CN106156114A

CN106156114A - Patent retrieval method and device

Info

Publication number: CN106156114A
Application number: CN201510159502.3A
Authority: CN
Inventors: 朱欣昱; 程序; 刘琦; 郭瑛媚; 孔文娟
Original assignee: Beijing Zhongxian Electronic Technology Development Center
Current assignee: Beijing Zhongxian Electronic Technology Development Center
Priority date: 2015-04-03
Filing date: 2015-04-03
Publication date: 2016-11-23

Abstract

The invention discloses a kind of patent retrieval method and device.Wherein, patent retrieval method includes: receive the retrieval request of user；Determine the first core word in retrieval request；Determining the first associational word according to the first core word and target information table, wherein, in target information table, storage has and has the associational word of incidence relation with the first core word；Search condition is determined according to the first core word and the first associational word；According to search condition patent searching file.By the present invention, solve the prior art Patent retrieval inaccurate problem of result.

Description

Patent retrieval method and device

Technical field

The present invention relates to searching field, in particular to a kind of patent retrieval method and device.

Background technology

In recent years, along with the increase of domestic amount of the application for patent, the quantity of patent constantly expands, and user is to patent application Concrete condition is the most increasingly paid close attention to.In prior art, user, can by inputting search condition in patent search system Statistical analysis is carried out with the amount of the application for patent to each field, each province and city and each enterprise etc., but, it is different from other retrievals, Patent retrieval is more strict to the requirement of retrieval result, so in retrieving, to retrieving the selection of key element and assembling, There is very professional skill set requirements.Prior art mainly has following three kinds of patent retrieval modes, specific as follows:

The first is the conventional treatment mode of current all patent retrieval platforms.When retrieval, by search field as far as possible How to be supplied to client, client assemble voluntarily.The retrieval result of this kind of patent retrieval mode is controlled, but to user's Retrieval ability requires higher, easily causes crucial retrieval information and omits, there is the inaccurate problem of patent search result.

The second is common synonym retrieval or company code's retrieval.When retrieval, according to the input of user or The selection of user, the built-in vocabulary of use system carries out limited spread, finally uses the retrieval key element of extension to retrieve. The retrieval result of this kind of patent retrieval mode is controlled, but can only carry out very limited amount of extension with in retrieval key element, Existence cannot meet retrieval key element association's demand of Detachable and solve the problem that search field assembles, and causes patent retrieval Result is inaccurate.Such as, retrieving washing machine, user needs to associate cylinder, intelligent control panel, but above-mentioned inspection Rope mode cannot associate cylinder, intelligent control panel.

The third is common semantic retrieval.When retrieval, user can be simply input vocabulary to be retrieved or description Statement.System carries out semantic analysis, extraction retrieval key element automatically according to input content, uses the mode of semantic matches certainly Move and retrieve.This kind of retrieval mode, whole retrieving user is the most uncontrollable, and retrieves result and do not possess and change Enter space, there is the inaccurate problem of patent search result.

Retrieve the inaccurate problem of result for prior art Patent, the most not yet propose effective solution.

Summary of the invention

Embodiments provide a kind of patent retrieval method and device, to solve prior art Patent retrieval result Inaccurate problem.

An aspect according to embodiments of the present invention, it is provided that a kind of patent retrieval method.

Patent retrieval method according to the present invention includes: receive the retrieval request of user；Determine in described retrieval request First core word；Determining the first associational word according to described first core word and target information table, wherein, described target is believed In breath table, storage has and has the associational word of incidence relation with described first core word；According to described first core word and described First associational word determines search condition；And according to described search condition patent searching file.

Further, described retrieval request includes content to be retrieved, determines the first core word in described retrieval request Including: to described content to be retrieved according to crucial morphological pattern key element, apply for that human-like key element and classification number type key element extract, Obtain one or more key element to be retrieved；Receive and select instruction；And according to described selection instruction determine one or A key element in multiple described key elements to be retrieved is described first core word.

Further, during described target information table includes first information table, the second information table and the 3rd information table at least it One, before receiving the retrieval request of user, described patent retrieval method also includes: set up according to patent initial data Basic search field tables of data, wherein, described patent initial data is published patent application document by application status Composition；According to the key word in described basic search field tables of data, set up described first information table, wherein, described First information table stores and is formed the key word of co-occurrence matrix at published described patent application document by any two Common occurrence number；According to classification number and described key word in described basic search field tables of data, set up described second Information table, wherein, comprises key word corresponding to each described classification number published described in described second information table Common occurrence number in patent application document；And according to applicant in described basic search field tables of data and described Key word, sets up described 3rd information table, wherein, comprises each described applicant corresponding in described 3rd information table The key word common occurrence number in published described patent application document.

Further, before determining the first associational word according to described first core word and target information table, described patent Search method also includes: judging the type of described first core word, wherein, the type of described first core word is crucial Morphological pattern, apply for human-like or classification number type；Wherein, in the feelings that the type judging described first core word is crucial morphological pattern Under condition, determine that described target information table includes described first information table, described second information table and described 3rd information table； In the case of the type judging described first core word is classification number type, determine that described target information table includes described Second information table；Judge the type of described first core word for application human-like in the case of, determine that described target is believed Breath table includes described 3rd information table.

Further, when the type of described first core word is crucial morphological pattern, described first associational word includes crucial morphological pattern Associational word, classification number type associational word and apply for human-like associational word, determine according to described first core word and target information table First associational word includes: pass through the first recommended models formula according to described first core word and described first information table recommend1_word=get (max_n1(fre (word, word'))) determine described crucial morphological pattern associational word, wherein, word For described first core word, recommend1_wordFor the title of described first recommended models formula, word' is described pass Keyword type associational word, n1 is the number preset and determine described crucial morphological pattern associational word, and fre (word, word') is the first core Heart word word and the crucial morphological pattern associational word word' common occurrence number in described first information table, get(max_n1(fre (word, word'))) it is that acquisition is common with described first core word word in described first information table N1 the crucial morphological pattern associational word that occurrence number is most；According to described first core word and described second information table by the Two recommended models formula recommend2_word=get (max_n2(fre (word, ipc))) determine described classification number type associational word, Wherein, ipc is described classification number type associational word, recommend2_wordFor the title of described second recommended models formula, N2 is the number preset and determine described classification number type associational word, and fre (word, ipc) is described first core word word and divides The class-mark type associational word ipc common occurrence number in described second information table, get (max_n2(fre (word, ipc))) be N2 the classification number type most with the described first common occurrence number of core word word is obtained in described second information table Associational word；And pass through the 3rd recommended models formula according to described first core word and described 3rd information table recommend3_word=get (max_n3(fre (word, appl))) determine the human-like associational word of described application, wherein, appl is The human-like associational word of described application, recommend3_wordFor the title of described 3rd recommended models formula, n3 is for presetting really The number of the fixed human-like associational word of described application, fre (word, appl) is that described first core word word is human-like with application Think the word appl common occurrence number in described 3rd information table, get (max_n3(fre (word, appl))) it is described 3rd information table obtains n3 the application human-like associational word most with the described first common occurrence number of core word word.

Further, when the type of described first core word is classification number type, described first associational word includes crucial morphological pattern According to described first core word and target information table, associational word, determines that the first associational word includes: according to described first core Word and described second information table are by the 4th recommended models formula recommend_ipc=get (max_n4(fre (ipc, word))) determine described crucial morphological pattern associational word, wherein, recommend_ipcFor the title of described 4th recommended models formula, ipc is described first core word, and word is described pass Keyword type associational word, n4 is the number preset and determine described crucial morphological pattern associational word, and fre (ipc, word) is the first core Word ipc and the crucial morphological pattern associational word word common occurrence number in described second information table, get(max_n4(fre (ipc, word))) jointly occur with described first core word ipc for obtaining in described second information table N4 the crucial morphological pattern associational word that number of times is most.

Further, when the type of described first core word is for applying for human-like, described first associational word includes crucial morphological pattern According to described first core word and target information table, associational word, determines that the first associational word includes: according to described core word and Described 3rd information table passes through the 5th recommended models formula recommend_appl=get (max_n5(fre (appl, word))) determine described crucial morphological pattern associational word, wherein, recommend_applFor the title of described 5th recommended models formula, appl is described first core word, and word is described Crucial morphological pattern associational word, n5 is the number preset and determine described crucial morphological pattern associational word, and fre (appl, word) is first Core word appl and the crucial morphological pattern associational word word common occurrence number in described 3rd information, get(max_n5(fre (appl, word))) for obtain jointly to go out with described first core word appl in described second information table N5 the crucial morphological pattern associational word that occurrence number is most.

Further, determine that search condition includes according to described first core word and described first associational word: obtain target The type of term, wherein, described target retrieval word is the selected word as retrieval elements, institute in lexical set Predicate collect be combined into described first core word and described first associational word composition set；According to described target retrieval word Described target retrieval word is classified by type, obtains different target types；And by under identical described target type Described target retrieval word logically or relation connect, and target retrieval word described under different described target types is pressed Relation according to logical AND connects generation search condition.

Further, described first associational word is multiple, is determining according to described first core word and target information table After one associational word, described patent retrieval method also includes: judge whether to receive association's instruction, wherein, described Being intended to refer to order is the instruction selecting arbitrary first associational word from multiple described first associational words；Described judging to receive In the case of association's instruction, determine that the first associational word selected by described association instruction is the second core word；And according to Described second core word and described target information table determine have the second associational word of incidence relation with described second core word, Wherein, in described target information table, also storage has and has described second associational word of incidence relation with described second core word； Wherein, determine that search condition includes according to described first core word and described first associational word: according to described first core Word, described second core word, described first associational word and described second associational word determine search condition.

Further, after determining the first associational word according to described first core word and target information table, described patent Search method also includes: receives and changes instruction；And redefine described first associational word according to described replacing instruction； Wherein, determine that search condition includes according to described first core word and described first associational word: according to described first core Word and redefine described first associational word and determine search condition.

Another aspect according to embodiments of the present invention, additionally provides a kind of patent retrieving device.

