CN102483749B - Method, system, and apparatus for delivering query results from an electronic document collection - Google Patents

Abstract

A method, system, and article are provided for efficiently and effectively searching an electronic document collection. Each of the documents in the collection is pre-divided into sub-sections. One or more profiles are created, each including a selection of at least one of the sections of the documents in the collection. In addition, a weight is assigned to each of the selected sections in the profile. Based upon the parameters of a query and selection of a profile, select sub-sections of each document are employed to compare query data to the underlying document collection. A compilation of documents is created with data matching the query data, and a relevancy score is computed for each document in the compilation. The relevancy score is then leveraged to sort the documents in a manner to convey relevancy to the query submission.

Description

The method of Query Result, system and equipment is transmitted from electronic literature set

Technical field

The present invention relates to electronic literature set, to this electronic literature set submit Query, and display Query Result.More specifically, the present invention relates to by generating retrieval to each chapters and sections specified weight of intellecture property document to be retrieved concise and to the point, and show this Query Result based on the relevance at least one being retrieved to concise and to the point returned Query Result.

Background technology

That submits to worldwide any Patent Office all must meet some requirements for all intellecture property documents examined, and it must be novel, practical in non-obvious that these conditions comprise each intellecture property document.In order to prepare the intellecture property document for examining rightly, the prior knowledge property right document (that is, prior art) understood in correlative technology field is helpful, this is because every invention can only authorize a patent right.Determine that the process of prior art is retrieval.Usually, energy focuses on and theme can be authorized maybe can to protect on theme by the writer that result for retrieval contributes to intellecture property application subsequently, and helps to formulate a set of rational strategy, to realize inventor or the proprietary target of intellecture property.

Known, before technological revolution enters into the current electronic information epoch, intellecture property retrieval is undertaken by manual.Searcher browses disclosure, and based on the classification of categorizing system determination the disclosure content, carries out the retrieval of document and record subsequently in this classification.Have recognized that, searcher browses the suitable chapters and sections of intellecture property document intuitively based on the limited range of carried out retrieval.Along with the appearance of infotech, because most of mandate intellecture property and disclosed application only Electronically exist, so manual information retrieval has not been suitable for most examination.Along with the appearance of the intellecture property document of electronic format, the Similar strategies that manual information retrieval adopts also can be used for retrieve electronic intellectual property data storehouse.

Different classes of retrieval can be adopted, to obtain different results.Such as, novelty search can be adopted, to determine whether the application submitting intellecture property to.Product can be adopted to evade index infringement, to determine whether product drops in the protection domain of the claim of existing intellecture property.Invalid retrieval can be adopted, to determine that whether the claim of the intellecture property of having authorized is effective etc.Different classes of retrieval do not supported by existing Electronic Intellectual Property Right literature search instrument.Thus, retrieval people (also referred to as searcher) needs to bear following workload: according to the scope of retrieval, limits the chapters and sections needing in intellecture property document to browse when retrieving.Because the quantity of the mandate intellecture property in database and disclosed unsettled intellecture property application constantly increases, so more pertinent literature browsed by each retrieval needs, thus add the burden of searcher.

Therefore, searcher needs to use a kind of instrument for identifying that inquiry submits to result to alleviate the workload relevant to this Query Result of assessment, and this instrument utilizes the advantage of the electronic format of intellecture property document.This instrument should enable searcher in retrieving, utilize the different chapters and sections of intellecture property document, thus more high-level efficiency and determine more effectively accurately, association and the result for retrieval expected.

Summary of the invention

The present invention includes for the method for efficiently and effectively patent searching literature collection, system and product.

In one aspect of the invention, a kind of computer implemented method for the result for retrieval specified associations to electronic literature set is provided.Collect to the set of patent documentation and index, described in each in described set, document has multiple chapters and sections.Identify in described set described in each document each described in chapters and sections.For described literature collection sets up retrieval briefly.Described retrieval briefly comprise from described set each document select at least one through identify chapters and sections.Concise and to the point for each, to chapters and sections specified weight selected by each.When to described set submit Query, select retrieval concise and to the point, and also specify the data in the chapters and sections of weight to compare through what identify data query and described literature collection.Mark to each document compute associations returned in documentation, described documentation is created on described submit Query.Based on described relevance scoring as calculated, to described documentation rating.Then, based on described grade, dynamic limitation is carried out to the result of described compilation.Based on applied described dynamic limitation, generate the first compilation of the described pertinent literature through classification.

In another aspect of this invention, provide a kind of computer system, described computer system comprises the processor communicated with storage medium, and electronic literature set stores on said storage.Described electronic literature set is the compilation of intellecture property document.The characteristic of knowledge based property right document, described in each in described set, document has multiple chapters and sections.Controller is used to collect to described literature collection and index.Described controller communicates with documentation management device, each chapters and sections of each document in set described in the identification of described documentation management device.In addition, concise and to the point manager is used for for described literature collection sets up retrieval briefly.Described concise and to the point manager communicates with described documentation management device, and being selected into each document in compilation set in described retrieval briefly through identification chapters and sections.Except concrete chapters and sections are selected into described concise and to the point in except, chapters and sections specified weight selected by each during described concise and to the point manager is also concise and to the point to each.Described weight is the reflection of the importance of related Sections.When inquiring about, inquiry manager is to described literature collection submit Query.Described inquiry comprise select at least one retrieval concise and to the point and by data query and described document described briefly in reflect each described in data in chapters and sections compare.After the described submission of described searching, managing device, generate the compilation of Patents document and returned.Described each document returned in compilation comprise with have that specified weight and relevance mark at least one through identifying the match query of the data in concise and to the point chapters and sections.Also arrange relevance omniselector to communicate with described inquiry manager, thus the result of described compilation dynamically limits based on described grade to the document rating in described compilation.Based on applied described dynamic limitation, generate the first compilation of the pertinent literature through classification.

In still yet another aspect of the present, provide a kind of product being provided with computer readable carrier, described computer readable carrier comprises computer program instructions, and it is for the result for retrieval specified associations to the electronic literature set on computer memory.Described computer readable carrier comprises the computer program instructions to described literature collection specified associations.Be provided for the instruction that the set of intellecture property document is collected and indexed.Patent documentation described in each in described set is divided into multiple chapters and sections.After described set is indexed, be provided for identifying each document in described set each described in the instruction of chapters and sections.Once identify the described chapters and sections of described document, be provided for as described literature collection sets up the concise and to the point instruction of retrieval.Described retrieval be briefly selected from each document in described set each through identifying chapters and sections.In addition, the instruction to chapters and sections specified weight described in each through identifying in described retrieval is briefly provided for.When to described literature collection submit Query, provide instruction for select at least one retrieval concise and to the point and by the described document in data query and described set described briefly in data in described chapters and sections through identifying compare.Then, provide instruction be return in documentation each document compute associations scoring and based on described scoring to the document rating in described set.Once complete described ranking, then based on described grade, instruction is provided to carry out the result collected described in dynamic limitation.Based on the described dynamic limitation being applied to described compilation, generate the first compilation through the pertinent literature of classification and returned.

