CN101341483A - User interface searching and displaying legal case histories - Google Patents

Abstract

An exemplary legal research system provides graphical user interfaces to client access devices, with the user interfaces specifically designed to facilitate access to procedural case histories. Each interface is dynamically customized based on the procedural relationships of a plurality of legal cases to provide hyperlinks to corresponding case law documents. The hyperlinks are positioned within indicator band of the interface to indicate corresponding tiers of a court system, but also to mirror the procedural and temporal relationships of legal cases corresponding to the case law documents. Additionally, the exemplary interface includes case validity indicators to signal the precedential value or status of corresponding legal cases. The graphical user interface, in some embodiments, provides analogous access to statutory information.

Description

Be used to search for and show the user interface of legal precedent history

Copyright statement

One or more parts of patent document comprise material protected by copyright.The copyright owner does not oppose that anyone opens up system to patent document or patent disclosure, as its in patent and trademark office patent file or record, occurred like that, in any case still keep all copyrights in other respects.Following statement is applicable to this document: copyright 2005, Thomson Legal andRegulatory, Inc.

Related application

The application requires in the U.S. Patent application No.11/370 of submission on March 6th, 2006,194 right of priority, and it is the U.S. Patent application No.11/182 that submits on July 13rd, 2005,028 continuation.The application is involved in the U.S. Patent application No.09/746 that submitted on Dec 22nd, 2000, and 557, it is the U.S. Patent application No.09/120 that submits on July 21st, 1998,170 continuation.All these applications all are bonded to this with for referencial use.

Technical field

Various embodiment of the present invention relates to and is used to make the user to carry out system, method and the software of (interface) alternately with document, and described document is such as being judicial opinion or legal precedent (case) document with complicated time, semanteme and/or legal relation.

Background

The legal system of the U.S. and more in the world other legal system depend critically upon written judicial opinion (being judge's written statement) and illustrate or explain the law that management resolves a dispute.Each bar judicial opinion is important for solving specific legal dispute not only, and also is very important for not solving or avoid similar dispute or legal precedent.Owing to this reason, the bench and bar in the American legal constantly studies the entity of the suggestion or the case law in the past of always expanding, to find those suggestions or the legal precedent the most relevant with solving new dispute.

For the ease of carrying out described search, Thomson Legaland Regulatory such as the St.Paul of Minnesota, Inc. the judicial opinion of the court in the entire United States scope is collected and announced in the company of (as Thomson West operation) and so on electronic form, and make described judicial opinion can be used for online researcher by its Westlaw research service.Many described suggestions announce that with reference citation or to the hyperlink of historical related opinions described historical related opinions is known as previous legal precedent, and it is from other courts that before all or part of identical dispute had ruling.

Described quoting with hyperlink makes the researcher can be at an easy rate visit related opinions electronically in described Westlaw research system.For example, usually will not only quote suggestion from the suggestion in the patent legal precedent of U.S.'s the last resort (being the last resort of the U.S.) from federal circuit court of appeals (for the inferior high court of patent legal precedent), but also will quote the suggestion of local federal court of the described patent legal precedent of beginning, thereby provide the Documentary Records of this legal precedent by the historical or progress of the United States Federal's judicial system.

Dispute is a particular importance by the historical or progress of court's system, and this is because the follow-up legal precedent in the chain may be overthrown one or more parts of previous legal precedent.Therefore, before the ruling in depending on any given legal precedent, the bench and bar need understand its history.For this purpose, described Westlaw research system comprises a kind of KeyCite feature, and it not only provides the order legal precedent tabulation of the history of any in fact given legal precedent of formation, but also the Short Description to the relation between each legal precedent is provided.For example, the history list with two legal precedents may be listed the motion (motion) that a legal precedent permission proposes in another legal precedent.This tabulation is very powerful instrument.

Yet the present inventor recognizes that some legal precedent history are quite complicated, and are difficult to follow sequential list.For list a legal precedent more than once, present the tabulation that between legal precedent, has the gap or list the history list of a plurality of legal precedents of the given previous legal precedent of influence especially like this.

Therefore, the present inventor recognizes, needs user and the mutual better mode of history-sensitive legal precedent of making especially.

The accompanying drawing summary

Fig. 1 is the block scheme that can be used to generate automatically the computer system of case law reference data in Law Text disposal system according to the present invention;

Fig. 2 A-2D illustrates the snapshot that is displayed to according to the example of the user's of case law cited system of the present invention user interface;

Fig. 3 shows the diagram according to the data flow in the case law cited system of the present invention;

Fig. 4 shows the diagram of the more details of the data flow in the case law cited system of Fig. 3;

Fig. 5 show the part that can be used as described case law cited system according to diagram of quoting recognition system of the present invention;

Fig. 6 shows the process flow diagram that is used to quote the method for identification and proof procedure according to of the present invention;

Fig. 7 shows the process flow diagram that is used to quote the method for checking according to of the present invention;

Fig. 8 shows the diagram according to negative historical deterministic process of the present invention;

Fig. 9 shows the process flow diagram that is used for determining to the method for the processing degree of depth of legal precedent according to of the present invention; And

Figure 10 shows the diagram that is used to distribute the method for subject content classification according to of the present invention.

The description of exemplary embodiment

This instructions with reference to and combine accompanying drawing and combine appendix and appended claims, its description has also illustrated one or more exemplary embodiment of the present invention.These embodiment are not only to be in order to carry out illustration and instruction in order to limit, and these embodiment are illustrated and describe with enough details, so that make those skilled in the art can make and use the present invention.Therefore, avoiding under the fuzzy suitable situation of the present invention, this instructions can omit some information that the technician in the association area has known.

Some embodiments of the present invention are specially adapted to computer implemented Law Text disposal system and are used for semi-automatically discerning the feature (such as quoting and quoting) in the juristic writing and discern this juristic writing and be stored in the method for the relation between other juristic writings of database.Described juristic writing can be legal precedent, decree, law review article, ALR article or law paper.In the situation of legal precedent, exemplary embodiment is described.Yet, will be appreciated that system and a method according to the invention has wider practicality, and can be used to different juristic writings, such as decree, legislation history and administrative procedures and patent.Some embodiment are applied to nonlegal document to the instruction here, such as scientific literature.

Before describing the preferred embodiments of the present invention, will be briefly described being used to describe term of the present invention.Any report judgement of supposing legal precedent is to this authority of law statement when this law is write.Subsequently, incident afterwards may influence the authority of the judgement of this legal precedent.These incidents afterwards can comprise the program afterwards of identical pendend lite or written judgement (for example directly historical), from by different way or use different reasonings to solve same problem and overthrown the judgement of the legal precedent afterwards of the different lawsuits of legal precedent in the early time or from differently solving same problem but clearly do not overthrow the judgement of legal precedent afterwards of the different lawsuits of described legal precedent.The direct history of legal precedent can comprise the record as the contact between the legal precedent of the part of identical lawsuit.Described direct history can have the different degrees of correlation, and can comprise positive historical (promptly keeping or support the authority of described legal precedent) or negative history (for example described legal precedent may no longer have its once all authority).The indirect history of legal precedent is legal precedent and be not the record of getting in touch between other legal precedents of a part of identical lawsuit.The indirect history of legal precedent can be the front or negative.The importance of specific legal precedent usually can be represented by (it can follow, overthrows or explain this legal precedent) the employed discussion quantity (being amount of text) in the process of the judgement that another legal precedent is discussed of legal precedent afterwards.As described below, this is known as the processing degree of depth to this legal precedent.One or more embodiment described herein can also realize in a system that such as the system of describing in the common pending trial U.S. Patent application of submitting on Dec 30th, 2,003 10/751,269, this patented claim is bonded to this with for referencial use.