Patent retrieval method according to the present invention includes: first receives unit, for receiving the retrieval request of user；The One determines unit, for determining the first core word in described retrieval request；Second determines unit, for according to described First core word and target information table determine the first associational word, and wherein, in described target information table, storage has and described the One core word has the associational word of incidence relation；3rd determines unit, for according to described first core word and described the One associational word determines search condition；And retrieval unit, for according to described search condition patent searching file.

Further, described retrieval request includes content to be retrieved, and described first determines that unit includes: abstraction module, For to described content to be retrieved according to crucial morphological pattern key element, apply for that human-like key element and classification number type key element extract, Obtain one or more key element to be retrieved；Receiver module, is used for receiving selection instruction；And first determine module, For selecting instruction to determine that a key element in one or more described key element to be retrieved is described first core according to described Heart word.

Further, during described target information table includes first information table, the second information table and the 3rd information table at least it One, described patent retrieving device also includes: first sets up unit, before the retrieval request reception user, and root Setting up basic search field tables of data according to patent initial data, wherein, described patent initial data is by application status Disclosed patent application document composition；Second sets up unit, for according to the pass in described basic search field tables of data Keyword, sets up described first information table, wherein, stores and formed co-occurrence matrix by any two in described first information table Key word in the common occurrence number of published described patent application document；3rd sets up unit, for according to institute State classification number and described key word in basic search field tables of data, set up described second information table, wherein, described Two information tables comprise common in published described patent application document of key word corresponding to each described classification number Occurrence number；And the 4th set up unit, for according to applicant and described pass in described basic search field tables of data Keyword, sets up described 3rd information table, wherein, comprises, in described 3rd information table, the pass that each described applicant is corresponding The keyword common occurrence number in published described patent application document.

Further, described patent retrieving device also includes: the first judging unit, for according to described first core Before word and target information table determine the first associational word, it is judged that the type of described first core word, wherein, described first The type of core word is crucial morphological pattern, applies for human-like or classification number type；Wherein, described first core word is being judged In the case of type is crucial morphological pattern, determine that described target information table includes described first information table, described second information Table and described 3rd information table；In the case of the type judging described first core word is classification number type, determine institute State target information table and include described second information table；It is to apply for human-like feelings in the type judging described first core word Under condition, determine that described target information table includes described 3rd information table.

Further, when the type of described first core word is crucial morphological pattern, described first associational word includes crucial morphological pattern Associational word, classification number type associational word and apply for human-like associational word, described second determines that unit includes: second determines module, For passing through the first recommended models formula according to described first core word and described first information table recommend1_word=get (max_n1(fre (word, word'))) determine described crucial morphological pattern associational word, wherein, word For described first core word, recommend1_wordFor the title of described first recommended models formula, word' is described pass Keyword type associational word, n1 is the number preset and determine described crucial morphological pattern associational word, and fre (word, word') is the first core Heart word word and the crucial morphological pattern associational word word' common occurrence number in described first information table, get(max_n1(fre (word, word'))) it is that acquisition is common with described first core word word in described first information table N1 the crucial morphological pattern associational word that occurrence number is most；3rd determines module, for according to described first core word and institute State the second information table by the second recommended models formula recommend2_word=get (max_n2(fre (word, ipc))) determine Described classification number type associational word, wherein, ipc is described classification number type associational word, recommend2_wordFor described second The title of recommended models formula, n2 is the number preset and determine described classification number type associational word, and fre (word, ipc) is institute State the first core word word and the classification number type associational word ipc common occurrence number in described second information table, get(max_n2(fre (word, ipc))) for obtain jointly to go out with described first core word word in described second information table N2 the classification number type associational word that occurrence number is most；And the 4th determine module, for according to described first core word and Described 3rd information table passes through the 3rd recommended models formula recommend3_word=get (max_n3(fre(word,appl))) Determining the human-like associational word of described application, wherein, appl is the human-like associational word of described application, recommend3_wordFor institute Stating the title of the 3rd recommended models formula, n3 is the number preset and determine the human-like associational word of described application, Fre (word, appl) is that described first core word word and the human-like associational word appl of application is in described 3rd information table Common occurrence number, get (max_n3(fre (word, appl))) it is to obtain and described first in described 3rd information table N3 the human-like associational word of application that the common occurrence number of core word word is most.

Further, when the type of described first core word is classification number type, described first associational word includes crucial morphological pattern Associational word, described second determines that unit includes: the 5th determines module, for according to described first core word and described the Two information tables pass through the 4th recommended models formula recommend_ipc=get (max_n4(fre (ipc, word))) determine institute State crucial morphological pattern associational word, wherein, recommend_ipcFor the title of described 4th recommended models formula, ipc is described First core word, word is described crucial morphological pattern associational word, and n4 is the number preset and determine described crucial morphological pattern associational word, Fre (ipc, word) is the first core word ipc and crucial morphological pattern associational word word jointly going out in described second information table Occurrence number, get (max_n4(fre (ipc, word))) it is to obtain and described first core word ipc in described second information table N4 the crucial morphological pattern associational word that common occurrence number is most.

Further, when the type of described first core word is for applying for human-like, described first associational word includes crucial morphological pattern Associational word, described second determines that unit includes: the 6th determines module, for according to described core word and described 3rd letter Breath table is by the 5th recommended models formula recommend_appl=get (max_n5(fre (appl, word))) determine described Crucial morphological pattern associational word, wherein, recommend_applFor the title of described 5th recommended models formula, appl is described First core word, word is described crucial morphological pattern associational word, and n5 is the number preset and determine described crucial morphological pattern associational word, Fre (appl, word) is the first core word appl and common in described 3rd information of crucial morphological pattern associational word word Occurrence number, get (max_n5(fre (appl, word))) it is to obtain and described first core word in described second information table N5 the crucial morphological pattern associational word that the common occurrence number of appl is most.

Further, the described 3rd determines that unit includes: acquisition module, for obtaining the type of target retrieval word, its In, described target retrieval word is the selected word as retrieval elements in lexical set, and described lexical set is described First core word and the set of described first associational word composition；Sort module, for the class according to described target retrieval word Described target retrieval word is classified by type, obtains different target types；And generation module, for by identical institute State described target retrieval word under target type logically or relation connect, and by described under different described target types Target retrieval word logically with relation be connected generation search condition.

Further, described first associational word is multiple, and described patent retrieving device also includes: the second judging unit, For after determining the first associational word according to described first core word and target information table, it may be judged whether receive association Instruction, wherein, described association instruction is the instruction selecting arbitrary first associational word from multiple described first associational words； 4th determines unit, in the case of judging to receive described association instruction, determines that described association instruction is selected The first associational word selected is the second core word；And the 5th determine unit, for according to described second core word and described Target information table determines have the second associational word of incidence relation, wherein, described target information with described second core word In table, also storage has and has described second associational word of incidence relation with described second core word；Wherein, described 3rd true Cell includes: the 7th determines module, for according to described first core word, described second core word, described first Associational word and described second associational word determine search condition.

Further, described patent retrieving device also includes: second receives unit, for according to described first core After word and target information table determine the first associational word, receive and change instruction；And the 6th determine unit, for according to Described replacing instruction redefines described first associational word, and wherein, the described 3rd determines that unit includes: the 8th determines mould Block, for according to described first core word with redefine described first associational word and determine search condition.

In embodiments of the present invention, the retrieval request of reception user is used；Determine the first core in described retrieval request Word；Determine the first associational word according to described first core word and target information table, wherein, described target information table is deposited Contain, with described first core word, there is the associational word of incidence relation；According to described first core word and described first association Word determines search condition；And according to described search condition patent searching file.By determining core according to retrieval request Word, and then the associational word relevant to core word can be determined according to the target information table pre-set, it is achieved that with core Heart word as the retrieval wish of user, for user have direction, can control to provide more relevant to core word and permissible For the associational word as search condition, and above-mentioned associational word can be that same type retrieves key element with core word Vocabulary, it is also possible to be and the vocabulary that core word is dissimilar retrieval key element so that user can have bigger selection empty Between determine search condition according to the associational word of core word and recommendation, and then carry out patent retrieval according to above-mentioned search condition, Solve the prior art Patent retrieval inaccurate problem of result, and then reach raising patent search result accuracy Effect.

Accompanying drawing explanation

Accompanying drawing described herein is used for providing a further understanding of the present invention, constitutes the part of the application, this Bright schematic description and description is used for explaining the present invention, is not intended that inappropriate limitation of the present invention.At accompanying drawing In:

Fig. 1 is the flow chart of patent retrieval method according to embodiments of the present invention；And

Fig. 2 is the schematic diagram of patent retrieving device according to embodiments of the present invention.

Detailed description of the invention

In order to make those skilled in the art be more fully understood that the present invention program, below in conjunction with in the embodiment of the present invention Accompanying drawing, is clearly and completely described the technical scheme in the embodiment of the present invention, it is clear that described embodiment It is only the embodiment of a present invention part rather than whole embodiments.Based on the embodiment in the present invention, ability The every other embodiment that territory those of ordinary skill is obtained under not making creative work premise, all should belong to The scope of protection of the invention.

It should be noted that term " first " in description and claims of this specification and above-mentioned accompanying drawing, " Two " it is etc. for distinguishing similar object, without being used for describing specific order or precedence.Should be appreciated that this Sample use data can exchange in the appropriate case, in order to embodiments of the invention described herein can with except Here the order beyond those illustrating or describing is implemented.Additionally, term " includes " and " having " and they Any deformation, it is intended that cover non-exclusive comprising, such as, contain series of steps or the process of unit, side Method, system, product or equipment are not necessarily limited to those steps or the unit clearly listed, but can include the clearest List or for intrinsic other step of these processes, method, product or equipment or unit.

According to embodiments of the present invention, it is provided that a kind of patent document retrieval embodiment of the method, it should be noted that Step shown in the flow chart of accompanying drawing can perform in the computer system of such as one group of computer executable instructions, and And, although show logical order in flow charts, but in some cases, can be to be different from order herein Step shown or described by execution.

Fig. 1 is the flow chart of patent retrieval method according to embodiments of the present invention.As it is shown in figure 1, the method includes step Rapid S102 is to step S110, specific as follows:

Step S102: receive the retrieval request of user.Specifically, retrieval request includes content to be retrieved, to be retrieved Content can be word, sentence, paragraph or chapter.