By the explanation to the preferred embodiment of the present invention carried out referring to accompanying drawing, other features and advantages of the present invention will be more obvious.

Accompanying drawing explanation

The accompanying drawing of reference here forms a part for instructions.Unless expressly stated, feature shown in the drawings only for illustration of some embodiment of the present invention, instead of illustrates all embodiments of the present invention.In addition, this is not containing the meaning in contrast.

Fig. 1 represents to identify that the chapters and sections of patent documentation are to generate the process flow diagram of one or more concise and to the point process.

Fig. 2 is the process flow diagram of the process being expressed as one or more concise and to the point generation time weight.

Fig. 3 represents to adopt time weight to reflect the process flow diagram producing the process of the position of string matching in each concise and to the point chapters and sections.

Fig. 4 represents to generate the secondary concise and to the point and process flow diagram of the process of result for retrieval specified weight that returns to submit Query.

Fig. 5 represents the process flow diagram to the secondary concise and to the point process of Query Result set application.

Fig. 6 is the process flow diagram of the process represented Query Result classification.

Fig. 7 be represent according to the preferred embodiment of the present invention to return and the process flow diagram of process of result specified associations through classification.

Fig. 8 represents that the process flow diagram of the process of dynamic limitation is carried out in the display to the Query Result of base document set.

Fig. 9 represents the process flow diagram utilizing graphic user interface as instrument, the Query Result of base document set dynamically to be arranged to the process of restriction.

Figure 10 is the block diagram of presentation graphic user interface embodiment.

Figure 11 represents for classifying and resolving the block diagram of one group of instrument of the Query Result in base document set.

Embodiment

Should be easily understood that, here, can arrange and design the present invention by different configurations and such as summarize in the accompanying drawings and illustrated assembly.Therefore, as shown in drawings, be hereafter only be selected from representational embodiment of the present invention to the detailed description of equipment of the present invention, system and method, instead of will the scope of protection of present invention be limited.

The functional unit described in this instructions is called manager and controller.Manager and/or controller can realize in the programmable hardware device such as such as field programmable gate array, programmable logic array, programmable logic device (PLD).The software that manager and/or controller also can be performed by various processor realizes.Such as, the appointment manager be made up of executable code and/or controller can comprise the one or more physical block or logical block that computer instruction forms, and these computer instructions may be constructed such such as object, program, function or other structures.However, the executable file of equivalent manager and/or controller does not have to be physically puts together, but the different instruction being stored in diverse location can be comprised, when these command logics ground is united, then their composition managers and/or controller realize the define objective of manager and/or controller.

In fact, the manager be made up of executable code and/or controller can be an instruction or many instructions, even can be distributed in multiple different code segment, different application programs and multiple different storer.Similarly, here, service data can be specified (be identified) and illustrate in the scope of manager and/or controller, and it may be embodied as any suitable form, also can be structured in the data structure of any suitable type.Service data can be integrated into individual data set, maybe can be distributed to the diverse location comprising different memory, can also be present on system or network as electronic signal at least in part.

" the selected embodiment " mentioned in whole instructions, " embodiment " or " embodiment " mean, comprise at least one embodiment of the present invention in conjunction with certain features, structure or feature described by this embodiment.Therefore, the term " selected embodiment " occurred everywhere at whole instructions, " in one embodiment " or " in an embodiment " not necessarily refer to same embodiment.

And, in one or more embodiments, described feature, structure or feature can be combined in any suitable manner.There is provided a large amount of detailed descriptions hereinafter, so that complete understanding embodiments of the invention.But person of ordinary skill in the field it will be appreciated that in the one or more specific detail of omission, or when with additive method, component, apparatus etc., also can realize the present invention.In other cases, for avoiding making inventive point of the present invention become obscure, well-known structure, apparatus or operation are then no longer shown specifically or are illustrated.

By reference to the accompanying drawing embodiment that the present invention may be better understood, wherein, in whole instructions, use the part that identical Reference numeral represents identical.Only sketch the equipment consistent with the invention protected at this, some preferred embodiment of system and method by the mode of citing below.

general introduction

Intellecture property literature collection is the compilation of authorizing announcement and disclosed application.Patent documentation set is the subset of intellecture property literature collection.Patent documentation comprises form and the publication forms of application of granted patent.What the difference between these two kinds of document classifications determined them can the value of exercise the right.More specifically, granted patent is the realized property power can exercised in law court, and publication application is undelegated application, is namely unsettled patent right.Each patent documentation is resolved into multiple chapters and sections, and each chapters and sections comprise written word or phrase (also referred to as string data).In order to can set be retrieved, based on the chapters and sections of each document in set, each document is resolved, and to each chapters and sections specified weight through parsing of intellecture property document.Weight is the numerical metric of the significance level of one or more particular chapter in document, for inquiry.Selected document chapters and sections together form with the weight being assigned to selected chapters and sections and retrieve briefly.Based on range of search, retrieval can be locked in the particular chapter of document, or different weights can be specified to the data of matching inquiry in each chapters and sections of document.For Query Result can be shown according to submit Query, dynamic limitation can be carried out to the relevance of this result.More specifically, and/or concise and to the point feature can be retrieved carry out the relevance of dynamic conditioning with this result based on the entirety of the statistical study of this Query Result, this Query Result.Therefore, quantification and display that concise and to the point generation and selection are directly involved in result for retrieval is retrieved.

ins and outs

Below, the accompanying drawing with reference to the part forming instructions illustrates embodiment, and accompanying drawing shows and can realize specific embodiment of the present invention.It should be noted, can structural change be carried out when not departing from scope of the present invention, thus adopt other embodiments.

It should be understood that and authorize the explanation document of announcement and disclosed intellecture property document to be divided into multiple chapters and sections.Each chapters and sections are necessary for the complete application of submission, and each chapters and sections have respective purposes.Here each chapters and sections of rudimentary knowledge property right will do not discussed in detail.But, based on disclosed object, need the different chapters and sections identifying patent (example as intellecture property document).For majority of case, each patented claim comprises title, priority date, summary, background technology, summary of the invention, Brief Description Of Drawings (if any), Figure of description (if any), embodiment and claim.

Based on the object of retrieval, in patent field, have employed different retrieval classifications.Such as, it is relevant to the term in claim that infringement and/or product evade index infringement, and the claim thus substantially comprised to literature collection is relevant.Effective and/or invalid retrieval is relevant to any known prior art, thus needs the priority date identifying patent documentation.When inventor submit to want to determine its novelty of an invention before or after patented claim time, inventor or its procurator or representative can adopt novelty search.This retrieval can not focus on claim, and pays close attention to the embodiment of invention.Therefore, as described herein, each retrieval is to the different chapters and sections specified weight of the patent documentation in literature collection.