Fig. 1 be wherein can specific implementation computer system 30 of the present invention block scheme.This system can semi-automatically discern the feature (such as quoting and quoting) in the legal precedent document, and generates subsequently about the information of this legal precedent in the context of other legal precedents.This computer system can comprise computing machine 32, server 34 and a plurality of client computer 36.Computing machine 32 can also comprise CPU (central processing unit) (CPU) 38, storer 40 and one or more process 42, and described process can be the software application that is stored in the storer 40.Described CPU controls the operation of described computing machine, and carries out the software application that is stored in the described storer.In operation, be fed in this computing machine corresponding to the multinomial electronic data of the text of announcing judgement of described legal precedent, and be temporarily stored in the storer 40.In the following discussion, the written comment of described legal precedent is known as legal precedent.Can utilize the described process that is comprised in the described software application that is included in the described storer that each electronic data (being each written comment of legal precedent) is handled automatically by described CPU, so that generate information about described legal precedent, as following described.For example, described CPU can carry out following operation: resolve the text of described legal precedent so that identification quotes the candidate of other legal precedents (being invalidated), and mark these quote for aftertreatment; Discern that candidate in the text of described legal precedent (being invalidated) quotes and the described text of mark correspondingly; Verify the source of quoting in the text of described legal precedent; Determine the processing degree of depth (promptly determining the importance of the legal precedent quoted) of the legal precedent quoted based on certain preassigned; Determine the negative processing of described legal precedent; And subject content text (such as annotations and comments) is distributed to respectively quoting in the described legal precedent according to the predtermined category system.Each described process can be carried out by the software application in the storer 40, and described software application is carried out by CPU 38.The details of each described process will be described below.

In case by process 42 finished handle after, just output is corresponding to the data recording 44 of specific legal precedent for computing machine 32, it comprises: about the information of the history of this legal precedent; Information about the processing degree of depth of respectively quoting of this legal precedent; The information of respectively quoting about this legal precedent; And about the information of every subject content text of quoting being assigned in this legal precedent (i.e. annotations and comments).The data recording that is generated for each legal precedent by computing machine 32 can be stored in the database 33 in the server 32.Subsequently, user's request of one of them of described a plurality of client computer 36 is about the information of a legal precedent, server 34 generates the user interface that comprises about the various information of the legal precedent of being asked based on being stored in each data recording in the database 33, and presents various information about this legal precedent to the user who consults this legal precedent.

An example of the user's who is provided for each client computer user interface is described below with reference to Fig. 2 A-2D.In this manner, the user of this client computer can ask the data about specific legal precedent, and system according to the present invention provides described data to this user.

When computing machine 32 receives electronic data corresponding to the text of every written comment of new law legal precedent, handle this legal precedent like that according to above-described, and the result of described processing is used as data recording 44 and is stored in the database 33 in the server 34.The user of each client computer can retrieve the data about specific legal precedent subsequently from server 34.Therefore, when one or more users of each client computer provided data about legal precedent, computing machine 32 can be handled each additional new law legal precedent and simultaneously in the database 33 that adds to corresponding to the information of this new law legal precedent in the server 34 at server 34.Below more detailed description is provided for the preferred, users interface of client computer and an example of information.

Fig. 2 A-2D shows according to of the present invention and is provided for client computer and by the snapshot of the example of the preferred, users interface of its demonstration and information.Fig. 2 A shows the computer screen 50 on the client computer, and this client computer is just showing the legal precedent that the user by this particular clients computing machine consults, and wherein this user interface has Windows form, tool bar, drop-down menu or the like.In this embodiment, this demonstration is the text that is known as the legal precedent of Pleasant v.Celli, and it is adjudicated by Gary Fu Niya the Court of Appeal.As mentioned above, have the clearly form of definition for any application of legal precedent, these are quoted thereby be convenient in the text of the written comment of this legal precedent identification.In order to visit the more information about shown legal precedent, the user of described client computer can be by clicking " KC " button 52 or clicking symbol 54 and select the service of quoting from services menu 51, and the described service of quoting is known as KeyCite ^TMKeyCite ^TMIt is assignee's of the present invention person who quote's (citator) trade mark.This symbol can be colour symbolism (a for example flag), and it provides the fast state of this legal precedent.As described below, red coloured flag can warn that certain part at least of this legal precedent of being consulted may not be good law, yellow coloured flag can show that this legal precedent has certain negative history, and perhaps another colour symbolism (such as the H of blueness) can show that this legal precedent has certain non-negative history.Yet, the invention is not restricted to the symbol or the color of any particular type.In case the user of described client computer has selected after the described person who quote system in some way, the screen shown in just can displayed map 2B.

Fig. 2 B shows the snapshot of an example of the computer screen 50 that the present invention can adopt, and it has control interface part 58 and display part 60.The control interface of this demonstration partly allows the shown information of customization.For example,, then can show the complete history of this legal precedent, comprising negative or positive direct history, negative indirect history and any relevant references if selected first radio button 62.If selected second radio button 64, then can only show negative direct and indirect history.If selected the 3rd radio button 66, the direct history that then can only show this legal precedent, thereby unrealistic any less important direct history (comprising list of references), direct history far away (such as the appeal after remanding a lawsuit for a new trial) and negative slightly indirect history.The control section 58 of described demonstration can also show the legal precedent number of the history that is regarded as this legal precedent.Control section 58 can also comprise the 4th radio button 68 and the number of references purpose of this shown legal precedent is indicated.When selecting the 4th radio button, show the tabulation of other documents.

In the example shown in Fig. 2 B, shown the complete history of described Pleasant v.Celli legal precedent.As shown in the figure, shown various types of history in display part 60, such as direct historical and negative indirect history, it is separated from each other by title.For each history, can indicate relation between listed legal precedent and the source legal precedent to the Short Description of described history or label, such as " suggestion quilt ... invalid ", " not by ... agree " or " with ... disagree ".In this embodiment, leading decision morning of identical court is invalid by described Pleasant legal precedent.(appendix demonstrate can with other example interface of describing by Fig. 2 B and/or use in combination at other local embodiment that describe of this instructions.) referring now to Fig. 2 C quoting described legal precedent described.

Fig. 2 C is an example that demonstrates the snapshot of computer screen 50, and it has control interface part 58 and display part 60.This screen display has gone out to quote each legal precedent of current this legal precedent of consulting (being the Pleasant v.Celli in this example).In this snapshot, selected the 4th radio button 68.Therefore, described control section can also have button 70, and it allows the user of described system to limit the shown type of quoting, as described below with reference to Fig. 2 D.Display part 60 can also show quotation marks symbol 72 and handle degree of depth symbol 74, described symbol be associated for quoting of described legal precedent, it refers to interested legal precedent.Quotation marks symbol 72 shows that the legal precedent of being quoted directly quotes from described interested legal precedent (promptly in this embodiment, Lubner v.City of Los Angeles comprises quoting from Pleasant v.Celli).To describe below and a kind ofly be used to discern the method for quoting and verifying described source of quoting according to of the present invention.Handling degree of depth symbol 74 for example can be one or many stars, and the number of its culminant star shows the degree of treatment of the written comment of this legal precedent, the amount of text at described legal precedent interested in the legal precedent suggestion of for example being quoted.The details of described processing depth assignment process will be described below in further detail.Describing one referring now to Fig. 2 D allows the user to be limited in the screen of quoting that shows in the described display part.

Fig. 2 D is an example that demonstrates the snapshot of computer screen 50, and it has control section 58 and display part 60.In this snapshot, suppose that the user of described system has selected at the restriction application button 70 shown in Fig. 2 C.As shown in the figure, the user of this system can limit shown application based on annotations and comments or theme, and this system will assess all quoting with reference to selected annotations and comments or theme, thereby only show the legal precedent that comprises selected annotations and comments or theme in the snapshot shown in Fig. 2 C.Annotations and comments can be the several sentence/paragraphs that are positioned at beginning place of legal precedent, and it shows the summary of law of the specific part of this legal precedent.The user interface of described system allows the researcher to carry out checking and matching function quickly and efficiently on legal precedent.To describe now according to of the present invention and be used to generate about the information of described legal precedent and the details of the system of described checking and matching function is provided.

Fig. 3 is the diagram of explanation the method according to this invention 100, and this method can be implemented on the computer system of Fig. 1 being used to handle legal precedent, thereby generates the information about this legal precedent, and described information can be used to checking and matching function.In order to help to understand described process, the action of single legal precedent will be described below.Yet, should be appreciated that and can handle a plurality of legal precedents simultaneously that this is because each legal precedent can be in the difference place in the described process.The electronic version 102 of the text of legal precedent (being called " WLLOAD " here) is fed to be quoted in the identifying 104 (ACITE), this is quoted the candidate that identifying discerns other legal precedents and other legal materials and quotes in the text of this legal precedent, and the described text of mark (promptly adding the signature symbol), thereby can identify described quoting in the back at an easy rate to described text.An example of label symbol can be the symbol combination " %v " that is placed on the beginning of quoting and place, end.It identifies described application to be used for the processing of back.

