JP5629976B2 - Patent specification evaluation / creation work support apparatus, method and program - Google Patents

Description

  The disclosed technique relates to a technique for supporting evaluation and creation work of a newly created patent specification.

Applying and patenting useful ideas as patents is very important in a company's patent strategy.
However, when an idea is licensed, it must be examined by the JPO, and it is often judged that a patent should not be granted by this examination. In this case, the JPO will send a notice of reasons for refusal to the inventor. However, the inventor can refute the notice of reasons for refusal and amend the claims. Submitted to the Agency. However, this process takes time and document creation costs. This cost can be reduced if it is approved as a patent in the initial examination, but at present, the strategy is often to consider the response after receiving a notice of reasons for refusal. This can be said to be a strategy based on the judgment that the cost of conducting a thorough known example search to clear the initial examination is likely to exceed the cost of the above-described process.

  In the notice of reasons for refusal, the description of the patent specification subject to examination, the reasons for refusal of the description, and the description in the well-known example that became the basis for it shall be stated, and the subject of examination cannot be recognized as a patent. It is clearly stated. From the perspective of the inventor, it is only necessary to first examine the contents listed as reasons for refusal and decide on countermeasures such as amendment of claims. An efficient response can be made. This is the basis for selecting the above-mentioned strategy, and conversely, the information described in the notice of reasons for refusal originally obtained by filing a request for examination is anticipated, and the information before filing a patent application It is very useful if you can get it.

  The following is known as the first prior art. That is, first, search for past patent gazettes with similar descriptions (referred to in this application as “patent gazettes” without distinguishing between published patent gazettes and patent publication gazettes). Is done. Then, based on the examination result in the search result, the examination result regarding the new patent specification is expected. A function for presenting a notice of reasons for refusal regarding past patent publications retrieved to the user is also disclosed.

As a second conventional technique, a technique for analyzing a description of claims and estimating a dependency relationship between claims is known.
As a third conventional technique, there is also known a technique for dividing a description into its constituent elements in the description of claims.

  The following prior art documents are disclosed as conventional techniques related to the technique disclosed in the present application.

JP 2008-1171010 A JP 2007-317164 A JP 2003-308318 A

  However, in the first prior art described above, it is often impossible to obtain the description content of the notice of reason for refusal that is expected for the input new patent specification. For example, a general example in a notice of reasons for refusal that a known example for each component described in a claim in the specification of a patent to be examined is given and the patent to be examined is merely a combination of existing technologies and cannot be patented. There is a pattern of rejection. As information corresponding to such a rejection pattern, it is necessary to obtain the description in the publicly known example for each described component. However, in the notice of reasons for refusal to the past patent publications whose contents are similar to the new patent specifications, it is not guaranteed that a publicly known example is given for each individual component. In addition, when individual components described in the new patent specification are described in different patent gazettes, in the similar search using the entire description of the new patent specification as a key, each component is There is a high possibility that the described patent publications will not hit.

  In addition, the description range of the patent specification to be examined referred to in the notice of reasons for refusal (hereinafter referred to as “pointed unit”) is determined only by a publicly known example found by the examiner. In the notice of reasons for refusal, from the case where all the claims of the patent to be examined are based on the same idea as a point of indication and the basis for refusal of right is given, each of the components described in the claims There are various possibilities up to the case where the grounds for refusal of rights are given based on known examples. In consideration of such circumstances, there is no technique for estimating the indicated unit in the notice of reasons for refusal expected for a new patent specification.

  Further, the second and third prior arts described above have the following problems. That is, for example, a claim in a new patent specification is composed of three elements A, B, and C, and a known example X having A and B as constituent elements and a known example including C as a constituent element. Consider the case where Example Y is found. In this case, it is appropriate for the examiner to point out the descriptions of the constituent elements A and B and the description of the constituent element C, respectively, and to give the description of the corresponding parts in the known examples X and Y as the grounds for the rejection. In the method of dividing such a rejection pattern into units of components and presenting similar search results using the description of each component as a key, the known example X is ranked higher than the similar search results. And the possibility of leaking from the search result increases. In particular, in the case where patentability is claimed where the component A and the component B are combined, it is fatal that the known example X leaks from the search result.

  In this way, when assuming a system that presents the contents of a notice of reasons for refusal that is expected for a new patent specification to the user, which part of the description of claims in the input specification is expected The prior art could not estimate whether it would be the indication unit in the notice. This is because it depends on the description content in the found public example.

  The problem to be solved by the disclosed technology is that the indication unit in the specification of the reason for refusal expected for the new patent specification is processed by the processing based on the unit of indication described in the previous notification of reasons for refusal. It is to be able to estimate appropriately.