Fig. 1 is expressed as to generate one or more concise and to the point and flow process Figure 100 to the process that the chapters and sections of patent documentation identify.For U.S.Patent & Trademark Office, according to present practice principle, each Patent Application Publication submitted to U.S.Patent & Trademark Office comprises following chapters and sections: the detailed description of title, background technology, summary of the invention, Brief Description Of Drawings, accompanying drawing, preferred implementation, claims and summary, wherein background technology comprises the explanation of technical field and prior art.In one embodiment, not all patent documentation all comprises accompanying drawing, such as, and Chemistry Literature or some foreign patents and patent documentation.Similarly, other countries and area Patent Office and before internal practice in, patent documentation may have the chapters and sections of varying number, or these chapters and sections may occur with different order.Therefore, for inquiry, before the one or more chapters and sections specified weight to the patent documentation in set, need to identify the tissue order of the source of document in set, the different chapters and sections of document and chapters and sections.

First, (step 102) is collected and indexs in patent documentation set.It should be understood that in the art, patent and patent publications are made up of multiple chapters and sections.After document is collected, identify each chapters and sections (step 104) of each patent in literature collection.By variable N _totalbe appointed as the quantity (step 106) of chapters and sections in patent documentation.For meeting different retrieval needs, generate different concise and to the point.By the various combination specified weight of the chapters and sections to patent documentation, and/or not paying attention to by one or more chapters and sections of document being ignored during retrieving, generate concise and to the point, wherein above-mentionedly to ignore by specifying 0 value to realize to these chapters and sections.In order to realize based on concise and to the point carried out retrieval, at least generate one briefly.But, in one embodiment, in order to realize concise and to the point selection to meet the needs of specific retrieval, generate briefly multiple.Once identify the chapters and sections of patent documentation in step 106, initialization to briefly identify relevant counting variable X, and be assigned therein as integer 1 (step 108), then the counting variable N relevant to the chapters and sections of patent documentation is appointed as integer 1 (step 110).From the chapters and sections of patent documentation set _nstart, judge whether chapters and sections _nconcise and to the point (concise and to the point as what generating _x) a part (step 112).If the judged result in step 112 is affirmative, then by chapters and sections _njoin concise and to the point _xin (step 114).At selection chapters and sections _nwhen, to chapters and sections _nspecify sovereignty heavy (step 116).Sovereignty are heavily digital values, and it is for representing compared with other chapters and sections of patent documentation set, chapters and sections _nfor briefly _ximportance, other chapters and sections above-mentioned comprise any previously selected chapters and sections and to be added in concise and to the point or treat the chapters and sections ignored from concise and to the point.If the judged result after step 116 or in step 112 is negative, then increase progressively the variable N (step 118) relevant to the chapters and sections of patent documentation.Judge that all of compiled and indexed set Patent Literature have identified whether chapters and sections are assessed, to be added briefly by these chapters and sections subsequently _xor from briefly _xin ignore (step 120).If the judged result in step 120 is affirmative, then terminate concise and to the point _xconcise and to the point generative process (step 122).On the contrary, if the judged result in step 120 is negative, then next step 112 is turned back to, for briefly _xconsider all the other chapters and sections in set.Subsequently, judge whether also for literature collection generates any other concise and to the point (step 124).If the judged result in step 124 is affirmative, then following incremental count variable X (step 126), and turn back to step 110.On the contrary, if the judged result in step 124 is negative, stop concise and to the point _xgeneration, and the numerical value of X is assigned to variable X _total(step 128).Therefore, can for patent documentation set generate briefly one or more, and each briefly in patent documentation set one or more through identify chapters and sections specify weight.

As shown in Figure 1, can generate briefly one or more, for emphasizing or weaken the producing level of chapters and sections selected by patent documentation in retrieving.Fig. 2 is the flow process Figure 200 representing the concise and to the point increase additional weight value generated to each.More specifically, the quantity of the matched character string in concise and to the point based on each in selected chapters and sections, can join additional weight (secondary weight) in weighted value or from weighted value and deduct.As shown in Figure 2, variable X _totalbe designated as representing generated concise and to the point quantity (step 202), and counting variable X is appointed as integer 1 (step 204).After this, as shown in Figure 2, variable Y _totalbe designated as representing concise and to the point _xmiddle appointment is had the right the quantity (step 206) of heavy chapters and sections.In order to assess each concise and to the point chapters and sections, counting variable Y is appointed as integer 1 (step 208).Subsequently, judge whether secondary weight to join briefly _xchapters and sections _yin (step 210).If the judged result in step 210 is negative, then next jump to step 230 with assess this concise and to the point in next chapters and sections (if these chapters and sections exist).On the contrary, if the judged result in step 210 is affirmative, then next inquire about whether the secondary weight judging to specify is hierarchy (step 212) by secondary.More specifically, each briefly can comprise graduate weighted value, and this depends on the quantity of the data character String matching returned during briefly being retrieved by selection.If the judged result in step 212 is negative, then next set the minimum threshold of the data character String matching that must return, with to chapters and sections _yspecify time weight (step 214).After step 214, for briefly _xchapters and sections _ysetting time weighted value (step 216).The input at step 214 and 216 places is used for setup parameter, and this parameter meets secondary weighting structure generated in the step 212.Therefore, for each concise and to the point chapters and sections, time weighted value can be set, for when exceeding matching threshold to result for retrieval specified weight.

Except setting single weighted value, chapters and sections selected by concise and to the point each can be configured to have graduate weight threshold.If the judged result in step 212 is affirmative, then next will specify concise and to the point _xchapters and sections _ythe quantity of graduation threshold value be appointed as variable Z _total(step 218), layering counting variable Z is set as integer 1 (step 220).After step 220, set the minimum threshold (step 222) of data character String matching that must return, with to briefly _xchapters and sections _ylayering _zspecify time weight, thus be concise and to the point _xchapters and sections _ylayering _zspecify time weighted value (step 224).Once be selected layering _zspecify weighted value, then increase progressively layering counting variable Z (step 226), then judge whether to be concise and to the point _xchapters and sections _yall layerings specify weighted value (step 228).If the judged result in step 228 is negative, then next return step 222.On the contrary, if the judged result in step 228 is affirmative or after step 116, then following incremental count variable Y, to assess (step 230) selected next chapters and sections briefly.Subsequently, judge whether to have carried out the assessment (step 232) of specifying of weight threshold by different level to selected all chapters and sections briefly.If the judged result in step 232 is negative, then next turn back to step 210, and if judged result in step 232 is affirmative, then next increase progressively concise and to the point counting variable X (step 234).After step 234, judge whether to all assessments (step 236) of specifying briefly having carried out secondary weight generated.If the judged result in step 236 is negative, then next turn back to step 206, if and the judged result in step 236 is affirmative, then next stop to generating concise and to the point selected by chapters and sections specify graduate weight threshold (step 238).Therefore, each briefly can have graduate weight, for chapters and sections selected by concise and to the point to each and concise and to the point in the quantity specified weight of matched character string.