In embodiments of the present invention, user can pass through the special of the terminal patent retrieval method to the execution embodiment of the present invention Profit searching system sends retrieval request.Being provided with input text area in patent search system, user can input at text By the related text content that the equipment inputs such as keyboard are to be retrieved in district, duplication, paste operation are supported in text input area, And allow the content that edit-modify inputs.If input content of text more, often more than 10 words, font size by No. 28 words are contracted to step by step until No. 12 words, and the greatest length that input content is supported by input text area is 1000 word Symbol.When the content of text of user's input is single vocabulary, it is provided that with the above-mentioned single vocabulary of input as prefix Relevant information selects for user.Such as: the content of text of input for " starting " time, it is provided that electromotor and The input prompt such as start the engine.

Step S104: determine the first core word in retrieval request.Specifically, the type of the first core word can be to close Keyword type, such as automobile；Can be applicant's type, such as company ABC；Can also be classification number type, such as GO1M 99/00(2011.01)。

Step S106: determine the first associational word according to the first core word and target information table, wherein, in target information table Storage has and has the associational word of incidence relation with the first core word.

Specifically, there is during target information table is multiple information tables pre-set the letter of incidence relation with the first core word Breath table, the corresponding different target information table of different types of first core word.The type of the first associational word equally can also It is divided into crucial morphological pattern, applies for human-like and classification number type.First associational word can be one, it is also possible to for multiple, and The type of each first associational word can be identical with the type of the first core word, it is also possible to different, i.e. each first Think that word and the first core word can be that same type retrieves key element, it is also possible to for dissimilar retrieval key element.

In embodiments of the present invention, by the associational word that can provide the user with core word is dissimilar retrieval key element, The range of choice of search condition be not confined between same type retrieval key element by user, true for subsequent user Determine search condition to provide and select the most flexibly, and then reach to improve the effect of user satisfaction.

Step S108: determine search condition according to the first core word and the first associational word.In embodiments of the present invention, inspection Rope condition can be collectively constituted by core word and associational word, can be only made up of core word, it is also possible to only by associative phrase Become, in the search condition whether collectively constituted by core word and associational word or the search condition being only made up of associational word Concrete associational word and the quantity of associational word can select according to demand to arrange.

Step S110: according to search condition patent searching file.Specifically, according to search condition patent searching file Range of search is whole patent datas.

In embodiments of the present invention, by determining core word according to retrieval request, and then can be according to the mesh pre-set Mark information table determine the associational word relevant to core word, it is achieved that using core word as the retrieval wish of user, for There is direction at family, can control to provide associational word that is more relevant to core word and that be used as search condition, and And the vocabulary that above-mentioned associational word can be and core word is same type retrieval key element, it is also possible to it is to be different from core word The vocabulary of type retrieval key element so that user can have bigger selection space true according to the associational word of core word and recommendation Determine search condition, and then carry out patent retrieval according to above-mentioned search condition, solve prior art Patent retrieval result Inaccurate problem, has reached to improve the effect of patent search result accuracy.

Specifically, in embodiments of the present invention, retrieval request includes content to be retrieved, can be the most true Determine the first core word in retrieval request:

Step 1041: to content to be retrieved according to crucial morphological pattern key element, apply for that human-like key element and classification number type key element are carried out Extraction, obtains one or more key element to be retrieved.Specifically, can be known by embedded keyword abstraction and entity Other algorithm automatically to content to be retrieved according to crucial morphological pattern key element, apply for that human-like key element and classification number type key element extract, Thus obtain the crucial morphological pattern information element, the classification number type information element that comprise in above-mentioned content to be retrieved and apply for human-like Information element, the above-mentioned key element being drawn into is key element to be retrieved, wherein, applies for that human-like information element can include public affairs Department's title or Personal name.

Step 1043: receive and select instruction.Specifically, instruction is selected to be used for allowing user to select step 1041 according to demand Which key element to be retrieved middle extraction obtains as core word.

Step 1045: determine that a key element in one or more key element to be retrieved is the first core according to selecting instruction Word.

In embodiments of the present invention, after content to be retrieved carried out extraction obtaining one or more key element to be retrieved, Also allowing for user selects a key element to be retrieved in said one or multiple key element to be retrieved as core according to demand Heart word, and non-immediate using said one or multiple key element to be retrieved is retrieved automatically as search condition, reach Improve the effect of retrieval accuracy, and then improve the satisfaction of user.

Specifically, target information table includes at least one of first information table, the second information table and the 3rd information table, Before receiving the retrieval request of user, the patent retrieval method that the embodiment of the present invention is provided is further comprising the steps of:

Step S1: setting up basic search field tables of data according to patent initial data, wherein, patent initial data is by Shen Please state be published patent application document composition.Specifically, in basic search field tables of data every be recorded as one Individual patent application, the content of every record includes that application number, key word, applicant and the IPC of single patent extraction divide Class-mark information.

In embodiments of the present invention, it has been disclosed that patent application document can acquire in several ways, such as: The modes such as CNIPR Patent Information Services platform or Soopat patent data search engine.

Step S3: according to the key word in basic search field tables of data, sets up first information table, wherein, the first letter Breath table stores and is formed the key word of co-occurrence matrix by any two and jointly go out occurrence at published patent application document Number.In embodiments of the present invention, first information table can also be called keyword data table, and common occurrence number can also Being called a frequency number, this step is it is, have recorded in keyword data table in basic search field tables of data and all appoint Anticipate two the piece frequency number that can form the key word of co-occurrence matrix in published patent application document.Such as, base Key word " automobile " in this search field tables of data and key word " computer " composition co-occurrence matrix, it is assumed that key word " automobile " and key word " computer " 100 patent application documents in 1000 published patent application documents In jointly occur in that, then the piece frequency number of key word " automobile " and key word " computer " (i.e. common occurrence number) It it is 100 times.The recording mode of the piece frequency number of the key word of other any two composition co-occurrence matrix is said with the example above As bright.

It should be noted that the same key word in basic search field tables of data can be with basic search field tables of data In other multiple key words separately constitute co-occurrence matrix, then need to record above-mentioned same key in keyword data table Word respectively with other key words common occurrence number in published patent application document.

Step S5: according to classification number and key word in basic search field tables of data, sets up the second information table, wherein, Second information table comprises key word corresponding to each classification number in published patent application document, jointly goes out occurrence Number.Specifically, the second information table is the information table set up for dimension with classification number, the key word that each classification number is corresponding For comprise this classification number patent application document in the key word that comprises.In embodiments of the present invention, the second information table is also Can be called classification number-keyword data table, this step is it is, in classification number-keyword data table with classification number be Dimension, have recorded each classification number and the key word corresponding with this classification number being total in published patent application document Same occurrence number.

It should be noted that a classification number may be comprised by multiple patent application documents, then said one classification number Corresponding key word is made up of, the most substantially the whole different key words comprised according to above-mentioned multiple patent application documents Same classification number in search field tables of data can be the most corresponding with multiple key words in basic search field tables of data, In classification number-keyword data table, so need to record above-mentioned same classification number every with above-mentioned multiple key words respectively The individual key word common frequency of occurrence in published patent application document.

Step S7: according to applicant and key word in basic search field tables of data, sets up the 3rd information table, wherein, 3rd information table comprises key word corresponding to each applicant in published patent application document, jointly goes out occurrence Number.Specifically, the 3rd information table is to apply for the information table that artificial dimension is set up, the key word that each applicant is corresponding For comprise the applicant patent application document in the key word that comprises.In embodiments of the present invention, the 3rd information table is also Applicant-keyword data table can be called, with the artificial dimension of application in this step namely applicant-keyword data table Degree, have recorded each applicant key word corresponding with the applicant jointly going out in published patent application document Occurrence number.

It should be noted that an applicant may be comprised by multiple patent application documents, then said one applicant Corresponding key word is made up of, the most substantially the whole different key words comprised according to above-mentioned multiple patent application documents Same applicant in search field tables of data can be the most corresponding with multiple key words in basic search field tables of data, In applicant-keyword data table, so need to record above-mentioned same applicant every with above-mentioned multiple key words respectively The individual key word common occurrence number in published patent application document.

Such as: apply for artificial company A, comprise and apply for that the published patent application document of artificial company A has two, Being patent application document B and patent application document C respectively, wherein, the key word comprised in patent application document B is Mobile phone and antenna, the key word comprised in patent application document C is computer, it is assumed that company A and mobile phone are jointly 25 Occurring in the published patent application document of a piece, company A and antenna are jointly in 20 published patent application documents Occurring, company A and computer occur jointly in 50 published patent application documents, then in applicant-key In word tables of data, record company A and the common occurrence number of mobile phone are 25 times, company A and antenna jointly go out occurrence Number is 20 times, and the common occurrence number of record company A and computer is 50 times.

Specifically, before determining the first associational word according to the first core word and target information table, embodiment of the present invention institute The patent retrieval method provided also includes step S9, specific as follows:

Step S9: judging the type of the first core word, wherein, the type of the first core word is crucial morphological pattern, applicant Type or classification number type；In the case of the type judging the first core word is crucial morphological pattern, determine target information table bag Include first information table, the second information table and the 3rd information table；It is classification number type in the type judging the first core word In the case of, determine that target information table includes the second information table；It is that application is human-like in the type judging the first core word In the case of, determine that target information table includes the 3rd information table.

In embodiments of the present invention, by judging the type of the first core word, and then can be according to the class of the first core word Type determines the scope of target information table, determines that with core word be dissimilar retrieval key element for follow-up, but has association The associational word of relation provides data basis.

According to foregoing, the type of the first core word can be three types, is crucial morphological pattern, classification respectively Number type and application are human-like, and the composition of the target information table that different types of core word is corresponding is different, individually below for the The type of one core word is for crucial morphological pattern, classification number type with when applying for human-like, to how according to the first core word and target Information table determines that the first associational word is specifically introduced explanation.