In brief, the described identifying of quoting is discerned the candidate and is quoted by identifying particular text pattern in the juristic writing, and these patterns and predetermined one group of list of references pattern are compared.Especially, can at first discern numeral in the text.Next, scan described text to find the abbreviation of contiguous described numeral, it is corresponding to known clerk's abbreviation, such as " Cal. " or " P. ".In case identify after one section text that the candidate quotes, control database 124 to determine that the candidate who is identified quotes effectively quoting in whether controlling database corresponding to this legal precedent with regard to inquiring about legal precedent with specific format and punctuate.If the candidate who is identified quotes with a certain quoting during this legal precedent is controlled database and matches well mutually, then carry out second around reason.If do not navigate to coupling, then can quote and carry out mark, to be used for carrying out afterwards hand inspection to this candidate who is identified.As mentioned above, each is quoted and has predetermined format.This form can be＜the legal precedent title〉＜reel number〉＜clerk's title abbreviation〉＜sequence number (if more than)〉＜page number in the volume 〉.For example, in " 18 Cal.App.4th 841 ", " Cal.App.4th " refers to " Gary Fu Niya appeal " clerk's the 4th sequence; " 18 " refer to the 18th volume; " 841 " refer to the 841st page, i.e. this page or leaf in the 18th volume that starts from Cal.App.4th is adjudicated in this legal precedent.

Example of quoting to legal precedent is " Pleasant v.Celli, 18 Cal.App.4th841,22 Cal.Rptr 2d 663 (1993) ", wherein first title part (the being Pleasant v.Celli) each side that expresses this legal precedent; Second secretary person's part (i.e. 18 Cal App.4th 841 and Cal.Rptr 2d 663) identifies the clerk who himself has aforesaid special characteristic form.

In case located after the text corresponding to clerk's title, just analyze the text adjacent, so that identify the described reel number of quoting, sequence number and page number and the time of the suggestion announced with this clerk's title.In case find after this information, just identification and the described candidate of mark quote as described above, quote so that it is identified as.The described identifying of quoting can be used process twice, wherein, at first identify such as the complete form of " Pleasant v.Celli; 18 Cal.App.4th 841; 22 Cal.Rptr2d 663 (1993) " and quote, itself and described legal precedent controlled database to be complementary, and places it in the table.In second time process, can go out the abbreviation of quoting based on the text identification that is included in the complete application in the described table such as Pleasant to described legal precedent.Should be noted that and at first do not identifying under each complete situation about quoting that these are quoted abbreviation and can't be automatically recognized.For with described table unmatched leave a question open quote abbreviation, can carry out tentative identification.

Quote identifying 104 output files 106 among Fig. 3, this document comprises the text that wherein marks any legal precedent of quoting.File 106 can be fed to subsequently quotes identifying 108 (IQUOTE), in this process, the text of this legal precedent is resolved, and respectively quoting identified and be labeled out in the text of this legal precedent, identifies described may originating of quoting in addition.At this moment, also do not verify described quoting of marking.Described quoting only is that the candidate quotes, and it must be further processed to verify.Below with reference to Fig. 4-6 described details of quoting identifying is described.This quotes identifying can output file 110, and this document comprises the text that wherein marks the legal precedent of quoting and quoting.At this moment, the text with this legal precedent of quoting and quoting mark is stored in the database for using afterwards, and can be fed in the plurality of processes.These processes can comprise quotes proof procedure 112, advanced treatment process 114 and negative processing procedure 116.As shown in the figure, these processes can be on same file executed in parallel, this is because information and the unpack that is generated by other processes and independent with it about described legal precedent that each process generated.To with reference to figure 7, Fig. 9 and Fig. 8 each described process be described in further detail respectively below.

In general, quote proof procedure 112 checking and quote really from described source (being the described legal precedent of quoting) by quoting the candidate that identifying 108 identifies, this is to quote quoting in the legal precedent and compare and realize by the candidate in the legal precedent of being quoted being quoted with described.This process generates data recording 118 subsequently, and described data recording comprises about the described information of quoting through checking.Advanced treatment process 114 uses the information (comprising described quoting through checking) that is generated by described system to generate advanced treating information, such as the occurrence number of quoting and the described characteristic of quoting based on its position (for example its be independently, the inside that is in the forefront of string or is in string).This process generates data recording 120 subsequently, and described data recording comprises every information about the processing depth information of quoting about being applied in the described interested legal precedent.Negative processing procedure 116 generates about by any information of quoting any negative processing of the described legal precedent interested that legal precedent receives, and in step 122, carry out artificial regeneration, so that reflect described negative processing to comprising about the database of information 124 of handled each legal precedent.As described below, can be by burster process 126 combining from described data recording 118,120 and annotations and comments distribute data record 128 (HNRESULT) that quote proof procedure and advanced treatment process respectively, so that generate the advanced treating information that comprises about handled legal precedent, the single data recording of quoting information and annotations and comments assignment information.This single data recording can be used to be created on the information that is displayed to the user on the computer screen subsequently, as shown in Fig. 2 A-2D.

The data recording 118 that comprises the information of verifying about the process in the described legal precedent of quoting can also be fed in the refers to identifying 130, this process attempts identifying the support text around the quoting and quoting in described legal precedent, so that generate refers to data recording 132.Described refers to data recording can be imported in subject content distribution and the threshold process process 134 subsequently, this process is complementary the word or expression in described the quoting and one or more annotations and comments or theme, and determine to select which annotations and comments based on a threshold value subsequently, as described below with reference to Figure 10.Subject content is distributed and threshold process 134 output datas record 128 (HNRESULT), and this data recording comprises selected subject content text (such as annotations and comments), and described text is fed in the burster 126, as described above.Therefore, according to the automatic information that generates about legal precedent of system of the present invention, and when being subjected to asking, utilize graphical user interface this information to be offered the user who uses described system subsequently.This user can be quickly and efficiently from the various information of single source location about described legal precedent, such as reference information, processing depth information, negative process information and subject content text (such as annotations and comments).Below with reference to the more details of Fig. 4 description about described system.

Fig. 4 shows the quoting identifying 108, quote proof procedure 112 of Fig. 3, advanced treatment process 114, negative processing procedure 116, refers to identifying 130 and subject content is distributed and the diagram of the more details of threshold process process 134.As shown in the figure, the output from each described process is fed in the system information database 33, as described above.

Quoting identifying 108 uses the described file that comprises the text that wherein marks the described legal precedent of quoting to discern as described above and mark is quoted.The text of this legal precedent comprises quoting of invalidated, and the file of quoting 144 that comprises through checking then is stored in the database 33.Described output of quoting identifying is many data recording, and wherein every data recording has quoting and described may originating of quoting of being identified.Can be combined with the text and the described described file of quoting that marks that comprise described legal precedent described output of quoting identifying, have the file of quoting and quoting 110 that marks so that produce, it is used as, and advanced treatment process 114, negative processing procedure 116, location recognition process 130 and subject content are distributed and the input of threshold process process 134.As following with reference to figure 5 and 6 in further detail as described in, during quoting identifying 108, carry out plurality of processes.At first, by scanning described text discerning for example symbol of quotation marks, thereby the candidate who identifies in the described text quotes, and described symbol shows the beginning or the end of quoting.Next, beginning of quoting and the end that identifier symbol (" %q ") comes mark to identify quoted in utilization.At last, discern described may originating of quoting, such as described legal precedent or other legal materials that is derived from wherein of quoting exploratoryly.As described below, during quoting proof procedure 112, verify described source of quoting subsequently.The output of quoting identifying 108 can comprise Qdata file 140 and Qtxt file 142, described Qdata file comprises about subsequently with reference to described every information of quoting may originating and be verified of quoting, and described Qtxt file comprises described actual text of quoting.

Described Qdata file 140 and Qtxt file 142 are fed to subsequently to be quoted in the proof procedure 112, and this process is used the electronic databank of available legal precedent to find and verified by described and quote may originating that each bar that identifying finds quotes.Quote for each bar, retrieve may originating that this quotes.Next, quoting quoting that identifying identifies and match well mutually, quote corresponding text with described in the described source so that locate with the described text that may originate by described.Be described source-verify described origin of quoting so just.Each bar for the source with process checking is quoted, and the data recording of quoting 144 through checking that comprises for a certain legal precedent is stored in the database 33.Subsequently, when comprising the legal precedent of quoting through checking and be shown as the quoting of a certain legal precedent, described quoting quoted symbol with comprising, and as mentioned above, the described symbol of quoting shows that this legal precedent has quoting through checking.Now described advanced treatment process will be described.