As shown in Figure 2, the secondary weight of graduation (i.e. layering) can be applicable to each concise and to the point independent chapters and sections, wherein, each weight be number of matches between data in inquiry string and the literature collection of just being resolved one or more threshold values based on.In another embodiment, as shown in Figure 3, secondary weight can reflect the position that string matching occurs in one or more concise and to the point chapters and sections.This time weight can be separated from each other with the secondary weight shown in Fig. 2, also can add in the secondary weight shown in Fig. 2.As shown in Figure 3, variable X _totalbe appointed as representing generated concise and to the point quantity (step 302), and counting variable X is appointed as integer 1 (step 304).After this, variable Y _totalbe appointed as representing concise and to the point _xin appointment have the quantity (step 306) of the chapters and sections of weight, and counting variable Y is appointed as integer 1 (step 308).Judge whether subsequently secondary weight to be joined briefly _xchapters and sections _yin (step 310).If the judged result in step 310 is affirmative, then next will be concise and to the point _xchapters and sections _ybe divided into multiple sub-chapters and sections (step 312).Division in step 312 can adopt different embodiments.Such as, in one embodiment, can be divided into three sub-chapters and sections, wherein the first sub-chapters and sections are defined as first sentence, and the 3rd sub-chapters and sections are defined as last sentence, and the second sub-chapters and sections are defined as in first and the 3rd all data between sub-chapters and sections.Similarly, in another embodiment, briefly _xchapters and sections _ymultiple chapters and sections can be divided into, the length of each chapters and sections with briefly whole _xchapters and sections _yin proportion relevant.No matter adopt which kind of method to determine the quantity of sub-chapters and sections, briefly _xeach chapters and sections _ycan be divided into plural sub-chapters and sections, the secondary weight of wherein specifying to sub-chapters and sections is not only concise and to the point for reflecting _xchapters and sections _yin matched character string, and for reflecting the position of above-mentioned coupling in selected sub-chapters and sections.

After step 312, variable Z _totalbe appointed as concise and to the point _xchapters and sections _ythe quantity (step 314) of middle generated sub-chapters and sections, and counting variable Z is appointed as integer 1 (step 316).To briefly _xchapters and sections _ysub-chapters and sections _zspecify time weight (step 318).After appointment in step 318, incremental count variable Z (step 320), then, judges briefly _xchapters and sections _yin whether also there is the sub-chapters and sections (step 322) of not carrying out the assessment that time weight is specified.If the judged result in step 322 is negative, then next turn back to step 318.On the contrary, if the judged result in step 322 is judged result in affirmative or step 310 is negative, then following incremental count variable Y (step 324).Subsequently, judge briefly _xin whether there are the chapters and sections (step 326) of the appointment assessment not carrying out time weight.If the judged result in step 326 is negative, then next turn back to step 310.On the contrary, if the judged result in step 326 is affirmative, then following incremental count variable X (step 328), and judge whether to all appointments assessment (step 330) briefly carrying out secondary weight.If the judged result in step 330 is negative, then next turn back to step 306, and if certainly, then stop the assignment procedure of time weight.Therefore, multiple sub-chapters and sections can being divided into based on the physical location of concise and to the point chapters and sections, wherein secondary weight being assigned to one or more sub-chapters and sections through identifying.

In Fig. 1 to Fig. 3, in order to retrieve the patent documentation of matched character string combination and each be had to the different chapters and sections specified weight of the document of coupling, formation base is concise and to the point.Based on the documentation with matched character string combination, also can adopt briefly secondary.More specifically, before showing result to searcher, consider based on secondary, can adopt secondary briefly to this result appointment time weight.For use time weight, can utilize the different characteristic of patent documentation, this includes but are not limited to priority date and/or publication date.In patent field, priority date represents the earliest time in patent families.More specifically, when submitting the patented claim describing certain invention in detail first to for priority date is set up in this invention.The publication date of patent documentation (a patent document) represents the authorization date of authorizing issued patents, and the publication date of patent publications (a patent publication) represents the publication date of unsettled patented claim.Secondaryly briefly can utilize in these record dates date or whole dates and generating.

Fig. 4 represents the process flow diagram (step 400) generating secondary concise and to the point process, this secondary concise and to the point based on the date key element relevant to the document date that submit Query returns to result for retrieval specified weight.In one embodiment, date key element can include but are not limited to publication date, the applying date and foreign priority day.First, set secondary concise and to the point " secondary-concise and to the point " (step 402).The quantity of document that submit Query obtains is appointed as variable N _total(step 404), and counting variable N is appointed as integer 1 (step 406).For the document returned in archives _n, its priority date is retrieved (step 408), following incremental count variable N (step 410).Then, the retrieval (step 412) of secondary-concise and to the point key element completing returned archives is judged whether.If the judged result in step 412 is negative, then next return step 408.On the contrary, if the judged result in step 412 is affirmative, then following secondary-concise and to the point key element based on extracting performs sorting algorithm, thus the document (step 414) in this result for retrieval of classifying.Sorting algorithm can adopt multiple different form, and therefore, the present invention will be not limited to any specific sorting algorithm.Once complete to the document classification in set, variable document _oLDbe appointed as in archives a document (step 416) with secondary-concise and to the point date the earliest, and variable document _nEWbe appointed as in archives a document (step 418) with up-to-date secondary-concise and to the point date.Variable date-scope is defined as variable document _oLDand document _nEWdifference (step 420), and by date-scope is divided into multiple chapters and sections (step 422).Different embodiment Division Dates-scope can be adopted in step 422.Such as, in one embodiment, can divide three sub-chapters and sections, wherein the first sub-chapters and sections are defined as closest to document _nEWthe document of relevant date, the 3rd sub-chapters and sections are defined as closest to document _oLDthe document of relevant date, and the second sub-chapters and sections are defined as the date at first and the 3rd all documents between sub-chapters and sections.Similarly, in another embodiment, date-scope can be divided into multiple chapters and sections, and wherein each chapters and sections have the Document metrology in equal document set.Therefore, no matter adopt which kind of method, can specify time weight to the sub-chapters and sections of each of archives, this weight has the relevance of the Query Result based on this weight.