Type one: when the type of the first core word is crucial morphological pattern, now, the first associational word includes crucial morphological pattern connection Think word, classification number type associational word and apply for human-like associational word, root can be completed by step S1061 to step S1065 Determining the first associational word according to the first core word and target information table, step S1061 is specific as follows to step S1065:

Step S1061: pass through the first recommended models formula according to the first core word and first information table recommend1_word=get (max_n1(fre (word, word'))) determine crucial morphological pattern associational word, wherein, word is the One core word, recommend1_wordBeing the title of the first recommended models formula, word' is crucial morphological pattern associational word, n1 For presetting the number determining crucial morphological pattern associational word, fre (word, word') is the first core word word and crucial morphological pattern The associational word word' common occurrence number in first information table, get (max_n1(fre (word, word'))) it is first Information table obtains the crucial morphological pattern associational word of n1 most with the first common occurrence number of core word word.In this step In Zhou, achieved by the first recommended models formula and obtain and the first common occurrence number of core word in first information table The crucial morphological pattern associational word of most n1.Specifically, the concrete numerical value of n1 can be arranged according to demand.

Assuming: the first core word word is automobile, and n1 is set as 5, now, above-mentioned first recommended models formula is recommend1_Automobile=get (max₅(fre (automobile, word'))), represent at first information table (i.e. keyword data table) Middle acquisition occurrence number common with " automobile " comes the key word of first 5, and the above-mentioned key word coming first 5 is Crucial morphological pattern associational word, be also and the associational word of the first the most relevant property of core word simultaneously.

In embodiments of the present invention, achieved with first by model formation according to the first core word, first information table Core word is the determination of same type retrieval key element.

Step S1063: pass through the second recommended models formula according to the first core word and the second information table recommend2_word=get (max_n2(fre (word, ipc))) determine classification number type associational word, wherein, ipc is classification number Type associational word, recommend2_wordBe the title of the second recommended models formula, for presetting, n2 determines that classification number type is associated The number of word, fre (word, ipc) is that the first core word word and classification number type associational word ipc is in the second information table Common occurrence number, get (max_n2(fre (word, ipc))) it is to obtain and the first core word word in the second information table N2 the classification number type associational word that common occurrence number is most.In this step, realized by the second recommended models formula N2 the classification number type associational word most with the first common occurrence number of core word is obtained in the second information table.Equally , the concrete numerical value of n2 can also be arranged according to demand.

Assuming: the first core word word is automobile, and n2 is set as 5, now, above-mentioned second recommended models formula is recommend2_Automobile=get (max₅(fre (automobile, ipc))), represent at the second information table (i.e. classification number-key word number According to table) in obtain the classification number that occurrence number common with " automobile " comes first 5, the above-mentioned classification coming first 5 Number it is classification number type associational word, is also and the associational word of the first the most relevant property of core word simultaneously.

In embodiments of the present invention, achieved with first by model formation according to the first core word, the second information table Core word is the determination of dissimilar retrieval key element.

Step S1065: pass through the 3rd recommended models formula according to the first core word and the 3rd information table recommend3_word=get (max_n3(fre (word, appl))) determine the human-like associational word of application, wherein, appl is Shen Type of asking someone associational word, recommend3_wordBe the title of the 3rd recommended models formula, for presetting, n3 determines that application is human-like The number of associational word, fre (word, appl) is that the first core word word and the human-like associational word appl of application is in the 3rd information Common occurrence number in table, get (max_n3(fre (word, appl))) it is to obtain and the first core in the 3rd information table N3 the human-like associational word of application that the common occurrence number of word word is most.In this step, by the 3rd recommended models It is human-like that formula achieves n3 the application that acquisition is most with the first common occurrence number of core word in the 3rd information table Think word.Same, the concrete numerical value of n3 can also be arranged according to demand.

Assuming: the first core word word is automobile, and n3 is set as 5, now, above-mentioned 3rd recommended models formula is recommend3_Automobile=get (max₅(fre (automobile, appl))), represent at the 3rd information table (i.e. applicant-key word number According to table) in obtain the applicant that occurrence number common with " automobile " comes first 5, the above-mentioned application coming first 5 People is the human-like associational word of application, is also and the associational word of the first the most relevant property of core word simultaneously.

In embodiments of the present invention, achieved with first by model formation according to the first core word, the 3rd information table Core word is the determination of dissimilar retrieval key element.

Type two: when the type of the first core word is classification number type, now, the first associational word includes crucial morphological pattern connection Think word, can complete to determine the first associational word, step according to the first core word and target information table by step S1067 S1067 is specific as follows:

Step S1067: pass through the 4th recommended models formula according to the first core word and the second information table recommend_ipc=get (max_n4(fre (ipc, word))) determine crucial morphological pattern associational word, wherein, recommend_ipcBeing the title of the 4th recommended models formula, ipc is the first core word, and word is crucial morphological pattern association Word, n4 is the number preset and determine crucial morphological pattern associational word, and fre (ipc, word) is the first core word ipc and crucial morphological pattern The associational word word common occurrence number in the second information table, get (max_n4(fre (ipc, word))) it is at the second letter Breath table obtains the crucial morphological pattern associational word of n4 most with the first common occurrence number of core word ipc.In this step, Acquisition in the second information table is achieved most with the first common occurrence number of core word by the 4th recommended models formula N4 crucial morphological pattern associational word.Same, the concrete numerical value of n4 can also be arranged according to demand.

Assuming: the first core word ipc is that AAA, n4 are set as 5, now, above-mentioned 4th recommended models formula is recommend_AAA=get (max₅(fre (AAA, word))), represent at the second information table (i.e. classification number-key word number According to table) in obtain the key word that occurrence number common with " AAA " comes first 5, the above-mentioned key word coming first 5 It is crucial morphological pattern associational word, is also and the associational word of the first the most relevant property of core word simultaneously.

When the type of type three: the first core word is for applying for human-like, now, the first associational word includes crucial morphological pattern association Word, can complete to determine the first associational word, step according to the first core word and target information table by step S1069 S1069 is specific as follows:

Step S1069: pass through the 5th recommended models formula according to core word and the 3rd information table recommend_appl=get (max_n5(fre (appl, word))) determine crucial morphological pattern associational word, wherein, recommend_applBeing the title of the 5th recommended models formula, appl is the first core word, and word is crucial morphological pattern connection Thinking word, n5 is the number preset and determine crucial morphological pattern associational word, and fre (appl, word) is the first core word appl and pass The keyword type associational word word common occurrence number in the 3rd information, get (max_n5(fre (appl, word))) it is Two information tables obtain the crucial morphological pattern associational word of n5 most with the first common occurrence number of core word appl.At this In step, achieve acquisition and the first core word in the 3rd information table by the 5th recommended models formula and jointly go out occurrence N5 the crucial morphological pattern associational word that number is most.Same, the concrete numerical value of n5 can also be arranged according to demand.

Assuming: the first core word appl is Samsung, and n5 is set as 5, now, above-mentioned 5th recommended models formula is recommend_Samsung=get (max₅(fre (Samsung, word))), represent at the 3rd information table (i.e. applicant-key word number According to table) in obtain the key word that occurrence number common with " Samsung " comes first 5, the above-mentioned key coming first 5 Word is crucial morphological pattern associational word, is also and the associational word of the first the most relevant property of core word simultaneously.

Preferably, realize determining inspection according to the first core word and the first associational word by step S1081 to step S1085 Rope condition includes:

Step S1081: obtaining the type of target retrieval word, wherein, target retrieval word is the most selected in lexical set As the word of retrieval elements, lexical set is the first core word and the set of the first associational word composition, and this step is namely Obtain user's selected type as retrieval elements word from the set that the first core word and the first associational word form.Need It is noted that selected retrieval elements includes the first core word and the first associational word, the first core can be only included Word, it is also possible to only include the first associational word, specifically can select to determine according to user's request.

In embodiments of the present invention, semantic computation association extension the not direct structure of retrieval key element (i.e. associational word) obtained Become retrieval type (i.e. retrieval elements) to participate in retrieval, as the selection of the associational word of retrieval elements, be open to user's choosing Select, thus reached the purpose making whole retrieving controlled.

Step S1083: according to the type of target retrieval word, target retrieval word is classified, obtain different target class Type, this step is it is, to selected the first core word as retrieval elements and the first associational word, the first core word Or the first associational word is classified, now, no longer core word and associational word are made a distinction.

It should be noted that owing to the type of the first core word and the first associational word is all same three types, respectively It is crucial morphological pattern, applies for human-like and classification number type, so target type is at most the most just divided into three types, distinguish equally It is crucial morphological pattern, applies for human-like and classification number type.

Step S1085: by target retrieval word under same target type logically or relation connect, and by difference mesh Under mark type target retrieval word logically with relation is connected and generates search condition, it is, each same target type Under target retrieval word between use logic " OR " annexation, between each target type use logic " AND " company Connect relation.Assume: the target type obtained in step S1083 has three kinds, be crucial morphological pattern respectively, apply for human-like With classification number type, crucial morphological pattern includes that two close target retrieval words, is automobile and computer respectively, applies for human-like including one Individual target retrieval word: Samsung, classification number type includes target retrieval word a: AAA, then above-mentioned target retrieval word forms Search condition be (automobile OR computer) AND Samsung AND AAA.

It should be noted that in performing the patent search system of patent retrieval method of the embodiment of the present invention, work as target When type is crucial morphological pattern, the range of search of the key word (i.e. target retrieval word) under this target type being included: mark Topic, summary, principal claim, claims, description full text and accompanying drawing explanation；When target type is classification number type, The range of search of the classification number (i.e. target retrieval word) under this target type being included: IPC code and main IPC；When When target type is for applying for human-like, the range of search of the applicant's (i.e. target retrieval word) under this target type being included: Applicant, inventor, patentee and related right people.In this patent searching system, user can be to the most selected Retrieval elements repeatedly revise, with send determine search instruction time the selected retrieval elements of correspondence be as the criterion generation retrieval Condition.

In embodiments of the present invention, can be automatically according to the selection of user, carry out search condition assembles relation, its knot Really user is adjustable, has reached the controllability of neither impact retrieval result, client can be helped again to assemble complex retrieval The effect of logic.

Preferably, the first associational word is multiple, according to the first core word and target information table determine the first associational word it After, the patent retrieval method that the embodiment of the present invention is provided also include step S11 to step S15, specific as follows:

Step S11: judging whether to receive association's instruction, wherein, association's instruction is for selecting from multiple first associational words Selecting the instruction of arbitrary first associational word, this step is it is, judge whether to receive selection from multiple first associational words Arbitrarily associational word carries out the instruction of association.