Advanced treatment process 114 can receive and comprise the file 110 that wherein marks the legal precedent text of quoting and quoting, and carry out plurality of processes at advanced treatment process described in the step 146, so that according to determining the described importance of quoting by certain mode one group of preassigned relevant with importance.These standards can be described the quote number of times that appears in the described legal precedent, the described type of quoting and quoting and described quote related through checking.At first, described advanced treatment process is read whole file 110, and identifies before by described and quote that identifying is tagged to be quoted.For quoting that each bar identified, the described type of quoting is defined as commonly quoting, goes here and there the forefront that the centre of quoting or string quote.Common quoting is typically to quote, and it appears in the legal precedent usually and do not have adjacent with it other and quotes.(inside) to quote be to appear at quoting of string centre of quoting or place, end in the middle of the string, in described string is quoted, in a sentence or paragraph a series of juristic writings quoted together.It is less that internal reference perceived as the degree of depth contribution that the legal precedent of being quoted is come into question usually.Quote the string forefront is to appear at quoting of beginning place that string uses, and it is perceived as the contribution of the described degree of depth more, and this is because all quoting of tool correlativity is placed on the place, forefront that string is quoted usually.Described advanced treatment process can also be quoted the available paging of thereon all of appearance and identifies described legal precedent page number for described, thus number of times that depth registration is written into and page break appear in the described legal precedent number of times as many.

Be output in the file 148 about every in the legal precedent information of quoting, this document is stored in the database 33, and wherein said information is such as being describedly to quote the total degree that appears in the described legal precedent document, every described type of quoting and corresponding to the page number of quoting appearance at every turn.Except with described arbitrary time occur being associated any who quotes through the quoting of checking, can also use above-mentioned information to generate above-described " to quoting of legal precedent " part and handle degree of depth symbol.Below the technology that is used to generate described processing degree of depth symbol will be described in more detail.

Negative processing procedure 116 can comprise automatic treatment step 150 and artificial verification step 152, and described step generates the tabulation of negative history corresponding to described legal precedent (promptly from disagree with current legal precedent or make other written comments of other invalid legal precedents of current legal precedent).During automatic treatment step 150, scan comprising the file that wherein marks the Law Text of quoting and quoting, so that identify the stem of some word, such as " overrule (invalid) ", " recede (recalling) ", " disapprove (disagreeing with) " or " distinguish (difference) ", it may show negative processing.The process of the root of the word " overrule (invalid) " in the text that is used for discerning legal precedent has been described as an illustration.When identifying example of root " overrule (invalid) ", as described below such, use one group of heuristic rules and determine to comprise in fact whether the sentence of this root that is identified be exactly invalid, as described below with reference to Fig. 8.Subsequently, during manual verification process 152, the operator of described system verifies the result of described automated procedure, and reality is added to legal precedent through checking invalid control database in 124.Operating personnel can also identify other the negative history about the legal precedent that can not discern automatically at an easy rate, as following described.It is invalid that described negative processing procedure helps operating personnel to identify apace.These are invalid to be the authoritative negative history of the reasoning of the described legal precedent of influence.

Location recognition process 130 is used the described file of quoting 144 that comprises the file that wherein marks the legal precedent text of quoting and quoting and comprise the process checking, it identifies any tagged one group of heuristic rules of quoting and use, so that quote near identification and select a part of text from each bar as described below, it may show by the described text of being supported of quoting.If quote in legal precedent and occur repeatedly for a certain, then make up the text of quoting corresponding to described on every side that occurs each time.In addition, if the described proof procedure of quoting has been verified any the quoting that is associated with this application as described above, then also combined other texts around the text of quoting of this process checking and described the quoting.Can use subsequently and be in each bar all texts that identify around quoting and determine to be applicable to described one or more annotating and commenting on or the subject content title of quoting.Described subject content title is classified to described quoting based on the subject content field of predetermined number, such as intellecture property or patent.Process 154 (Headqf) is read with given and is quoted all adjacent texts that identify, and generate the Natural Language Search inquiry in case in existing database the coupling of search and the text that is identified, as following described.In U.S. Patent No. 5,265, described natural language querying process has been described generally in 065 and 5,418,948, described patent is transferred to the assignee identical with the application, and is bonded to this with for referencial use.Headqf process 154 generates the file 156 that comprises described natural language querying.Utilize described natural language querying, subject content allocation procedure step 158 is carried out described natural language querying at the annotation data storehouse, so that identify the subject content title (such as annotations and comments) that may be complementary with the text around described the quoting.For the subject content title of each coupling, described inquiry also generates the confidence value of the matching degree that shows described subject content title and described text.The subject content title that is closely related most and its confidence of predetermined number are provided to threshold process process steps 160.

Described threshold process step is used the described subject content title that identifies and is carried out various calculating, and described calculating is considered the confidence of the ordering of described subject content title, described subject content title and with reference to the number of quoting of this subject content title.After having carried out described calculating, the top of predetermined number annotations and comments are hit and whether passed through the sign of described threshold process with being stored in the database 33 to described link of quoting corresponding to these annotations and comments that show of each annotations and comments.These subject content titles allow to utilize these subject content titles quoting in the legal precedent classified and to search for, as top with reference to figure 2D described.Now the described identifying of quoting will be described in further detail.

Fig. 5 and Fig. 6 show about diagram of quoting the more details of identifying 108 according to the present invention.Quote identifying 108 and can comprise lexical scanner process 170, paragraph impact damper 172 and major cycle process 174, it is used for receiving the text of legal precedent and generates automatically and comprises that every of being identified quotes and corresponding to the every file of quoting that may originate.Lexical scanner 170 resolves into logic fragment (being known as pictograph) to document, and uses these pictographs to discern by major cycle process 174 subsequently and quote.The pictograph that is identified by described lexical scanner can comprise capitalized words, may finish the punctuation mark of a sentence, the white space such as one or more spaces, legal precedent title, footnote be with reference to, the beginning of quotation marks and the end of quotation marks.Employed described lexical scanner process can be based in the middle of the multiple software application that can buy any, and such as the application that is known as FLEX, it can be from Sun Microsystems Inc, MountainView, and CA obtains.Described lexical scanner accepts to specify the pattern of grammer, and identifies an action when navigating to AD HOC.Especially, can be divided into legal precedent the paragraph of particular type based on predetermined one group of standard (such as a rule set) according to lexical scanner of the present invention: 1) may comprise the paragraph of quoting; 2) paragraph of quoting as predetermined piece; 3) comprise the paragraph of important information (such as the beginning of document, the sequence number of document or the end of document) about document; And 4) for the described nonsensical paragraph of identifying of quoting, such as annotations and comments, title or the like.Can use multiple various criterion and rule to discern these paragraphs.

An example of the rule set that can be used by the present invention will be described now.This rule set can comprise following rule: identify and do not comprise any paragraph of quoting and it is stored in rule in the described paragraph impact damper, described paragraph is override by next paragraph at this place; And corresponding to the rule of the paragraph of quoting with possibility, wherein said lexical scanner returns label to described major cycle, and this label shows that described paragraph is that plain text paragraph, predetermined block are quoted paragraph or shown that the described text of quoting appears in the footnote.In case determined after the type of described paragraph, described lexical scanner is just handled text in this paragraph according to identical mode, so that identify any pictograph in each paragraph.

In each paragraph, described lexical scanner can be discerned following pictograph: capitalized words, non-capitalized words, the numeric character string, abbreviation, correct title (i.e. " Mr.Smith "), legal precedent is quoted, chapters and sections are with reference to (i.e. " Section 150 "), legal precedent title (being Roe v.Wade), the reference of embedding, any end of the sentence punctuate, the colon in space followed in any other punctuation character, back, branch or comma, single or multiple white space characters, the beginning of quoting, the end of quoting, footnote number, left parenthesis and right parenthesis, left square bracket and right-hand bracket, a left side brace and right brace, corresponding to the mark of quoting, and corresponding to the mark of the reference that embeds.More details about the operation of described FLEX software application and modification can be from Sun Microsystems Inc.Reference Manual, Programmer ' s Overview Utilities and Libraries, Chapter9, pp, 203-226 obtains, and it is bonded to this with for referencial use.