Based at least one secondary data standard to after the document classification in Query Result, variable Z _totalbe appointed as date-the chapters and sections quantity (step 424) of scope, and counting variable Z is appointed as integer 1 (step 426).To date-scope _zspecified weight (step 428), next increases progressively variable Z (step 430).After step 430, judge whether to be each sub-chapters and sections specified weight (step 432).If the judged result of step 432 is negatives, then next return step 428.On the contrary, if the judged result of step 432 is affirmatives, then stop as each sub-chapters and sections specified weight generated.Therefore, in order to secondary key element can be given prominence to further before display data, can generate secondary concise and to the point next to result set appointment time weight.

Query Result to literature collection is applied secondary key element and is not limited as the date.Fig. 5 represents that this result set does not adopt the relevant date of any one patent documentation to the secondary concise and to the point process flow diagram (step 500) of result set application.When starting retrieval, select one or more literature collection to realize inquiring about (step 502).In one embodiment, literature collection can be the form of intellecture property literature collection.Similarly, in one embodiment, literature collection can be the form of country variant, such as, by the set of the publication such as U.S.Patent & Trademark Office, Japan Office, EUROPEAN PATENT OFFICE.Once select document set, then have selected concise and to the point (step 504) for retrieving.Concise and to the point embodiment is illustrated in Fig. 1 to Fig. 3 above.Based on completing of selecting in step 502 and 504, open inquiry, and inquiry is committed to this concise and to the point and selected literature collection (step 506).In one embodiment, inquiry is character string inquiry.Determine the quantity that the document of query event at least one times occurs in gathering, this quantity is appointed as variable X _total(step 508), and the counting variable X mating document is appointed as integer 1 (step 510).In addition, variable N _totalbe appointed as the concise and to the point chapters and sections quantity (step 512) selected by submit Query, and selected concise and to the point counting variable N is appointed as integer 1 (step 514).For each document _xchapters and sections _ncalculate scoring.In one embodiment, score calculation becomes chapters and sections _nthe number of middle match query with to chapters and sections _nthe product (step 516) of the number of the point of specifying.In one embodiment, chapters and sections are specified in _npoint represent the significance level of this particular chapter in set.

After step 516, increase progressively variable N (step 518), next judge whether these concise and to the point all chapters and sections evaluated (step 520).If the judged result of step 520 is negatives, then return step 516.On the contrary, if the judged result of step 520 is affirmatives, then variable X (step 522) is increased progressively.Then, all documents evaluated (step 524) under this quantity are judged whether.If the judged result of step 524 is affirmatives, then stop carrying out fraction assessment (step 526) to returned document.On the contrary, if the judged result of step 524 is negatives, then returns step 514, by concise and to the point chapters and sections, next document is marked.

Once to all documents and selected concise and to the point appointment scoring, then brief calculation overall score (step 526) selected by returning for each document and according to submit Query.As shown in Figure 5, in compilation, each document all has scoring, this scoring be based on string matching quantity and concise and to the point in the numeric form of associated weight of specifying.

Should be understood that in the process performing patent retrieval, importantly judge which result for retrieval is more relevant.Such as, except marking except the contribution key element of match query, also utilize scoring to coupling document given level.This grade represents: compared with other return document, and the document which returns is rated as more relevant.This ranking utilizes different key element, and these key elements can comprise the combination based on the grade of marking and/or grade and secondary key element.

Fig. 6 represents the process flow diagram 600 to the process that the document returned by inquiry is classified, and this classification returns the appointment scoring of document and each concise and to the point chapters and sections based on each.As Fig. 6 calculate, variable X _totalbe appointed as the total quantity (step 602) of carrying out the document that submit Query at least one times returns.Then, follow sorting algorithm and carry out classified documents (step 604).In one embodiment, can from the highest scoring to minimum scoring, or from minimum scoring to the series classification document of the highest scoring.Sorting algorithm can adopt multiple different form, and therefore, the present invention will be not limited to any specific sorting algorithm.Once complete to the document classification in whole set, also can classify to selected each the concise and to the point literature collection of inquiry, thus in each chapters and sections, produce a class document.In one embodiment, the assorting process returning document can regard the ranking process of appraisal result as.Variable N _totalrepresent the quantity (step 606) of the concise and to the point middle chapters and sections selected by retrieval.Chapters and sections counting variable N is initialized as integer 1 (step 608), and document counting variable X is initialized as variable 1 (step 610).For chapters and sections _n, correspond to each document inquiring about input at least one times _xbe divided into the first document _xto last document _xTotal(step 612).Once to chapters and sections _nclassify complete, then increase progressively variable N (step 614), next judge whether to carry out classification assessment (step 616) to selected all chapters and sections briefly.If the judged result of step 616 is negatives, then next return step 612.On the contrary, if the judged result of step 616 is affirmatives, then represents and by selected all chapters and sections briefly, all documents are classified.Therefore, the classification of Query Result performs in two ranks, and first level is the inquiry to its entirety, and second level by form this concise and to the point selected by the classification of chapters and sections.

Once the classification of literature collection is complete, then different instrument is taked to transmit classified inquiry result.More specifically, after completing the inquiry of Query Result and classification, the data sending inquiry submitter to are the result relevances based on whole inquiry, and/or based on the relevance of concise and to the point each chapters and sections of the inquiry through submitting to.Fig. 7 be represent to return and the process flow diagram 700 of process of result for retrieval specified associations through classification.This quantity returning the layering of result for retrieval is appointed as variable T _total(step 702).In one embodiment, variable T _totalit is static variable.But, in another embodiment, variable T _totalit can be dynamic variable.Relevance assessment can perform in two ranks, and first level is based on all documents in Query Result, and the assessment of second level is concise and to the point based on each of literature collection.Variable X _totalrepresent all documents (step 704) returned according to inquiry institute and through classifying, and variable X _totalbe layered quantity T _totalremove, to calculate the quantity QS (step 706) of the Query Result being assigned to each layering (T represents).For Query Result is assigned to layering _t, layering counting variable T is initialized as integer 1 (step 708), and counting variable X represents the document being assigned to layering, and is initialized as integer 1 (step 710).After step 708 and 710 carry out initialization, by document _xbe assigned to layering _t(step 712).After step 712 carries out above-mentioned appointment, increase progressively variable X (step 714), then judge whether layering _tfill up Query Result (step 716).If the judged result of step 716 is negatives, then next return step 712.On the contrary, if the judged result of step 716 is affirmatives, then represents and complete layering _tquery Result specify.Then, increase progressively variable T (step 718), next judged whether that all Query Results to layering specify (step 720).If the judged result of step 720 is negatives, then next return step 710.On the contrary, if the judged result of step 720 is affirmatives, then stop the layering given query result to generating.It should be noted that can from being related to most the most incoherent mode classification from top to down to the sorted table of layering given query result, or from be least related to maximally related mode to lower and Shangdi the sorted table to layering given query result.Similarly, in one embodiment, in classification and graduate result, occurring flex point, and adjacent layering is divided at this flex point place.Therefore, in order to the outstanding selected relevance through classified documents, to layering given query result.