Step S13: in the case of judging to receive association's instruction, determines the first association selected by association's instruction Word is the second core word, i.e. if receiving the finger selecting any associational word to carry out association from multiple first associational words Order, then using the first associational word selected by association's instruction as new core word.

Step S15: determine, with the second core word, there is the second of incidence relation according to the second core word and target information table Associational word, wherein, in target information table, also storage has and has the second associational word of incidence relation with the second core word, this Step, with step S106, specifically determines the same process determining the first associational word of process of the second associational word, with specific reference to upper State content, explanation is not repeated herein.

Now, determine that search condition includes according to the first core word and the first associational word: according to the first core word, second Core word, the first associational word and the second associational word determine search condition.In embodiments of the present invention, according to the first core Word, the second core word, the first associational word and the second associational word determine that the process of search condition is with the first core word and first Associational word determines the process of search condition, is not repeated.

It should be noted that when the second associational word is multiple, it is also possible to judge whether to receive from multiple second associations Word selects any associational word carry out the instruction of association, select to appoint from multiple second associational words if it is judged that receive Meaning associational word carries out the instruction of association, then the second associational word selected by association can being instructed as the 3rd core word, And then can determine, with the 3rd core word, there is the 3rd associational word of incidence relation according to the 3rd core word and target information table, So the like, multiple core word can be obtained with repeated execution of steps S11 to step S15 and has with this core word Relevant associational word.

In embodiments of the present invention, by associational word can be changed into new core word, thus obtain new with above-mentioned Core word has the associational word of incidence relation, it is achieved that astride hierarchy is associated, and is not limited to the effect of the association of same level, enters One step has reached to improve the effect of user satisfaction, and provides good basis for later retrieval result accuracy.

In performing the patent search system of patent retrieval method of the embodiment of the present invention, core word and associational word are all can Depending on change search interface show, specifically, centered by core word, associational word successively with circular configuration arranged distribution at core Around heart word.Specifically, core word shows with No. 12 words, when core word is more than 10 words, and the most only display three Point.In this patent search system, user can also perform search operation, amendment operation and check operation core word. Search operation is for obtaining preliminary examination result using this core word as search condition, retrieval, and above-mentioned preliminary examination result can show All retrieve the entry number of results, and show first five content of retrieval result, specifically include title, application number, when Front statutory status, patent type and summary info.This preindexing result is shown by the form of floating window, and user is permissible By clicking on the button of packing up in floating window, floating window is packed up.Amendment operation is used for returning under initial text entry mode, Namely return to input text area, thus content of text can be edited again.Check that operation is for showing this The full text content to be retrieved of secondary input, is mainly used in the full text when content of text is longer and checks.

Same, in this patent searching system, user can also carry out certain operations to each associational word, is specially Amendment operation, search operaqtion, association's operation and deletion action.Amendment operation is for carrying out currently available associational word Edit-modify；Search operation, for using certain currently available associational word as search condition, is retrieved and is obtained preliminary examination result, Same, this preliminary examination result can show the entry number all retrieving result, and shows first five content retrieving result, Specifically include title, application number, Current statutory state, patent type and summary info.This preindexing result is same Can be shown by the form of floating window, floating window can be packed up by user by clicking on the button of packing up in floating window.Association behaviour Act on using certain currently available associational word as new core word, continue to obtain the core word new with this and have and associate The associational word of relation；Deletion action is for deleting certain currently available associational word.

In this patent searching system, the fixing display all the time of the core word before current core word is at visual search interface Fixed area, such as in the upper left corner, a upper core word dotted line of current core word and current core word connect, Other connects with dotted line, to show hierarchical relationship between the core word of fixed area.Along with going deep into step by step of association, Can form an association tree in above-mentioned fixed area, this association tree have recorded the whole expansion process of this search operaqtion, The node of association tree is the core word every time carrying out association's operation, clicks on the arbitrary node on this association tree, above-mentioned arbitrarily The core word that node is corresponding can launch display again, and connects this node and its father node, the relation of child node with dotted line. In this patent system, by forming the association tree about core word, it is achieved the omnidistance record to search operaqtion process, And serve reversibility and the effect of repeatability ensureing associative process.

Preferably, after determining the first associational word according to the first core word and target information table, embodiment of the present invention institute The patent retrieval method provided also includes: receives and changes instruction；The first associational word is redefined according to changing instruction, its In, determine that search condition includes according to the first core word and the first associational word: according to the first core word with redefine the One associational word determines search condition.

In embodiments of the present invention, user can change instruction by sending, it is achieved associates having with current core word The effect that the associational word of relation is replaced, and expand the follow-up range of choice choosing retrieval elements of user.Need Bright, in addition to core word, all of associational word all can be replaced, the quantity that concrete each type associational word is changed Determining according to the model formation that foregoing provides, now, search condition is according to the association after current core word and replacing Word determines.Particularly, user can send and repeatedly change instruction, and search condition by current core word and is sent out for the last time Associational word corresponding after changing instruction is sent to determine.

It should be noted that for aforesaid each method embodiment, in order to be briefly described, therefore it is all expressed as one it be The combination of actions of row, but those skilled in the art should know, the present invention not limiting by described sequence of movement System, because according to the present invention, some step can use other orders or carry out simultaneously.Secondly, art technology Personnel also should know, embodiment described in this description belongs to preferred embodiment, involved action and module Not necessarily necessary to the present invention.

Through the above description of the embodiments, those skilled in the art is it can be understood that arrive according to above-mentioned enforcement The method of example can add the mode of required general hardware platform by software and realize, naturally it is also possible to by hardware, but In the case of Hen Duo, the former is more preferably embodiment.Based on such understanding, technical scheme substantially or Saying that the part contributing prior art can embody with the form of software product, this computer software product is deposited Storage is in a storage medium (such as ROM/RAM, magnetic disc, CD), including some instructions with so that a station terminal Equipment (can be mobile phone, computer, server, or the network equipment etc.) performs described in each embodiment of the present invention Method.

The embodiment of the present invention additionally provides a kind of patent retrieving device, and this patent retrieving device may be used for performing the present invention The patent retrieval method of embodiment.

Fig. 2 is the schematic diagram of patent retrieving device according to embodiments of the present invention.As in figure 2 it is shown, the inspection of this patent document Rope device includes: first receives unit 10, first determine unit 20, second determine that unit the 30, the 3rd determines unit 40 and retrieval unit 50, wherein:

First receives unit 10 for receiving the retrieval request of user.Specifically, retrieval request includes content to be retrieved, Content to be retrieved can be word, sentence, paragraph or chapter.

In embodiments of the present invention, user can be by terminal to patent retrieving device special with the embodiment of the present invention Profit searching system sends retrieval request.Being provided with input text area in patent search system, user can input at text By the related text content that the equipment inputs such as keyboard are to be retrieved in district, duplication, paste operation are supported in text input area, And allow the content that edit-modify inputs.If input content of text more, often more than 10 words, font size by No. 28 words are contracted to step by step until No. 12 words, and the greatest length that input content is supported by input text area is 1000 word Symbol.When the content of text of user's input is single vocabulary, it is provided that with the above-mentioned single vocabulary of input as prefix Relevant information selects for user.Such as: the content of text of input for " starting " time, it is provided that electromotor and The input prompt such as start the engine.

First determines that unit 20 is for the first core word determining in retrieval request.Specifically, the type of the first core word Can be key word type, such as automobile；Can be applicant's type, such as company ABC；Can also be classification number type, Such as GO1M 99/00 (2011.01).

Second determines that unit 30 is for determining the first associational word, wherein, target according to the first core word and target information table In information table, storage has and has the associational word of incidence relation with the first core word.

Specifically, there is during target information table is multiple information tables pre-set the letter of incidence relation with the first core word Breath table, the corresponding different target information table of different types of first core word.The type of the first associational word equally can also Being divided into crucial morphological pattern, apply for human-like and classification number type, the first associational word can be one, it is also possible to for multiple, and The type of each first associational word can be identical with the type of the first core word, it is also possible to different, i.e. each first Think that word and the first core word can be that same type retrieves key element, it is also possible to retrieve key element for dissimilar class.

3rd determines that unit 40 is for determining search condition according to the first core word and the first associational word.Implement in the present invention In example, search condition can be collectively constituted by core word and associational word, can be only made up of core word, it is also possible to only by Associational word forms, the search condition whether collectively constituted by core word and associational word or the inspection being only made up of associational word Concrete associational word and the quantity of associational word in rope condition can select to arrange according to demand.

Retrieval unit 50 is for according to search condition patent searching file.Specifically, according to search condition patent searching literary composition The range of search of part is whole patent datas.

In embodiments of the present invention, by determining core word according to retrieval request, and then can be according to the mesh pre-set Mark information table determine the associational word relevant to core word, it is achieved that using core word as the retrieval wish of user, for There is direction at family, can control to provide associational word that is more relevant to core word and that be used as search condition, and And the vocabulary that above-mentioned associational word can be and core word is same type retrieval key element, it is also possible to it is to be different from core word The vocabulary of type retrieval key element so that user can have bigger selection space true according to the associational word of core word and recommendation Determine search condition, and then carry out patent retrieval according to above-mentioned search condition, solve prior art Patent retrieval result Inaccurate problem, and then reached to improve the effect of patent search result accuracy.

Specifically, in embodiments of the present invention, retrieval request includes content to be retrieved, and first determines that unit 20 includes Abstraction module, receiver module and first determine module, wherein:

Abstraction module for content to be retrieved according to crucial morphological pattern key element, apply for that human-like key element and classification number type key element are entered Row extraction, obtains one or more key element to be retrieved.Specifically, embedded keyword abstraction and entity can be passed through Recognizer automatically to content to be retrieved according to crucial morphological pattern key element, apply for that human-like key element and classification number type key element are taken out Take, thus obtain crucial morphological pattern information element, classification number type information element and the application comprised in above-mentioned content to be retrieved Human-like information element, the above-mentioned key element being drawn into is key element to be retrieved, wherein, applies for that human-like information element includes public affairs Department's title or Personal name.