Paragraph impact damper 172 is that storage is quoted if identify in this paragraph about by the nearest place of the pictograph of the paragraph of scanning of described lexical scanner before being handled by major cycle 174, and then described pictograph is written out in the output file subsequently.Major cycle 174 can determine for each pictograph that returns by described lexical scanner take what action, the described paragraph impact damper of management and decision when from described paragraph impact damper, abandon data corresponding to last paragraph, for be present on several physics paragraphs quote several physics paragraphs be linked in the virtual paragraph together, determine interruption between the sentence in paragraph where appearance and determine when handle virtual paragraph by one group of heuristic rules, such as will be described below.

Fig. 6 is a process flow diagram of quoting identifying 108 according to of the present invention.In step 180, land the described legal precedent text of scanning piecemeal, and identify sentence and pictograph in this paragraph for each paragraph.In step 182, each pictograph in the paragraph is used one group of heuristic rules, quote so that determine whether to identify.Described lexical scanner and described one of them of paramount importance function of quoting identifying are exactly beginning and the end that identification is quoted.This is very difficult, because each writer may use different slightly forms for the beginning of quoting with the end.Therefore, need several rules to discern the beginning and the end of quoting.An example that can be used to the heuristic rules collection of realizing this identification will be described below.These rules can be used described lexical scanner to discern traditional beginning and quote punctuation mark (such as " or "), discern traditional end delimiter (such as " or ") of quoting, and perhaps quote beginning/end symbol in the longer character string of identification.For example, a certain rule can be attempted discerning wherein traditional end symbol of quoting and be embedded in string in the sentence.For each bar in the middle of these rules, can check described pictograph character on every side, be actually the beginning or the end of quoting so that guarantee this pictograph.

In case used after the described rule for each pictograph in the paragraph, the described identifying of quoting just determines whether to have another paragraph, and be recycled to step 180 to handle new paragraph in step 184.In case analyzed after all paragraphs, in step 186, the output of described process comprises quoting of being identified and the data recording that may originate that these are quoted.The described proof procedure of quoting will be described now.

Fig. 7 shows the process flow diagram that is used to verify the method 112 of quoting according to of the present invention.In step 200, the described proof procedure of quoting is read and to be quoted identifying 108 and be identified as the text string of quoting, and identifies separator (when existing) in the middle of the predetermined group separater from described text string.Described separator can comprise suspension points, parenthesized expression formula and stop phrase.The described phrase that stops to comprise various law phrases and other phrases that is helpless to discern the source of quoting, such as " citation (s) omitted (omission is quoted) ", " sic (sic) ", " emphasis provided " or the like.When having described separator, described separator is used to described text string is resolved to fragment, and wherein each fragment comprises the content that appears between a pair of separator.In step 202, described text string is resolved, so that determine its length, this is that wherein stopping word is the word that does not contain content, such as article and preposition because it for example can be 6 non-words that stop that length is quoted in I checking.Also described text string is resolved so that compress any word that comprises ellipsis or other punctuation marks, for example " T] hen ".Drop in two different classes of one of them through the text string of quoting of resolving: (1) has the text string of single fragment; Perhaps (2) have the text string of a plurality of fragments.Therefore, in step 204, described system determines whether described text string has single fragment.If text string has single fragment, then in step 205, determine for the contrary document frequency (IDF) of the set normalization of each (word) in this single fragment of text string.The document frequency values is represented the frequency of particular item in typical literature collection, and IDF then equals the inverse (i.e. 1/ document frequency) of described document frequency, perhaps in other words, equals a certain the rare degree in file set.In a preferred embodiment, if the occurrence number of a word greater than 0, then can the described set normalization of following calculating contrary document frequency (IDF):

$\mathrm{idf}\_\mathrm{score}=\frac{\log \left(\frac{\mathrm{Collection}\_\mathrm{Docs}}{\mathrm{Doc}\_\mathrm{Occurences}}\right)}{\log (\mathrm{Collection}\_\mathrm{Docs})}$

Wherein, Doc_Occurrences is that described given item is present in document number wherein, and Collection_Docs is the total number of files in the described set.Good the purpose that described IDF is used to be identified for mating, this usually may be unique because of a rare word, and to provide about quoting be the good indication of originating from the candidate.

In case calculated after the described IDF for each, just can be by every sort (step 206) with selected threshold value the highest following IDF value of IDF value to selected number (for example 6), and it is inserted in " template " (being storage array) (step 207), described template shows each position in described text string.Do not use any with high abnormally IDF value (for example greater than 0.80), this is because this item that seldom occurs usually is the word of misspelling.If there are several to have identical IDF value, then described every alphanumeric order can be used as at described template described every less important key that sorts.If still there is the item (item that for example has the spelling of identical IDF value and alphanumeric) that is equal to, then described every position in described text string can be used as in the described template described every the 3rd key that sorts.Can compare described quote of described template subsequently, so that the position based on described each high IDF item determines whether to occur accurate coupling in step 208 with the document of originating from described candidate.If accurately mate, then output being quoted and it is fed in the described database, as described above through what verifies in step 210.If accurately coupling do not occur, then in step 212, generate checking it fails to match message, and described quoting is not stored in the described database.

In step 204, if having a plurality of fragments, described text string (promptly in text string, comprises one or more separator items, such as " The roof fell in...crashing down "), then described process proceeds to step 214, in this step, determine corresponding to each the IDF in each required fragment.Subsequently, by the item sort (step 215) of IDF to the selected number in each fragment (for example 4) with the highest IDF value below the described threshold value, and it is inserted (step 216) in the template, so that determine the purpose (step 208) of described every position in described fragment to be used to mate.For the text string that has more than 4 fragments, can use preceding two and latter two fragment to come to mate (step 217-218) at the described candidate document of originating.In this manner, quote quoting of identifying identification automatically and verified automatically, and go out any through the quoting of checking, as described above by quoting symbol logo by described.To describe now according to negative processing procedure of the present invention.

Fig. 8 shows the diagram that is used for the negative processing of definite legal precedent according to of the present invention.The file 110 that comprises the text that wherein marks the legal precedent of quoting and quoting is imported into automatic negative processing procedure 150.This automatic negative processing procedure can be carried out following operation: the 1) appearance of identification stem " overrule (invalid) " in legal precedent; 2) determine described stem and the adjacency of quoting; And 3) get rid of any bad legal precedent.At identification stem " overrule (invalid) " before, can check that legal precedent controls database 124, and if had any history corresponding to this legal precedent, then described automatic processing stops.In order to discern the appearance of stem " overrule ", the text of this legal precedent is scanned, and determine the verb time sequence of any appearance of this stem.The verb time sequence of described stem shows and describedly invalidly is meant the invalid previous legal precedent of current legal precedent, or certain other types is invalid.A heuristic rules collection can be sought specific verb time sequence, and takes action based on described verb time sequence subsequently.

To describe an example of employed rule set now, but the invention is not restricted to any ad hoc rules collection.For example, rule can find " overrule " or " overrules " in the sentence, and scanning reaches 4 words backward.If find " not " or " never ", then abandon this sentence, invalid because it does not refer to any reality.If find " we ", then this sentence to be added in the invalid tabulation of possibility, this tabulation is checked by the mankind.If do not find any one described phrase in described scan period backward, then this sentence added in the described tabulation yet.

Another rule can find " overrule ", and scanning reaches 5 words backward subsequently, so that attempt to find non-legal precedent word, described non-legal precedent word will show that the other guide that is different from described legal precedent is disabled, thereby this sentence not added in the described tabulation.Several examples of these non-legal precedent words comprise " request ", " motion ", " objection ", " claim " and " verdict ".If described rule finds " point " or " points ", then can scan the end of this sentence forward up to this sentence, if and found " case ", " cases " or " supra ", then the state of this sentence would be unknown, and it is passed to human overlooker.

Another rule can find " overrule " and backward or the line scanning of marching forward, and it is possible invalid to refuse or accept based on other words of this word of next-door neighbour " overrule ", and this is because these additional words will be provided at the context that wherein uses this word " overrule ".For example, in case find " overrule ", then can scan this word 4 words before, and when finding following word, take following action: 1) if find " we ", and " we " word before is " that ", then ignore " we " (only discussing) about invalid, if but do not find word " that ", then this sentence is possible invalid: 2) modify if this verb is expressed probabilistic word, such as " rather ", " might " or the like, then refuse this sentence, this is because court may only show that it may deactivate this legal precedent; 3) if any word shows discusses to invalid, then this sentence is rejected; 4) described invalid if a certain word shows that another person has implemented, then this sentence is rejected; And 5) if find " will " or " should ", then described process sees that backward 5 words are to seek the front word, so that accept this sentence.For the infinitive of verb " overrules ", this verb and the passive voice of this verb similar rule set can also be arranged.