As mentioned above, the concise and to the point of submit Query can not be considered, with rough principle classified inquiry result.But, also can carry out layering appointment according to concise and to the point principle (profile basis) (also referred to as refinement principle).More specifically, can with reference to concise and to the point feature, the relevance order according to returned document is briefly classified to each in submit Query.In order to realize concise and to the point utilization, can the model split shown in Fig. 7 be also layering to each archives of briefly specifying.The concise and to the point layering of this refinement is specified and Query Result can be made to transmit further based on concise and to the point feature.

As mentioned above, each patent in literature collection Query Result can be resolved, to provide the display of result based on relevance.In one embodiment, this result can be shown emphasize to emphasize the matched data value returned in the appointment chapters and sections of compilation of intellecture property document or to remove.Similarly, in one embodiment, based on relevance, the display of result for retrieval is limited.With regard to regard to Query Result appointment layering, only can view the layering of selection, the layering of wherein this selection can be that those are considered to comprise the layering of more relevant query result.Similarly, with regard to Query Result scoring, the result that restriction makes only to show submitted to inquiry in those regulation marks can be set.The display restriction of Query Result should not be confined to embodiment described herein, can adopt other forms that the Query Result viewed is defined as the result that those only have certain relevance score.

In one embodiment, the Query Result through classifying is shown as the compilation of pertinent literature statically.But, in another embodiment, based on the classification of returned document, can the returning of dynamic limitation literature collection.This dynamic aspect support changes correlation criterion to reflect Query Result.Fig. 8 is the process flow diagram 800 representing an embodiment of the display of Query Result being carried out to dynamic limitation.As mentioned above, based on the relevance quantitative factor of submit Query key element, to each document classification (step 802) returned according to inquiry.Based on this numeric data (that is, to each return the quantitative factor of the relevance that document is specified), to compilation application curves fitting routine (step 804) of returned document.This curve fitting routine calculates the theory function of assemble data.More specifically, this curve fitting routine determines this theory function based on the raw value factor of relevance.Based on this curve fitting routine, some documents in compilation and the curves of this theory function or close to this theory function curve (step 806).Calculate one or more derivatives (step 808) of theoretical function.In order to dynamically limit the result of compilation, the quantity (step 810) of Choice Theory function derivative.More specifically, in order to the result of compilation being defined as most pertinent literature, this Dynamic Selection will be limited in the first order derivative of the theory function of this curve fitting routine.Similarly, obtain greater amount document to expand the result of compilation, this Dynamic Selection will extend to second derivative (or more high-order).Based on the quantity of selected derivative, the documentation in derivative range of choice is returned (step 812).Therefore, based on the degree of approximation of the theory function of document and curve fitting routine, dynamic conditioning is carried out to the compilation returning document.

Dynamic selection processes shown in Fig. 8 and instrument represent the embodiment limiting compiled results.In another embodiment, utilize graphic user interface as the upper strata (veneer) of source code, thus realize user to the mutual of the conventional result of the compilation through classification and amendment.Fig. 9 represents the process flow diagram by using graphic user interface dynamically to limit the process of the result of compilation.As mentioned above, based on the relevance quantitative factor of submit Query key element, classify (step 902) to according to inquiring about each document returned.Result for retrieval is depicted as figure (904).Multi-form figure can be adopted.In one embodiment, this figure can be X-Y scheme form, be wherein an axle with returned quantity of document, and the quantitative factor of correlativity is another axle.Interface provides the mechanism (step 906) that quantity of document can be made to be defined in selected relevance numerical value.In one embodiment, provide vernier (slider) on a user interface, and by pointing tool (pointing tool), this vernier can move to any relating value (step 908) of figure.Based on the movement of vernier, dynamically change the quantity of pertinent literature and be considered to relevant specific document.More specifically, this vernier plays marginal effect, wherein specifying has all documents of the relevance above this slider position be identified as pertinent literature and returned (step 910), specifies and has all documents of the relevance below this slider position then not return (step 912).In one embodiment, appointment has all documents of the relevance of slider position be identified as pertinent literature and return.On the contrary, in one embodiment, appointment has all documents of the relevance of slider position be identified as uncorrelated and do not return.Therefore, the vernier of this graphic user interface can be moved, be identified as relevant document to adjust and returned in documentation.

As shown in Figure 9, graphic user interface can be adopted provide the instrument being convenient to pertinent literature Dynamic Selection.Figure 10 is the block diagram 1000 of presentation graphic user interface embodiment.More specifically, computer system 1000 has processing unit 1002, and described processing unit 1002 is connected to storer 1006 by bus structure 1008.In one embodiment, although only illustrate a processing unit 1002, more processing unit can be provided in expansion design.Illustrated system 1000 communicates to connect with the storage medium 1040 being configured to store literature collection 1042.In one embodiment, electronic literature set comprises the compilation of patent documentation, and this patent documentation comprises granted patent and disclosed patented claim.Storage medium 1040 and processing unit 1002 communicate to connect.In addition, this system illustrated and visual display unit 1050 communicate to connect, with display of visually data.Input equipment 1052 is adopted to communicate with visual display unit 1050.Can adopt multiple multi-form input equipment, it includes but not limited to keyboard, mouse, tracking ball, electronic pen etc.The scoring of relevance as calculated of carrying based on single result and the quantity of document forming this compilation, visual display unit 1050 provides graphic user interface 1054, to transmit the figure display of the compilation of Query Result.In one embodiment, graphical interfaces user 1054 is used as the upper strata of the source code run in processing unit 1002.The figure mechanism 1060 by input equipment access is provided, to realize the Dynamic Selection of the subset of Query Result in this graphic user interface.In one embodiment, figure mechanism 1060 is forms of vernier, the separatrix in the figure that vernier represents Query Result represents.Move along with figure mechanism 1060 represents in scope at figure, revise and will fall into the ad hoc inquiry result of compilation.In one embodiment, select to quote and be contained in Query Result from figure mechanism all documents of 1060 sides and/or all documents fallen in this figure mechanism, and get rid of all documents quoted from figure mechanism 1060 opposite sides.Therefore, the figure mechanism 1060 of this graphic user interface applies to the compilation of Query Result the instrument on-the-fly modified.