Receiver module is used for receiving selection instruction.Specifically, instruction is selected to be used for allowing user's selecting extraction according to demand In module, which key element to be retrieved extraction obtains as core word.

First determines that module is for being the according to the key element selecting instruction to determine in one or more key element to be retrieved One core word.

Specifically, target information table includes at least one of first information table, the second information table and the 3rd information table, this The patent retrieving device that inventive embodiments is provided also include first set up unit, second set up unit, the 3rd set up single Unit and the 4th sets up unit, wherein:

First sets up unit for before receiving the retrieval request of user, sets up basic retrieval according to patent initial data Field data table, wherein, patent initial data is that published patent application document forms by application status.Specifically, In basic search field tables of data, every is recorded as a patent application, the content of every record include application number, single Key word, applicant and the IPC code information of patent extraction.

Second sets up unit for according to the key word in basic search field tables of data, sets up first information table, wherein, First information table stores and is formed common at published patent application document of the key word of co-occurrence matrix by any two Occurrence number.In embodiments of the present invention, first information table can also be called keyword data table, common occurrence number Can also be called a frequency number, this unit is it is, have recorded in keyword data table in basic search field tables of data All any two can form the key word of co-occurrence matrix piece frequency number in published patent application document.Example As, the key word " automobile " in basic search field tables of data and key word " computer " composition co-occurrence matrix, it is assumed that Key word " automobile " and key word " computer " 100 patent Shens in 1000 published patent application documents File please be occurred in that jointly, then the piece frequency number of key word " automobile " and key word " computer " (goes out the most jointly Occurrence number) it is 100 times.The recording mode of the piece frequency number of the key word of other any two composition co-occurrence matrix is with upper As stating illustration.

3rd sets up unit for according to classification number and key word in basic search field tables of data, sets up the second information table, Wherein, the second information table comprises common in published patent application document of key word corresponding to each classification number Occurrence number.Specifically, the second information table is the information table set up for dimension with classification number, and each classification number is corresponding Key word is the key word comprised in the patent application document comprising this classification number.In embodiments of the present invention, the second letter Breath table can also be called classification number-keyword data table, this unit it is, in classification number-keyword data table with point Class-mark is dimension, have recorded each classification number and the key word corresponding with this classification number at published patent application document In common occurrence number.

4th sets up unit for according to applicant and key word in basic search field tables of data, sets up the 3rd information table, Wherein, the 3rd information table comprises common in published patent application document of key word corresponding to each applicant Occurrence number.Specifically, the 3rd information table is to apply for the information table that artificial dimension is set up, and each applicant is corresponding Key word is the key word comprised in the patent application document comprising the applicant.In embodiments of the present invention, the 3rd letter Breath table can also be called applicant-keyword data table, with application in this unit namely applicant-keyword data table Artificial dimension, have recorded each applicant key word corresponding with the applicant in published patent application document Common occurrence number.

Specifically, the patent retrieving device that the embodiment of the present invention is provided also includes the first judging unit, wherein, first Judging unit is for before determining the first associational word according to the first core word and target information table, it is judged that the first core word Type, wherein, the type of the first core word is crucial morphological pattern, applies for human-like or classification number type；Judging first In the case of the type of core word is crucial morphological pattern, determine target information table include first information table, the second information table and 3rd information table；In the case of the type judging the first core word is classification number type, determine that target information table includes Second information table；Judge the type of the first core word for application human-like in the case of, determine that target information table includes 3rd information table.

Type one: when the type of the first core word is crucial morphological pattern, now, the first associational word includes crucial morphological pattern connection Think word, classification number type associational word and apply for human-like associational word, second determine unit 30 include second determine module, the 3rd Determine that module and the 4th determines module, wherein:

Second determines that module is for passing through the first recommended models formula according to the first core word and first information table recommend1_word=get (max_n1(fre (word, word'))) determine crucial morphological pattern associational word, wherein, word is the One core word, recommend1_wordBeing the title of the first recommended models formula, word' is crucial morphological pattern associational word, n1 For presetting the number determining crucial morphological pattern associational word, fre (word, word') is the first core word word and crucial morphological pattern The associational word word' common occurrence number in first information table, get (max_n1(fre (word, word'))) it is first Information table obtains the crucial morphological pattern associational word of n1 most with the first common occurrence number of core word word.At this mould In block, achieved by the first recommended models formula and obtain and the first common occurrence number of core word in first information table The crucial morphological pattern associational word of most n1.Specifically, the concrete numerical value of n1 can be arranged according to demand.

3rd determines that module is for passing through the second recommended models formula according to the first core word and the second information table recommend2_word=get (max_n2(fre (word, ipc))) determine classification number type associational word, wherein, ipc is classification number Type associational word, recommend2_wordBe the title of the second recommended models formula, for presetting, n2 determines that classification number type is associated The number of word, fre (word, ipc) is that the first core word word and classification number type associational word ipc is in the second information table Common occurrence number, get (max_n2(fre (word, ipc))) it is to obtain and the first core word word in the second information table N2 the classification number type associational word that common occurrence number is most.In this module, realized by the second recommended models formula N2 the classification number type associational word most with the first common occurrence number of core word is obtained in the second information table.Equally , the concrete numerical value of n2 can also be arranged according to demand.

4th determines that module is for passing through the 3rd recommended models formula according to the first core word and the 3rd information table recommend3_word=get (max_n3(fre (word, appl))) determine the human-like associational word of application, wherein, appl is Shen Type of asking someone associational word, recommend3_wordBe the title of the 3rd recommended models formula, for presetting, n3 determines that application is human-like The number of associational word, fre (word, appl) is that the first core word word and the human-like associational word appl of application is in the 3rd information Common occurrence number in table, get (max_n3(fre (word, appl))) it is to obtain and the first core in the 3rd information table N3 the human-like associational word of application that the common occurrence number of word word is most.In this module, by the 3rd recommended models It is human-like that formula achieves n3 the application that acquisition is most with the first common occurrence number of core word in the 3rd information table Think word.Same, the concrete numerical value of n3 can also be arranged according to demand.

Type two: when the type of the first core word is classification number type, now, the first associational word includes crucial morphological pattern connection Thinking word, second determines that unit 30 includes that the 5th determines module, wherein:

5th determines that module is for passing through the 4th recommended models formula according to the first core word and the second information table recommend_ipc=get (max_n4(fre (ipc, word))) determine crucial morphological pattern associational word, wherein, recommend_ipcBeing the title of the 4th recommended models formula, ipc is the first core word, and word is crucial morphological pattern association Word, n4 is the number preset and determine crucial morphological pattern associational word, and fre (ipc, word) is the first core word ipc and crucial morphological pattern The associational word word common occurrence number in the second information table, get (max_n4(fre (ipc, word))) it is at the second letter Breath table obtains the crucial morphological pattern associational word of n4 most with the first common occurrence number of core word ipc.In this module, Acquisition in the second information table is achieved most with the first common occurrence number of core word by the 4th recommended models formula N4 crucial morphological pattern associational word.Same, the concrete numerical value of n4 can also be arranged according to demand.

When the type of type three: the first core word is for applying for human-like, now, the first associational word includes crucial morphological pattern association Word, second determines that unit 30 includes that the 6th determines module, wherein:

6th determines that module is for passing through the 5th recommended models formula according to core word and the 3rd information table recommend_appl=get (max_n5(fre (appl, word))) determine crucial morphological pattern associational word, wherein, recommend_applBeing the title of the 5th recommended models formula, appl is the first core word, and word is crucial morphological pattern connection Thinking word, n5 is the number preset and determine crucial morphological pattern associational word, and fre (appl, word) is the first core word appl and pass The keyword type associational word word common occurrence number in the 3rd information, get (max_n5(fre (appl, word))) it is Two information tables obtain the crucial morphological pattern associational word of n5 most with the first common occurrence number of core word appl.At this In module, achieve acquisition and the first core word in the 3rd information table by the 5th recommended models formula and jointly go out occurrence N5 the crucial morphological pattern associational word that number is most.Same, the concrete numerical value of n5 can also be arranged according to demand.

Preferably, the 3rd determines that unit 40 includes acquisition module, sort module and generation module, wherein:

Acquisition module is for obtaining the type of target retrieval word, and wherein, target retrieval word is the most selected in lexical set As the word of retrieval elements, lexical set is the first core word and the set of the first associational word composition, and this module is namely Obtain user's selected type as retrieval elements word from the set that the first core word and the first associational word form.Need It is noted that selected retrieval elements includes the first core word and the first associational word, the first core can be only included Word, it is also possible to only include the first associational word, specifically can select to determine according to user's request.

Sort module, for classifying target retrieval word according to the type of target retrieval word, obtains different target class Type, this module is it is, to selected the first core word as retrieval elements and the first associational word, the first core word Or the first associational word is classified, now, no longer core word and associational word are made a distinction.

Generation module for by target retrieval word under same target type logically or relation connect, and by difference mesh Under mark type target retrieval word logically with relation is connected and generates search condition, it is, each same target type Under target retrieval word between use logic " OR " annexation, between each target type use logic " AND " company Connect relation.Assume: the target type obtained in sort module has three kinds, be crucial morphological pattern respectively, apply for human-like and Classification number type, crucial morphological pattern includes that two close target retrieval words, is automobile and computer respectively, applies for human-like including one Target retrieval word: Samsung, classification number type includes target retrieval word a: AAA, then above-mentioned target retrieval word composition Search condition is (automobile OR computer) AND Samsung AND AAA.

It should be noted that in there is the patent search system of patent retrieving device of the embodiment of the present invention, work as target When type is crucial morphological pattern, the range of search of the key word (i.e. target retrieval word) under this target type being included: mark Topic, summary, principal claim, claims, description full text and accompanying drawing explanation；When target type is classification number type, The range of search of the classification number (i.e. target retrieval word) under this target type being included: IPC code and main IPC；When When target type is for applying for human-like, the range of search of the applicant's (i.e. target retrieval word) under this target type being included: Applicant, inventor, patentee and related right people.In this patent searching system, user can be to the most selected Retrieval elements repeatedly revise, with send determine search instruction time the selected retrieval elements of correspondence be as the criterion generation retrieval Condition.

