding method and device based on associated application | |
| CN108170665B (en) | Keyword expansion method and device based on comprehensive similarity | |
| CN108830302B (en) | Image classification method, training method, classification prediction method and related device | |
| CN114722086A (en) | Method and device for determining search rearrangement model | |
| CN108021641B (en) | The method and apparatus that the association keyword of application is expanded | |
| CN112015911B (en) | Method for searching massive knowledge maps | |
| Jalali et al. | A new clustering approach based on graph partitioning for navigation patterns mining | |
| CN116610853A (en) | Search recommendation method, search recommendation system, computer device, and storage medium |
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 |