Another rule set can seek show discussion whether will be invalid, whether court have invalid authority or invalid word in the past, these sentences are not owing to comprise invalid being rejected of reality.Another rule set can refuse to show that other people are implementing the sentence of invalid (another court of promptly passing by).Another rule set can be sought " overruling " and refuse or accept this sentence based on the sentence around the described word subsequently, just as described above.

Can also have and be independent of other rules that the special characteristic of sentence sought in verb " overrule ".For example, if phrase " COURT: " is positioned at the beginning of sentence, then it shows that this sentence can be accepted from directly the quoting of judge.If word " Congress " is positioned at the beginning of sentence, then it can show that act of Congress is disabled or shows Congress itself just in an invalid legal precedent, and this sentence can be rejected.If in sentence, near word " overrule ", find word " circulated ", then this sentence can be accepted so that catch uncommon language, such as " becausethe decision overrules an opinion of this court; it was circulated to all activejudges... ", it can't be automatically recognized in certain other modes.Another rule can be sought " overrule " and refuse this sentence in the string of being quoted, this be because its normally by another court to the legal precedent of quoting by current court make invalid.

Except stem " overrule ", can also search for and discern other synonyms.For example, described rule can also detect stem " abrogat " for Gary Fu Niya legal precedent, term " abrogated " is used in described Gary Fu Niya legal precedent, and described rule can also detect phrase " receded from " for the Florida legal precedent, this be because these terms in each corresponding state, be used to represent invalid.Can use these verb time sequence rules according to any order, and the invention is not restricted to any specific execution sequence of any ad hoc rules collection or described rule.

From the output of the described verb time sequence rule set of described automatic negative processing procedure be possible invalid tabulation.Subsequently, for every may inactive application the adjacency rule, invalidly whether be applicable to specific legal precedent to determine this.For example, described adjacency rule can be removed possible invalid under following situation: the sentence that comprises described stem does not comprise to be quoted, previous or next sentence does not comprise to be quoted, the sentence that perhaps has stem " overrul " does not comprise the word that is used to refer to for legal precedent, such as " case ", " opinion ", " holding ", " precedent ", its plural form or " progeny " or " v. ", " ex rel ", " ex parte " or " supra ".Any sentence that comprises stem " overrul " and satisfy described adjacency rule all is added in the invalid tabulation 222 of hint of described legal precedent.In the artificial review process step 152 of being undertaken, check subsequently and to check that these hint invalid by the mankind.During described artificial review process, people determine that also the legal precedent that is disabled and specified data are imported into legal precedent and control database in 124, and this legal precedent is controlled database and followed the tracks of legal precedent in the described legal precedent database.

According to this negative processing procedure on the other hand, described automated procedure can also be discerned the relation except being disabled, such as " disting " in the legal precedent (expression " distinguished ") or " apposite ", this is to realize by described method being expanded to the language that characterizes described other relations.In brief, described negative processing procedure helps the personnel of checking to determine possible invalid in the legal precedent, thereby this is to significantly reduce and must be realized by the actual amount of text of checking of people by the invalid possible position in automatically definite legal precedent.Therefore, described negative processing procedure has improved and can discern invalid in the legal precedent and add it in the negative history of this legal precedent speed.Now described advanced treatment process will be described in more detail.

Fig. 9 is the process flow diagram according to advanced treatment process 114 of the present invention, wherein gives the advanced treating allocation of symbols every in the legal precedent to quote, thereby uses the people of described system can determine to discuss specially the specific number of texts of quoting apace.This information can be used as the expression for the described correlativity of quoting, and this is to be used for quoting of height correlation because court will and discuss more text.

In step 230, as mentioned above, text and the described file of quoting and quoting that marks with described legal precedent are received by described advanced treatment process.In step 223, this advanced treatment process is discerned quoting in this legal precedent, and determines reference type subsequently in step 234.Quote for every in this legal precedent and can be: 1) be in quoting of the string forefront of quoting; 2) there are not other to follow quoting of quoting; 3) be in quoting of the string inside of quoting; Perhaps 4) pro forma history is quoted (promptly only being cited as quoting for the auxiliary historical reference of one of them legal precedent of quoting) in himself right in the context of the document.The importance of each in the middle of these reference types is all different.It is for example, independent that to quote or be in quoting of forefront that string quotes often more important than quoting of the centre that is in this string.

Next described advanced treatment process determines in the described legal precedent whether any additional quoting is arranged in step 236, and is circulated back to step 232 and quotes with next bar of handling in this legal precedent.In case identified all references in this legal precedent and be classified into above-described wherein in one type after, just they are fed to as shown in Figure 3 burster process 126.After described burster process, in step 238, described advanced treatment process is for the different sum of quoting every kind of reference type in definite described legal precedent of each bar.For example, for quoting of Pleasant v.Celli, have altogether 5 and quote in described legal precedent, wherein three are in the forefront that string is quoted, and two are in the inside that string is quoted.In step 240, use about every in the described legal precedent this information of quoting and about determining to be assigned to this specific degree of depth symbol of quoting with the specific any data of quoting that are associated of quoting through checking.In case quote and distributed after the described degree of depth symbol for every, described advanced treatment process is finished.

Use description to distribute an example of the technology of degree of depth symbol now, but the invention is not restricted to be used to distribute any particular technology of described degree of depth symbol into application-specific.In addition, the invention is not restricted to the degree of depth symbol of any particular type.In this embodiment, in described legal precedent, quoting in this legal precedent of 1 to 3 time occur and be assigned with two stars (for example * *) with any reference type (be independent reference, be in the string forefront of quoting quote or be in quoting of centre that string uses), in this legal precedent, 4 to 8 times quote occur and be assigned with three stars (for example * * *), and occurred 9 times or quoted more frequently being assigned with four stars (for example * * * *) with any reference type with any reference type.For further these distribution of refinement, with any reference type occurred 3 times and have associated therewith through checking quote quote and be assigned with three stars (for example * * *), if only have inner string reference type, then quote and subtract a star from this and quote.Therefore, by the automatic degree of depth symbol that distributes corresponding to the application-specific in the described legal precedent of system according to the present invention.The user that described degree of depth symbol helps described system determines more apace which is quoted and may have more correlativity.To describe now according to subject content text assigning process of the present invention.

Figure 10 shows according to of the present invention and is used for the process flow diagram of distributing to the described method of quoting 250 of described legal precedent from one section text of the legal precedent of being quoted.In the example that is described below, the text of the annotations and comments in the legal precedent of being quoted is distributed to described quoting, but can use together with the different text chunks in the legal precedent of being quoted according to text assigning process of the present invention.In step 252, distribute corresponding to every refers to of quoting (promptly probably corresponding to quoting the text filed of the text supported) by described according to the rule set that is discussed below.

In order to discern described refers to, can use several text resolution rules, some rules wherein are better than other rules, but it will probably identify described text as a whole.In order to allow the different validity of described Different Rule, can according to the text that is extracted be the likelihood of the part of correct text position come the text that is extracted be divided into " height ", " in " and " low " three groups.Described rule comprises:

Classification	Text
		High
	1, is included in any non-quoted material that comprises in the sentence that base quotes.2, all first non-quoted sentences before described base is quoted in identical paragraph.If 3 do not have the 2nd class sentence, then be all the first non-quoted sentences after described quoting in the identical paragraph.4, can be identified as all texts of quoting from the legal precedent of being quoted.
		In	5, appear at the next one that described base quotes at all sentences of preceding quoting between (not and described base quote in abutting connection with) and the 2nd class sentence.6, appear at the next one that described base quotes after quote all sentences between (not and described base quote in abutting connection with) and the 3rd class sentence.If 7 do not have the 2nd class or the 3rd class sentence, and comprising the paragraph that described base quotes and finish with colon or comma, then is to comprise the integral body of the next one of this paragraph that this base quotes at the back paragraph.If 8 do not have the 2nd class or the 3rd class sentence and do not have the 6th class paragraph, then be to comprise the integral body of the next one of this paragraph that described base quotes in previous paragraphs.If the 9 any described text filed annotations and comments reference markers of being discerned by any rule that comprise then comprise described annotations and comments and crucial row thereof.
Low	10, the integral body that comprises this paragraph that described base quotes.
		Generally	If 11 described quoting appear in the footnote, then it is treated as the position that appears at described footnote position and described footer mark.
Do not charge to	12, quote and only appear in another follow-up history of quoting.