Preferably, the first associational word is multiple, and the patent retrieving device that the embodiment of the present invention is provided also includes: second Judging unit, the 4th determine that unit and the 5th determines unit, wherein:

Second judging unit is for after determining the first associational word according to the first core word and target information table, it is judged that be The no association that receives instructs, and wherein, association's instruction is the finger selecting arbitrary first associational word from multiple first associational words Order, this unit is it is, judge whether to receive and select any associational word to carry out association from multiple first associational words Instruction.

4th determines that unit in the case of judging to receive association's instruction, determines the selected by association's instruction One associational word is the second core word, i.e. select any associational word to join if received from multiple first associational words The instruction thought, then using the first associational word selected by association's instruction as new core word.

5th determines that unit is for determining have incidence relation with the second core word according to the second core word and target information table The second associational word, wherein, in target information table also storage have with the second core word have incidence relation second association Word, this unit determines unit with second, specifically determines the same process determining the first associational word of process of the second associational word, With specific reference to foregoing, explanation is not repeated herein.

Now, the 3rd determines that unit includes that the 7th determines module, the 7th determine module for according to the first core word, Two core words, the first associational word and the second associational word determine search condition.In embodiments of the present invention, according to the first core Heart word, the second core word, the first associational word and the second associational word determine that the process of search condition is with the first core word and the One associational word determines the process of search condition, is not repeated.

It should be noted that when the second associational word is multiple, it is also possible to judge whether to receive from multiple second associations Word selects any associational word carry out the instruction of association, select to appoint from multiple second associational words if it is judged that receive Meaning associational word carries out the instruction of association, then the second associational word selected by association can being instructed as the 3rd core word, And then can determine, with the 3rd core word, there is the 3rd associational word of incidence relation according to the 3rd core word and target information table, So the like, can repeat to call the second judging unit, the 4th determine that unit and the 5th determines unit, obtain many Individual core word and there is the associational word of incidence relation with this core word.

In having the patent search system of patent retrieving device of the embodiment of the present invention, core word and associational word are all can Depending on change search interface show, specifically, centered by core word, associational word successively with circular configuration arranged distribution at core Around heart word.Specifically, core word shows with No. 12 words, when core word is more than 10 words, and the most only display three Point.In this patent search system, user can also perform search operation, amendment operation and check operation core word. Search operation is for obtaining preliminary examination result using this core word as search condition, retrieval, and above-mentioned preliminary examination result can show All retrieve the entry number of results, and show first five content of retrieval result, specifically include title, application number, when Front statutory status, patent type and summary info.This preindexing result is shown by the form of floating window, and user is permissible By clicking on the button of packing up in floating window, floating window is packed up.Amendment operation is used for returning under initial text entry mode, Namely return to input text area, thus content of text can be edited again.Check that operation is for showing this The full text content to be retrieved of secondary input, is mainly used in the full text when content of text is longer and checks.

Preferably, the patent retrieving device that the embodiment of the present invention is provided also includes that the second reception unit and the 6th determines list Unit, wherein, second receives unit is used for after determining the first associational word according to the first core word and target information table, Receive and change instruction；6th determines that unit is for redefining the first associational word according to replacing instruction；Now, the 3rd is true Cell includes that the 8th determines module, and the 8th determines that module is for determining retrieval according to the first core word and the first associational word Condition includes: according to the first core word with redefine the first associational word and determine search condition.

As can be seen from the above description, the present invention solves the retrieval of prior art Patent result is inaccurate asks Topic, and then reached to improve the effect of patent search result accuracy.

The invention described above embodiment sequence number, just to describing, does not represent the quality of embodiment.

In the above embodiment of the present invention, the description to each embodiment all emphasizes particularly on different fields, and does not has in certain embodiment The part described in detail, may refer to the associated description of other embodiments.

In several embodiments provided herein, it should be understood that disclosed technology contents, can be passed through other Mode realize.Wherein, device embodiment described above is only schematically, the division of the most described unit, Can be that a kind of logic function divides, actual can have other dividing mode, the most multiple unit or assembly when realizing Can in conjunction with or be desirably integrated into another system, or some features can be ignored, or does not performs.Another point, institute The coupling each other shown or discuss or direct-coupling or communication connection can be by some interfaces, unit or mould The INDIRECT COUPLING of block or communication connection, can be being electrical or other form.

The described unit illustrated as separating component can be or may not be physically separate, shows as unit The parts shown can be or may not be physical location, i.e. may be located at a place, or can also be distributed to On multiple unit.Some or all of unit therein can be selected according to the actual needs to realize the present embodiment scheme Purpose.

It addition, each functional unit in each embodiment of the present invention can be integrated in a processing unit, it is also possible to It is that unit is individually physically present, it is also possible to two or more unit are integrated in a unit.Above-mentioned integrated Unit both can realize to use the form of hardware, it would however also be possible to employ the form of SFU software functional unit realizes.

If described integrated unit is using the form realization of SFU software functional unit and as independent production marketing or use, Can be stored in a computer read/write memory medium.Based on such understanding, technical scheme essence On the part that in other words prior art contributed or this technical scheme completely or partially can be with software product Form embodies, and this computer software product is stored in a storage medium, including some instructions with so that one Platform computer equipment (can be for personal computer, server or the network equipment etc.) performs each embodiment institute of the present invention State all or part of step of method.And aforesaid storage medium includes: USB flash disk, read only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), portable hard drive, magnetic disc or CD Etc. the various media that can store program code.

The above is only the preferred embodiment of the present invention, it is noted that for the ordinary skill people of the art For Yuan, under the premise without departing from the principles of the invention, it is also possible to make some improvements and modifications, these improve and profit Decorations also should be regarded as protection scope of the present invention.

Claims

1. a patent retrieval method, it is characterised in that including:

Receive the retrieval request of user；

Determine the first core word in described retrieval request；

The first associational word, wherein, described target information table is determined according to described first core word and target information table Middle storage has and has the associational word of incidence relation with described first core word；

Search condition is determined according to described first core word and described first associational word；And

According to described search condition patent searching file.

Patent retrieval method the most according to claim 1, it is characterised in that described retrieval request includes to be retrieved Content, determines that the first core word in described retrieval request includes:

To described content to be retrieved according to crucial morphological pattern key element, apply for that human-like key element and classification number type key element are taken out Take, obtain one or more key element to be retrieved；

Receive and select instruction；And

According to described select instruction to determine a key element in one or more described key element to be retrieved is described the One core word.

Patent retrieval method the most according to claim 1, it is characterised in that described target information table includes the first letter At least one of breath table, the second information table and the 3rd information table, before receiving the retrieval request of user, described Patent retrieval method also includes:

Setting up basic search field tables of data according to patent initial data, wherein, described patent initial data is by Shen Please state be published patent application document composition；

According to the key word in described basic search field tables of data, set up described first information table, wherein, institute State first information table stores and formed the key word of co-occurrence matrix in published described patent application by any two The common occurrence number of file；

According to classification number and described key word in described basic search field tables of data, set up described second information table, Wherein, described second information table comprises key word corresponding to each described classification number in published described patent Common occurrence number in application documents；And

According to applicant and described key word in described basic search field tables of data, set up described 3rd information table, Wherein, described 3rd information table comprises key word corresponding to each described applicant in published described patent Common occurrence number in application documents.

Patent retrieval method the most according to claim 3, it is characterised in that according to described first core word and mesh Before mark information table determines the first associational word, described patent retrieval method also includes:

Judging the type of described first core word, wherein, the type of described first core word is crucial morphological pattern, Shen Please human-like or classification number type；

Wherein, in the case of the type judging described first core word is crucial morphological pattern, described target is determined Information table includes described first information table, described second information table and described 3rd information table；Described in judging In the case of the type of the first core word is classification number type, determine that described target information table includes described second information Table；Judge the type of described first core word for application human-like in the case of, determine described target information table Including described 3rd information table.

Patent retrieval method the most according to claim 4, it is characterised in that the type of described first core word is for closing During keyword type, described first associational word includes crucial morphological pattern associational word, classification number type associational word and applies for human-like Think word, determine that the first associational word includes according to described first core word and target information table:

The first recommended models formula is passed through according to described first core word and described first information table recommend1_word=get (max_n1(fre (word, word'))) determine described crucial morphological pattern associational word, wherein, Word is described first core word, recommend1_wordFor the title of described first recommended models formula, word' For described crucial morphological pattern associational word, n1 is the number preset and determine described crucial morphological pattern associational word, Fre (word, word') is that the first core word word and crucial morphological pattern associational word word' is at described first information table In common occurrence number, get (max_n1(fre (word, word'))) for obtaining and institute in described first information table State n1 the crucial morphological pattern associational word that the first common occurrence number of core word word is most；

The second recommended models formula is passed through according to described first core word and described second information table recommend2_word=get (max_n2(fre (word, ipc))) determine described classification number type associational word, wherein, ipc For described classification number type associational word, recommend2_wordFor the title of described second recommended models formula, n2 is Presetting the number determining described classification number type associational word, fre (word, ipc) is described first core word word and divides The class-mark type associational word ipc common occurrence number in described second information table, get(max_n2(fre (word, ipc))) it is to obtain in described second information table with described first core word word altogether With n2 the classification number type associational word that occurrence number is most；And

The 3rd recommended models formula is passed through according to described first core word and described 3rd information table recommend3_word=get (max_n3(fre (word, appl))) determine the human-like associational word of described application, wherein, Appl is the human-like associational word of described application, recommend3_wordFor the title of described 3rd recommended models formula, N3 is the number preset and determine the human-like associational word of described application, and fre (word, appl) is described first core word Word and the application human-like associational word appl common occurrence number in described 3rd information table, get(max_n3(fre (word, appl))) it is to obtain and described first core word word in described 3rd information table N3 the human-like associational word of application that common occurrence number is most.

Patent retrieval method the most according to claim 4, it is characterised in that the type of described first core word is for dividing During class-mark type, described first associational word includes crucial morphological pattern associational word, believes according to described first core word and target Breath table determines that the first associational word includes:

The 4th recommended models formula is passed through according to described first core word and described second information table recommend_ipc=get (max_n4(fre (ipc, word))) determine described crucial morphological pattern associational word, wherein, recommend_ipcFor the title of described 4th recommended models formula, ipc is described first core word, and word is institute Stating crucial morphological pattern associational word, n4 is the number preset and determine described crucial morphological pattern associational word, and fre (ipc, word) is First core word ipc and the crucial morphological pattern associational word word common occurrence number in described second information table, get(max_n4(fre (ipc, word))) it is that acquisition is common with described first core word ipc in described second information table N4 the crucial morphological pattern associational word that occurrence number is most.