The process as shown in Figures 1 to 9 of employing and/or instruction to literature collection submit Query, and in response to this set of this query parse.But the present invention should not be limited to process or instruction set, in one embodiment, the hardware element communicated to connect with literature collection can be comprised.Figure 11 represents based on the concise and to the point submission of the retrieval of submit Query by Query Result classification and the block diagram 1100 of one group of instrument resolving to one or more layering, comprising the different chapters and sections specified weight to the intellecture property document of this retrieval briefly through identifying.As shown in the figure, computer system 1102 has processing unit 1104, and described processing unit 1104 is connected to storer 1106 with bus structure 1108.In one embodiment, although only illustrate a processing unit 1104, more processing unit can be provided in expansion design.Illustrated system 1102 communicates to connect with the storage medium 1140 being configured to store literature collection 1142.In one embodiment, electronic literature set comprises the compilation of patent documentation, and this patent documentation comprises granted patent and disclosed patented claim.Storage medium 1140 and processing unit 1104 communicate to connect.In addition, illustrated system and visual display unit 1150 communicate to connect, with display of visually data.The various piece support herein illustrated and illustrate is committed to the inquiry of literature collection 1142.

It is local that controller 1160 is arranged on computer system 1102, and communicate to connect with storer 1106 and processing unit 1104.Controller 1160 is responsible for collecting to literature collection 1142 and indexing.Controller 1160 and documentation management device 1162 communicate to connect, and documentation management device 1162 is for identifying each chapters and sections of each document in set.As mentioned above, when patent documentation set, each patent or disclosed patented claim are made up of chapters and sections that are specific, unity of form.But not all patent documentation set all has unified form (layout).Therefore, documentation management device 1162 is for identifying the chapters and sections of set Literature, and in one embodiment, documentation management device 1162 is for identifying the DISPLAY ORDER through identifying chapters and sections.Concise and to the point manager 1164 is arranged to communicate to connect with documentation management device 1162.Concise and to the point manager 1164 is that literature collection 1142 sets up retrieval briefly.More specifically, concise and to the point manager 1164 contributes to selecting one or more chapters and sections of document, and selected by each chapters and sections specified weight.Wherein, selected chapters and sections are the chapters and sections identified by documentation management device 1162 comprised in inquiry.In one embodiment, weight is digital value, to represent the importance of matched data in selected chapters and sections.Therefore, the retrieval set up by concise and to the point manager 1164 briefly provides the summary of chapters and sections associated with the query in literature collection.

Inquiry manager 1166 and concise and to the point manager 1164 communicate to connect, and inquiry manager 1166 is arranged on computer system 1102 this locality, and communicate to connect with storer 1106.Inquiry manager 1166 is responsible for by selecting at least one retrieval concise and to the point to literature collection 1142 submit Query.More specifically, the data in data query and literature collection 1142 chapters and sections compare by inquiry manager 1166, and these chapters and sections identify and have specified weight in concise and to the point.Inquiry manager 1166 and relevance omniselector 1168 communicate to connect.This relevance omniselector is used for marking as the document rating in compilation based on relevance, and dynamically limits the result in this compilation based on this grade.By inquiry manager 1166 carry out relatively and be combined relevance omniselector 1168, generate the compilation of the Patents document based on applied dynamic limitation.In one embodiment, this compilation is presented on visual display unit 1150.Similarly, in one embodiment, this compilation can be kept in volatibility or permanent memory.Conveniently to the transmission of inquiry submitter, inquiry manager and class manager communicate to connect, with the grade based on document classification evaluation submit Query result.

In one embodiment, controller 1160, documentation management device 1162, concise and to the point manager 1164 and inquiry manager 1166 can be arranged in the storer 1106 of computer system 1102 this locality.But, the invention is not restricted to this embodiment.Such as, in one embodiment, controller, documentation management device, concise and to the point manager and inquiry manager 1160-1166 each can be used as the outside that hardware tools resides in local storage 1106, or they can implement in the combination of hardware and software.Similarly, in one embodiment, controller and manager 1160-1166 can reside in the remote system communicated to connect with storage medium 1140.Therefore, controller and controller may be embodied as Software tool or hardware tools, submit one or more inquiry to, to generate the compilation of Patents document for support to electronic literature set.

In one embodiment, the present invention is with implement software, and described software includes but not limited to firmware, resident software, microcode etc.The present invention can be the form of computer program by computer usable medium or computer-readable medium access, this computer usable medium or computer-readable medium provide the program code used by computing machine or any instruction execution system, or provide the program code communicated to connect with computing machine or any instruction execution system.For purposes of illustration, computer usable medium or computer-readable medium can be can hold, store, communicate, propagate or the device of transmission procedure, and said procedure is used by instruction execution system, equipment or device or and instruction executive system, equipment or device communicate to connect.

Embodiment in the scope of the invention also comprises the product of manufacture, and this product comprises the program storage device wherein with encoded program code.This program storage device can be any available medium by universal or special computer access.For example, this program storage device can include but not limited to RAM, ROM, EEPROM, CD-ROM or other optical disc memorys, magnetic disk memory or other magnetic storage apparatus or can be used for storing expect that program code means also can by any other medium of universal or special computer access.The combination of said apparatus also should be included in the scope of this program storage device.

Above-mentioned medium can be electronic system, magnetic system, optical system, electromagnetic system, infrared system or semiconductor system (or equipment or device).The example of computer-readable medium comprises semiconductor or solid-state memory, tape, moveable computer format floppy, random access memory (RAM), ROM (read-only memory) (ROM), hard disk and CD.Current example of optical disks comprises read-only compact disk B (CD-ROM), read/write compact disk B (CD-R/W) and DVD.

The data handling system being applicable to storage and/or executive routine code at least comprises one is connected to memory element directly or indirectly processor by system bus.Memory element can be included in the local storage, mass storage and the buffer memory that use when program code is actual to be performed.This buffer memory stores at least some program code temporarily, thus can reduce in the process of implementation replace synchronous codes number from mass storage.

I/O or I/O equipment (including but not limited to keyboard, display, pointing device etc.) directly or by middle I/O controller can be coupled to system.Network adapter also can be coupled to system, to make data handling system can be coupled by middle individual or public network and other data handling system or remote printer or memory device.

Software tool can be the form of the computer program can accessed by computer usable medium or computer-readable medium, and this computer usable medium or computer-readable medium are for providing program code that is that used by computing machine or any instruction execution system or that communicate to connect with computing machine or any instruction execution system.

the advantage of hinge structure

Known, each intellecture property document has the chapters and sections for meeting the restriction general picture that legal submission condition requires in the art.Generate briefly one or more, to contribute to literature collection submit Query.Each briefly applies weight to the one or more chapters and sections through identifying in document.Weight represents the importance of the chapters and sections through identifying, and applies numerical value to each document returned in compilation.Not all retrieval is not always the case.Such as, have recognized that, even if the intellecture property document in chemical field is when having accompanying drawing, also only to there is a limited number of accompanying drawing.Therefore, the inquiry in chemical field can remove emphasizing accompanying drawing, and improves emphasizing penman text.Different retrievals is submitted to, to obtain different results to set.Therefore, generate briefly multiple, and each briefly selects different from identifying chapters and sections, and specify different weight to chapters and sections selected by different, make it possible to efficiently and submit Query effectively, to produce paid close attention to documentation result.