Subsequently, in step 254, according to the rule that is used to discern described refers to, utilize the height correspond respectively to 2.0,1.0 and 0.5 weight, in or low coupling to every being weighted at described refers to place.May need different rule sets to determine described weight for dissimilar document (such as legal precedent or law review article).In case identified each section text and distributed after the confidence value, in step 255, each section text that is identified and each section text that may be in the document of being quoted have been complementary.In an example, each section text that is in the document quoted may be annotations and comments, but the invention is not restricted to any particular type of text that each section text that is identified is complementary with it.Described coupling can realize that described patent is had by the application's assignee by using the natural language querying as describing in the U.S. Patent No. 5,265,065 and 5,418,948 of reference formerly, and is incorporated in this with for referencial use.The result of described search is the tabulation of the possible text chunk (such as annotations and comments) from the legal precedent of being quoted, and it can be assigned to described the quoting in the described legal precedent, and can distribute the confidence corresponding to each possibility text chunk.

Next, in step 258, the passing threshold processing procedure selects to be assigned to described described one or more snippets text of quoting.Described threshold process process is quoted for each bar based on described confidence each section text is sorted.Whenever following amount equals or exceeds at 0.5 o'clock, each section text can be logged in the described database:

$\frac{1}{-({\mathrm{\&beta;}}_0+{\mathrm{\&beta;}}_1\sqrt{\mathrm{belief}-0.4}+{\mathrm{\&beta;}}_2\mathrm{In}\left(\mathrm{freq}\right)+{\mathrm{\&beta;}}_3\mathrm{In}(\mathrm{belief}-\mathrm{<figure-callout\; id="2"\; label="lag"\; filenames="A20068003362701461.png"\; state="\{\{state\}\}">lag</figure-callout>}2+0.0001)_{1+\mathrm{\&sigma;}}}$

β value corresponding to this equation is as follows:

Document	Ordering	β 0	β 1	β 2	β 3
						Non-ALR	1	4.0451	3.1975	0.8477	.9033
Non-ALR	2	0.5573	9.0220	1.0348	0.6743
						Non-ALR	3	-2.0421	11.2619	0.8949	0.2954
ALR	1	-1.4256	50.4929	0.3488	0.0000
						ALR	2	-2.8199	65.6148	0.6207	0.0000
ALR	3	-2.3701	40.8479	1.2445	0.0000
						ALR	4	-3.3474	60.8075	1.7349	0.0000
ALR	5	-3.0805	55.6003	1.3188	0.0000

Wherein, each row that is marked as " ALR " comprise variable corresponding to the ALR article, and as mentioned above, it has than the higher confidence of non-ALR document.Each row that is labeled as " non-ALR " comprise variable corresponding to non-ALR document.

In described equation, Freq quotes right total reference frequency corresponding to described, and lag2 is second confidence the back candidate (if perhaps do not have this candidate then be 0.4) when described candidate being arranged by confidence according to descending.In case finished described threshold process and described one or more snippets text distributed to after in the described legal precedent every quotes, in step 260 as described above one or more snippets text storage in described database, thereby can be that the user obtains described text when being subjected to asking.

In brief, described subject content assigning process comes to generate one or more snippets text for quoting in the described legal precedent based on each section text in the legal precedent of being quoted (such as annotations and comments).Described process at first automatically identifies the support text in the described legal precedent and confidence value is distributed to described support text, its all described each section texts are complementary with each section text in the legal precedent of being quoted, and subsequently distributing to described specific quoting automatically from one section text in the legal precedent of being quoted (such as annotations and comments).These subject contents are distributed and are allowed to classify or selects the quoting of described legal precedent by described subject content, and this understands helpful in described alignment processes.

Therefore, the system of Realization by Machine according to the present invention handles the document such as legal precedent automatically, and generates the information about described document, and described information can provide the useful information about the content of the document for the user of this system.On the other hand, in traditional system, this information about described document will be generated by reading described document and taking notes about described document by the mankind, this be very slowly, expensive, error-prone process.For legal precedent, the information of the advanced treating of quoting that described system can generate information about the negative history of this legal precedent automatically, do about this legal precedent, be derived from the information of quoting of particular source and about being assigned to the information of specific one or more annotations and comments of quoting in this legal precedent about being verified as in this legal precedent.Therefore, the operator of described system can generate this information about described legal precedent apace, and the user of described system can find this information apace, and this is because can visit this information at an easy rate from graphical user interface.

Appendix

Lantern slide 1

Lantern slide 2

Lantern slide 3

Lantern slide 4

Lantern slide 5

Lantern slide 6

Lantern slide 7

Lantern slide 8

Lantern slide 9

Lantern slide 10

Lantern slide 11

Lantern slide 12

Lantern slide 13

Lantern slide 14

Lantern slide 15

Lantern slide 16

Lantern slide 17

Lantern slide 18

Lantern slide 19

Lantern slide 20

Lantern slide 21

Lantern slide 22

Lantern slide 23

Lantern slide 24

Lantern slide 25

Lantern slide 26

Lantern slide 27

Lantern slide 28

Lantern slide 29

Lantern slide 30

Lantern slide 31

Lantern slide 32

Lantern slide 33

Lantern slide 34

Lantern slide 35

Lantern slide 36

Lantern slide 37

Lantern slide 38

Lantern slide 39

Lantern slide 40

Lantern slide 41

Lantern slide 42

Lantern slide 43

Lantern slide 44

Lantern slide 45

Lantern slide 46

Lantern slide 47

Lantern slide 48

Lantern slide 49

Lantern slide 50

Lantern slide 51

Lantern slide 52

Lantern slide 53

Lantern slide 54

Lantern slide 55

Lantern slide 56

Lantern slide 57

Lantern slide 58

Lantern slide 59

Lantern slide 60

Lantern slide 61

Lantern slide 62

Lantern slide 63

Lantern slide 64

Lantern slide 65

Lantern slide 66

Lantern slide 67

Lantern slide 68

Lantern slide 69

Lantern slide 70

Lantern slide 71

Lantern slide 72

Lantern slide 73

Lantern slide 74

Lantern slide 75

Lantern slide 76

Lantern slide 77

Lantern slide 78

Lantern slide 79

Lantern slide 80

Lantern slide 81

Lantern slide 82

Lantern slide 83

Lantern slide 84

Lantern slide 85

Lantern slide 86

Lantern slide 87

Lantern slide 88

Lantern slide 89

Lantern slide 90

Lantern slide 91

Lantern slide 92

Lantern slide 93

Lantern slide 94

Lantern slide 95

Lantern slide 96

Lantern slide 97

Lantern slide 98

Lantern slide 99

Lantern slide 100

Lantern slide 101

Lantern slide 102

Lantern slide 103

Lantern slide 104

The figure KeyCite that is used for legal precedent

Legal precedent frame layout ... ... ... ... ... ... ... ... ... 111

Procedural motion frame layout ... ... ... ... ... ... ... ... ..111

Situation ... ... ... ... ... ... ... ... ... ... 112

Rule-summary ... ... ... ... ... ... ... ... ... ..145

Legal precedent frame layout

Icon SlugQuote court's line date file wire or UPO file wire or UPO ... The historical processing QuoteThe historical processing Quote ...

Procedural motion frame layout

[icon] The historical processingCourt's line date [the legal precedent frame of KeyCite]

Situation

Rule-summary

The ■ parental generation is in its filial generation centered beneath.

Non-the remanding a lawsuit for a new trial of ■ concerns that arrow should enter from frame bottom.

■ is when when existing T shape to intersect between parental generation and the filial generation, and the bottom vertical line will be offset.

The ■ relation is separated and will be taken place on court's rank of parental generation.

The legal precedent that ■ sends downwards will begin the top of the legal precedent that sent to this legal precedent from the right side of this legal precedent downwards.

■ as more than a legal precedent by when single parental generation sends downwards, will continue to hold additional legal precedent from the line on the right side of this parental generation.

The parental generation that ■ has the top that arrives this legal precedent will have the line that sends and be connected to its filial generation from its right side.

■ is on the identical court rank and its relation when being in horizontal lawsuit tabulation and going up when parental generation and filial generation, and parental generation will be plotted in the left side of filial generation.

If the ■ parental generation can't be in the filial generation centered beneath, then it will be offset below filial generation.

■ is in order to draw its all parental generation relations, and the width of filial generation will be lengthened to necessary point.