Patent retrieval method the most according to claim 4, it is characterised in that the type of described first core word is Shen During type of asking someone, described first associational word includes crucial morphological pattern associational word, believes according to described first core word and target Breath table determines that the first associational word includes:

The 5th recommended models formula is passed through according to described core word and described 3rd information table recommend_appl=get (max_n5(fre (appl, word))) determine described crucial morphological pattern associational word, wherein, recommend_applFor the title of described 5th recommended models formula, appl is described first core word, and word is Described crucial morphological pattern associational word, n5 is the number preset and determine described crucial morphological pattern associational word, fre (appl, word) It is the first core word appl and the crucial morphological pattern associational word word common occurrence number in described 3rd information, get(max_n5(fre (appl, word))) it is to obtain in described second information table with described first core word appl altogether With n5 the crucial morphological pattern associational word that occurrence number is most.

Patent retrieval method the most according to claim 1, it is characterised in that according to described first core word and described First associational word determines that search condition includes:

Obtaining the type of target retrieval word, wherein, described target retrieval word is for have selected conduct in lexical set The word of retrieval elements, described lexical set is described first core word and the set of described first associational word composition；

According to the type of described target retrieval word, described target retrieval word is classified, obtain different target class Type；And

By target retrieval word described under identical described target type logically or relation connect, and by difference institute State described target retrieval word under target type logically with relation be connected generation search condition.

Patent retrieval method the most according to claim 1, it is characterised in that described first associational word is multiple, After determining the first associational word according to described first core word and target information table, described patent retrieval method also wraps Include:

Judging whether to receive association's instruction, wherein, described association instruction is for from multiple described first associational words Select the instruction of arbitrary first associational word；

In the case of judging to receive described association instruction, determine first selected by described association instruction Think that word is the second core word；And

Determine, with described second core word, there is incidence relation according to described second core word and described target information table The second associational word, wherein, in described target information table, also storage has to have with described second core word and associates Described second associational word of system；

Wherein, determine that search condition includes according to described first core word and described first associational word: according to described First core word, described second core word, described first associational word and described second associational word determine search condition.

Patent retrieval method the most according to claim 1, it is characterised in that according to described first core word and mesh After mark information table determines the first associational word, described patent retrieval method also includes:

Receive and change instruction；And

Described first associational word is redefined according to described replacing instruction；

Wherein, determine that search condition includes according to described first core word and described first associational word: according to described First core word and redefine described first associational word and determine search condition.

11. 1 kinds of patent retrieving devices, it is characterised in that including:

First receives unit, for receiving the retrieval request of user；

First determines unit, for determining the first core word in described retrieval request；

Second determines unit, for determining the first associational word according to described first core word and target information table, its In, in described target information table, storage has and has the associational word of incidence relation with described first core word；

3rd determines unit, for determining search condition according to described first core word and described first associational word； And

Retrieval unit, for according to described search condition patent searching file.

12. patent retrieving devices according to claim 11, it is characterised in that described retrieval request includes to be retrieved Content, described first determines that unit includes:

Abstraction module, for described content to be retrieved according to crucial morphological pattern key element, apply for human-like key element and classification Number type key element extracts, and obtains one or more key element to be retrieved；

Receiver module, is used for receiving selection instruction；And

First determines module, for determining in one or more described key element to be retrieved according to described selection instruction A key element be described first core word.

13. patent retrieving devices according to claim 11, it is characterised in that described target information table includes the first letter At least one of breath table, the second information table and the 3rd information table, described patent retrieving device also includes:

First sets up unit, for, before receiving the retrieval request of user, setting up base according to patent initial data This search field tables of data, wherein, described patent initial data is published patent application literary composition by application status Part forms；

Second sets up unit, for according to the key word in described basic search field tables of data, sets up described the One information table, wherein, stores in described first information table and is formed the key word of co-occurrence matrix by any two The common occurrence number of disclosed described patent application document；

3rd sets up unit, is used for according to classification number and described key word in described basic search field tables of data, Set up described second information table, wherein, described second information table comprises the key that each described classification number is corresponding The word common occurrence number in published described patent application document；And

4th sets up unit, is used for according to applicant and described key word in described basic search field tables of data, Set up described 3rd information table, wherein, described 3rd information table comprises the key that each described applicant is corresponding The word common occurrence number in published described patent application document.

14. patent retrieving devices according to claim 13, it is characterised in that described patent retrieving device also includes:

First judging unit, for according to described first core word and target information table determine the first associational word it Before, it is judged that the type of described first core word, wherein, the type of described first core word is crucial morphological pattern, Shen Please human-like or classification number type；

15. patent retrieving devices according to claim 14, it is characterised in that the type of described first core word is for closing During keyword type, described first associational word includes crucial morphological pattern associational word, classification number type associational word and applies for human-like Thinking word, described second determines that unit includes:

Second determines module, for recommending mould according to described first core word and described first information table by first Type formula recommend1_word=get (max_n1(fre (word, word'))) determine described crucial morphological pattern associational word, Wherein, word is described first core word, recommend1_wordFor the title of described first recommended models formula, Word' is described crucial morphological pattern associational word, and n1 is the number preset and determine described crucial morphological pattern associational word, Fre (word, word') is that the first core word word and crucial morphological pattern associational word word' is at described first information table In common occurrence number, get (max_n1(fre (word, word'))) for obtaining and institute in described first information table State n1 the crucial morphological pattern associational word that the first common occurrence number of core word word is most；

3rd determines module, for recommending mould according to described first core word and described second information table by second Type formula recommend2_word=get (max_n2(fre (word, ipc))) determine described classification number type associational word, its In, ipc is described classification number type associational word, recommend2_wordFor the title of described second recommended models formula, N2 is the number preset and determine described classification number type associational word, and fre (word, ipc) is described first core word word With the classification number type associational word ipc common occurrence number in described second information table, get(max_n2(fre (word, ipc))) it is to obtain in described second information table with described first core word word altogether With n2 the classification number type associational word that occurrence number is most；And

4th determines module, for recommending mould according to described first core word and described 3rd information table by the 3rd Type formula recommend3_word=get (max_n3(fre (word, appl))) determine the human-like associational word of described application, its In, appl is the human-like associational word of described application, recommend3_wordName for described 3rd recommended models formula Claiming, n3 is the number preset and determine the human-like associational word of described application, and fre (word, appl) is described first core Word word and the application human-like associational word appl common occurrence number in described 3rd information table, get(max_n3(fre (word, appl))) it is to obtain and described first core word word in described 3rd information table N3 the human-like associational word of application that common occurrence number is most.

16. patent retrieving devices according to claim 14, it is characterised in that the type of described first core word is for dividing During class-mark type, described first associational word includes crucial morphological pattern associational word, and described second determines that unit includes:

5th determines module, for recommending mould according to described first core word and described second information table by the 4th Type formula recommend_ipc=get (max_n4(fre (ipc, word))) determine described crucial morphological pattern associational word, Wherein, recommend_ipcFor the title of described 4th recommended models formula, ipc is described first core word, word For described crucial morphological pattern associational word, n4 is the number preset and determine described crucial morphological pattern associational word, fre (ipc, word) It is the first core word ipc and the crucial morphological pattern associational word word common occurrence number in described second information table, get(max_n4(fre (ipc, word))) it is that acquisition is common with described first core word ipc in described second information table N4 the crucial morphological pattern associational word that occurrence number is most.

17. patent retrieving devices according to claim 14, it is characterised in that the type of described first core word is Shen During type of asking someone, described first associational word includes crucial morphological pattern associational word, and described second determines that unit includes:

6th determines module, for public by the 5th recommended models according to described core word and described 3rd information table Formula recommend_appl=get (max_n5(fre (appl, word))) determine described crucial morphological pattern associational word, wherein, recommend_applFor the title of described 5th recommended models formula, appl is described first core word, and word is Described crucial morphological pattern associational word, n5 is the number preset and determine described crucial morphological pattern associational word, fre (appl, word) It is the first core word appl and the crucial morphological pattern associational word word common occurrence number in described 3rd information, get(max_n5(fre (appl, word))) it is to obtain in described second information table with described first core word appl altogether With n5 the crucial morphological pattern associational word that occurrence number is most.

18. patent retrieving devices according to claim 11, it is characterised in that the described 3rd determines that unit includes:

Acquisition module, for obtaining the type of target retrieval word, wherein, described target retrieval word is at word finder The selected word as retrieval elements in conjunction, described lexical set is described first core word and described first association The set of word composition；

Sort module, for described target retrieval word being classified according to the type of described target retrieval word, To different target types；And

Generation module, for by target retrieval word described under identical described target type logically or relation even Connect, and by target retrieval word described under different described target types logically with relation be connected generation search bar Part.

19. patent retrieving devices according to claim 11, it is characterised in that described first associational word is multiple, institute State patent retrieving device also to include:

Second judging unit, for according to described first core word and target information table determine the first associational word it After, it may be judged whether receiving association's instruction, wherein, described association instruction is for from multiple described first associational words Select the instruction of arbitrary first associational word；

4th determines unit, for, in the case of judging to receive described association instruction, determining described association The first associational word selected by instruction is the second core word；And

5th determines unit, for determining and described second according to described second core word and described target information table Core word has the second associational word of incidence relation, and wherein, in described target information table, also storage has and described the Two core words have described second associational word of incidence relation；

Wherein, the described 3rd determines that unit includes: the 7th determines module, for according to described first core word, Described second core word, described first associational word and described second associational word determine search condition.

20. patent retrieving devices according to claim 11, it is characterised in that described patent retrieving device also includes:

Second receive unit, for according to described first core word and target information table determine the first associational word it After, receive and change instruction；And

6th determines unit, for redefining described first associational word according to described replacing instruction,

Wherein, the described 3rd determines that unit includes: the 8th determines module, for according to described first core word and Redefine described first associational word and determine search condition.