Once generate concise and to the point and select at least one briefly for submit Query, then next step shows Query Result by mode briefly arranged side by side selected by making.In one embodiment, this inquiry produces documentation, and this compilation is then classified and is placed in the layering of graduation classification.This makes directly to show relevance while display Query Result.In another embodiment, based on chapters and sections selected in concise and to the point, Query Result can be transmitted further, wherein based on the classification of the document in represented single chapters and sections in concise and to the point and each chapters and sections, show the second group polling result.Therefore, utilize concise and to the point selection to carry out both generated query results, show Query Result based on relevance in mode briefly arranged side by side selected by making again.

embodiment

Although should be appreciated that for purposes of illustration and herein illustrating specific embodiments more of the present invention, various amendment can be done under prerequisite without departing from the spirit and scope of the present invention.Especially, there is multi-form intellecture property document, comprise patent, trade mark and literary property.In the classification of patent documentation, can classify further to document, comprise granted patent, disclosed patented claim, abridgments of specifications and utility model registration.Some documents in these documents can comprise the chapters and sections sorted with same sequence of equal number, and some other document can comprise the chapters and sections of varying number and/or different orders.Generate independently briefly based on included chapters and sections, and these chapters and sections putting in order in base document need not be concerned about.

In addition, although specifically understand the electronic literature set (comprise granted patent and disclosed patented claim, trade mark registration and application and literary property register and apply for) relevant to intellecture property document, but the present invention should not be limited to these specific classification of electronic literature.In one embodiment, electronic literature set can comprise the document of any type with limited multiple chapters and sections.Document is resolved to limited chapters and sections by enabling supvr by this, has the briefly multiple of respective weights for one or more limited chapters and sections create, and to having selected concise and to the point literature collection submit Query.As mentioned above, dynamic corrections can be carried out to the concise and to the point selection of inquiry.In one embodiment, the order that can change the relevance that the document that returns in compilation and document present in compilation while query contents to the concise and to the point correction of inquiry is being kept.Therefore, protection scope of the present invention is only by appended claim and equivalents thereof.

Claims (11)

1. the method for the result for retrieval specified associations to electronic literature set performed by computing machine, it comprises:

Collect to the document in the set of intellecture property document and index, in described set, described in each, document has multiple chapters and sections;

Identify the described chapters and sections in described set in document described in each;

For described literature collection sets up retrieval briefly, wherein, described retrieval briefly comprises at least one chapters and sections through identification of selection;

In set up retrieval is concise and to the point, to the chapters and sections specified weight through identifying selected by each;

When inquiry, to patent documentation set submit Query, described inquiry comprises selects at least one retrieval concise and to the point, and data query and the data being used for the chapters and sections represented in selected retrieval is briefly compared;

To in documentation each document compute associations scoring, described documentation be in response to described submit Query generate;

Based on described relevance scoring as calculated, to the document rating in described documentation;

Based on described grade, to the document application dynamic limitation in described compilation;

Based on applied dynamic limitation, the document in described compilation is classified, to generate the compilation through classification; And

To select from the described chapters and sections through identifying through identifying chapters and sections application time weight, to revise the weight being assigned to these chapters and sections through identifying based on the quantity of the searching character string matched with these chapters and sections through identifying.

2. method according to claim 1, comprises further, generates the subset of the compilation through classification, and is that described subset calculates the second relevance scoring based on the concise and to the point middle secondary standard occurred of described retrieval.

3. method according to claim 2, comprises further:

I () is classified to described subset;

(ii) to the dynamically specified associations restriction of the described subset through classification, and

(iii) based on described relevance restriction of specifying, returning of Query Result is limited.

4. method according to claim 1, comprises further, generates the described figure of compilation through classification represent and the quantity of the document represented of marking with different relevance in described compilation based on described relevance scoring as calculated.

5. method according to claim 1, comprises further, and to the described compilation application curves fitting routine through classification, wherein said routine calculates the theory function of the data of the described compilation through classification, and calculates at least one derivative of described function.

6. method according to claim 5, wherein, comprises the step of the document application dynamic limitation in described compilation: select described function derivative, and returns the data fallen within the scope of selected described function derivative.

7., for a system for the result for retrieval specified associations to electronic literature set, it comprises:

Processor, itself and storer and storage medium communicate to connect, and wherein, preserve intellecture property literature collection on said storage, and described in each in described set, document has multiple chapters and sections;

Controller, it is connected with described processor communication, and collects to the document in described intellecture property literature collection and index;

Documentation management device, itself and described controller communicate to connect, and for identifying the chapters and sections in described set in each document;

Concise and to the point manager, itself and described documentation management device communicate to connect, and be that described literature collection sets up retrieval briefly, wherein, at least one chapters and sections through identifying that described retrieval briefly comprises selection, described concise and to the point manager further in set up retrieval is concise and to the point to the chapters and sections specified weight through identifying selected by each;

Inquiry manager, its inquire about time to described literature collection submit Query, described inquiry comprises selects at least one retrieval briefly and by data query and the data being used for the document chapters and sections represented in selected retrieval briefly to compare, described inquiry makes to submit in response to described inquiry the compilation of pertinent literature producing and returned by described inquiry manager to, and pertinent literature described in each has the match query submitted to at least one data through identifying in chapters and sections with specified weight and relevance scoring; And

Relevance omniselector, itself and described inquiry manager communicate to connect, and based on described relevance scoring to the described document rating in described compilation, and based on the result of described grade to the document application dynamic limitation in described compilation,

Wherein, described storer stores the compilation through classification, and the described compilation through classification generates by classifying to the document in described compilation based on applied dynamic limitation, and

Described concise and to the point manager to select from the described chapters and sections through identifying through identifying chapters and sections application time weight, to revise the weight being assigned to these chapters and sections through identifying based on the quantity of the searching character string matched with these chapters and sections through identifying.

8. system according to claim 7, comprise further: class manager, itself and described relevance omniselector communicate to connect, and based on the second relevance scoring of the subset of the described compilation through classification, described subset is classified, wherein said second relevance scoring determines based on the concise and to the point middle secondary standard occurred of described retrieval.

9. system according to claim 7, comprises: visual display unit further, and it is for showing the described figure of compilation through classification of marking based on described relevance and represent and the quantity of the document represented of marking with different relevance in described compilation.

10. system according to claim 7, wherein, described processor runs the instruction of the curve fitting routine being used for the described compilation through classification, and described curve fitting routine calculates the theory function of the data of the described compilation through classification, and calculates at least one derivative of described function.

11. systems according to claim 10, wherein, described document in the compilation of classification is limited within the scope of the described function derivative of selection by described relevance omniselector, and returns the data fallen within the scope of selected described function derivative.