■ walks around a filial generation when arriving another filial generation when parental generation with needs, and this parental generation will be offset.

■ motion corresponding to identical parental generation on identical court rank is laminated to each other.

The stacked procedural motion of ■ is ranked into the bottom that is positioned at the earliest, the latest be positioned at the top.

When ■ had the phase same date when each procedural motion, any order that receives them according to us was carried out stacked to it.

■ is when from left to right sorting to each filial generation, and the whole procedure motion is stacked gets its date of member the earliest.

The order of each filial generation of ■ will be filial generation the earliest on a left side, filial generation is on the right side the latest.

When ■ need intersect when line, online crossing point was on the horizontal line projection will be arranged.

Handle if the ■ process block has different history to be shown, it should be thus lifted to the essence frame.

The legal precedent that ■ has the filial generation of remanding a lawsuit for a new trial and laterally lawsuit filial generation should have two lines that send from its frame right side.

■ only carries out stacked to the procedural motion with public single parental generation.

Emphasis

Exceed and quote inspection: Graphical KeyCite delineates out the picture of program legal precedent history

A picture surpasses thousands and thousands of words.For legal research person, the value of powerful new Graphical KeyCite may be also higher, and this is because it illustrates the program history of case law definitely.

Thomson West is Thomson Corporation (NYSE:TOC; TSX:TOC) company under, its Graphical KeyCite that introduces now is Key Recent innovation, and described Key Service has caused the revolution of quoting the inspection field when being introduced into for the first time.

Then, the law librarian of american association of law libraries has praised KeyCite sign intuitively, it makes legal research person know the law whether a certain judicial opinion remains at once, and shows in other suggestions described processing degree of depth star and symbol to the degree of dependence of this legal precedent.First of history that KeyCite still allows the researcher to probe into legal precedent like a dream quotes the service of inspection.

Graphical KeyCite has taken these innovations to a new height, and this is to realize by the picture of the direct history of drawing out legal precedent definitely.Along with described legal precedent rises to higher court, described feature is Reference-links motion, defense and rudimentary court decision finally.If this exclusive KeyCite feature helps the researcher to see legal precedent at once along with the time moves in the system of court, and understands the influence on each rank apace.

The product development of KeyCite is responsible for Jon Medin and is said: " the historical of court case is illustrated for the first time, thereby can help the researcher to understand its influence quickly ".

Medin adds, KeyCite says something the analysis from the law editor of Thomson West with technology is combined, such as quote legal precedent that the legal precedent discussion quotes how much.Medin mentions: " created West company the identical lawyer-editor of a large amount of collections of authoritative precedent method annotations and comments has also been distributed KeyCite sign and symbol, lawyer and judge depend on these and indicate and symbol is quoted the law that whether remains to learn ".

Compared with any other legal research service, the more multichain that the document on the Westlaw is included in relevant sources connects.Medin mentions, and when court's document was mobile in justice system, GraphicalKeyCite utilized these links and uses proprietary technology to illustrate contact between the described court document.In addition, the researcher can click chart simply and open full copy document on the Westlaw.

The senior executive Mike Bernstein of Westlaw market department of Thomson West says: " in our test, the researcher of use Graphical KeyCite can understand the direct history of legal precedent faster and more exactly.For quoting anyone of research, the GraphicalKeyCite picture all definitely surpasses thousands and thousands of words ".

Though the description of front is with reference to specific embodiment of the present invention, but those skilled in the art will be appreciated that, can make a change described embodiment under the situation that does not deviate from principle of the present invention and spirit, scope of the present invention is limited by appended claims.

Claims (19)

1, a kind of server that is used for the legal research system comprises:

Be used to receive the device that sends by one or more client access device at the request of program legal precedent history; And

Be used in response to one of them received request and the device of definition user interface on corresponding to one of them described client access device of described one of them received request at least in part, described user interface comprises:

Two or more zones, each zone wherein is corresponding to the corresponding level of system of court; And

Two or more hyperlink legal precedent tabulations, each hyperlink legal precedent tabulation wherein can be selected to start the retrieval to corresponding legal precedent document, one of them hyperlink legal precedent tabulation is positioned at first described zone, and second hyperlink legal precedent tabulation is positioned at second described zone.

2, the described server of claim 1, wherein, described user interface also comprises one or more route indicators, one of them described route indicator is indicated first hyperlink legal precedent tabulation and second the hyperlink legal precedent path between tabulating.

3, the described server of claim 1, wherein, described user interface also comprises one or more legal precedent validity designators, and wherein the position of each designator is adjacent with corresponding one of them described hyperlink legal precedent tabulation, and indicates the precedent value or the state of corresponding legal precedent.

4, the described server of claim 1, wherein, the described device that is used for defining at least in part user interface comprises:

Rule set is used for organizing described hyperlink legal precedent tabulation according to the grid with row and column; And

Be used to push one or more described row so that dwindle the device of the visible perception of described grid.

5, the described server of claim 1, wherein, each in the middle of described two or more zones is rectangle basically, and is presented with the level sensing.

6, the described server of claim 1, wherein, described graphical user interface utilizes Microsoft Windows Graphical Device Interface to present.

7, the described server of claim 1, wherein, each hyperlink legal precedent tabulation is associated with the rolling interface characteristics, and it causes exporting the enlarged image around the rectangular area of described hyperlink legal precedent tabulation.

8, a kind of method comprises:

Reception is by the request at program legal precedent history that client access device sent; And

In response to received request, on this client access device, define user interface at least in part, described user interface comprises:

Two or more zones, each zone wherein is corresponding to the corresponding level of system of court; And

Two or more hyperlink legal precedent tabulations, each hyperlink legal precedent tabulation wherein can be selected to start the retrieval to corresponding legal precedent document, one of them hyperlink legal precedent tabulation is positioned at first described zone, and second hyperlink legal precedent tabulation is positioned at second described zone.

9, the described method of claim 8, wherein, defining described user interface comprises: described user interface is defined as comprises one or more route indicators, one of them described route indicator is indicated the path of first hyperlink legal precedent tabulation between tabulating with second hyperlink legal precedent.

10, the described method of claim 8, wherein, defining described user interface comprises: described user interface is defined as comprises one or more legal precedent validity designators, wherein the position of each designator is adjacent with corresponding one of them described hyperlink legal precedent tabulation, and indicates the precedent value or the state of corresponding legal precedent.

11, a kind of graphical user interface that is used for the legal research system, described interface comprises:

Two or more zones, each zone wherein is corresponding to the corresponding level of system of court; And

Two or more hyperlink legal precedent tabulations, each hyperlink legal precedent tabulation wherein can be selected to start the retrieval to corresponding legal precedent document, wherein first hyperlink legal precedent tabulation is positioned at first described zone, and second hyperlink legal precedent tabulation is positioned at second described zone.

12, the described interface of claim 11 also comprises one or more route indicators, and one of them described route indicator is indicated first hyperlink legal precedent tabulation and second the hyperlink legal precedent path between tabulating.

13, the described interface of claim 11, wherein, described user interface also comprises one or more legal precedent validity designators, and wherein the position of each designator is adjacent with corresponding one of them described hyperlink legal precedent tabulation, and indicates the precedent value or the state of corresponding legal precedent.

14, the described interface of claim 11, wherein, each in the middle of described two or more zones is rectangle basically, and is presented with the level sensing.

15, the described interface of claim 11, wherein, described interface utilizes vector graphics to present.

16, the described interface of claim 11, wherein, each hyperlink legal precedent tabulation is associated with the rolling interface characteristics, and when the cursor that utilizes indicating equipment rolled this feature, it caused exporting the enlarged image around the rectangular area of described hyperlink legal precedent tabulation.

17, a kind of graphical user interface that is used for the legal research system, described interface comprises:

Two or more hyperlink legal precedent tabulations, each hyperlink legal precedent tabulation wherein can be selected to start the retrieval to corresponding legal precedent document; And

One or more route indicators, one of them described route indicator are indicated first hyperlink legal precedent tabulation and second the hyperlink legal precedent path between tabulating.

18, the described interface of claim 17 also comprises:

Two or more zones, each zone wherein are corresponding to the corresponding level of system of court, and wherein first hyperlink legal precedent tabulation is positioned at first described zone, and second hyperlink legal precedent tabulation is positioned at second described zone.

19, the described interface of claim 18, wherein, each in the middle of described two or more zones is rectangle basically, and is presented with the level sensing.

Images