  With regard to the description of the scope of claims in the application documents, the disclosed technology associates and accumulates past applications for which examination requests have been made and notices of reasons for refusal issued for the applications subject to examination. By searching the patent examination database, it is realized as an apparatus, method, or program that supports the evaluation and creation of the application documents. When the apparatus is realized as an apparatus, it has the following configuration.

  The data preparation unit extracts the notice of reasons for refusal from the patent examination database, and from the result of morphological analysis of the notice of reasons for refusal taken and the predetermined rules, the application to be examined and its application described in the notice of reasons for refusal Information that extracts the description of the claim in which the reason for refusal is indicated in the application, the text that is the basis for the reason for refusal, the cited document indicated by the reason for refusal, and the wording described in the cited document, and links the extracted information And store it in the rejection reason database.

  The refusal reason database retrieval subsystem reads the application documents, searches the description of the claims of the refusal database similar to the description of the claims of the application documents as a key, and describes the claims of the application documents. The information extracted from the notice of reasons for refusal issued for the application having the description of the claims similar to the description of the scope of claims searched in the database of reasons for refusal is presented.

  According to the disclosed technology, the indication unit in the specification in the notification of reasons for refusal expected for the new patent specification is appropriately estimated by processing based on the indication unit in the notification of reasons for refusal in the past. It becomes possible.

  Furthermore, by presenting the search result of the reason database for rejection with respect to the estimated indication unit, the description in the known example corresponding to the indication unit, and further, the citation that became the basis for refusing the right of the description in the known example Presents useful information for reviewing and correcting the input new patent specifications, such as descriptions in the literature, corrections to the description in the above-mentioned known examples, objections to the rejection notice, and final disposal results. For example, by knowing how a description in a publicly known example similar to the indicated description has been corrected and registered, it is easy to appropriately correct the indicated portion in the new patent specification. Can be done.

It is a block diagram of embodiment of a patent specification evaluation and preparation work assistance apparatus. It is an operation | movement flowchart which shows a rejection reason DB storage data creation process. It is an operation | movement flowchart which shows the process of a rejection reason DB search subsystem. It is a figure which shows the example of the notice of reasons for refusal. It is a figure which shows the example of a written opinion. It is a figure which shows the example of Claim 1 in the new patent specification made into a process target. It is a figure which shows the example of an output result (initial state) of the system of embodiment. It is a figure which shows the example of an output result (search result display) of the system of embodiment. It is Table 1 which shows the information extracted from the notice of reasons for refusal of FIG. It is Table 2 which shows the information which added the extraction result from the opinion book of FIG. 3 to Table 1. FIG. Table 2 shows information obtained by analyzing pointer information, deleting terminology, and assigning final disposal results. 5 is Table 4 showing new information created from information 2 in Table 3. It is Table 5 which shows the example of the data stored in rejection reason DB. 7 is a table 6 showing an example of a search unit management table corresponding to the claim example of FIG. 6. 10 is a table 7 illustrating an example of a search result management table. It is a hardware block diagram which implement | achieves the system of embodiment.

Hereinafter, embodiments will be described in detail.
FIG. 1 is a configuration diagram of an embodiment of a patent specification evaluation / creation work support apparatus.
This apparatus includes a patent examination database (patent examination DB 101) 101, a rejection reason database (rejection reason DB 102) 102, a data creation unit 103 that creates data to be stored in the rejection reason DB 102 , and a rejection reason DB search subsystem 104. The

Further, the data creation unit 103 includes a rejection reason notification analysis unit 103-1, a pointer information analysis unit 103-2, a data shaping unit 103-3, and an opinion document analysis unit 103-4.
The rejection reason notification analysis unit 103-1 analyzes the reason for rejection notification and extracts information necessary for creating information to be stored in the rejection reason DB 102.

The pointer information analysis unit 103-2 acquires the substance of information stored in the rejection reason DB 102 based on the information extracted by the rejection reason notification analysis unit 103-1.
The data shaping unit 103-3 shapes the information stored in the rejection reason DB 102 by the information created by the pointer information analysis unit 103-2. Then, the data shaping unit 103-3 divides the description content of the published patent publication of the application for which the request for examination rejected by a plurality of cited documents is made in accordance with the description of the cited documents, and creates a new DB registration record. create.

The opinion document analysis unit 103-4 analyzes the opinion document or the correction document, and extracts information stored in the rejection reason DB 102.
The rejection reason DB search subsystem 104 includes a claim description division unit 104-1, a search unit generation unit 104-2, a search unit management table 104-3, a similarity search unit 104-4, a search result management table 104-5, an indication A unit estimation unit 104-6 and a result display unit 104-7 are included.

The claim description dividing unit 104-1 extracts the claim description of the input new patent specification, determines the minimum unit for searching the rejection reason DB 102 by dividing the description.
The search unit generation unit 104-2 generates the minimum unit created by the claim description division unit 104-1, and all combinations thereof.

The search unit management table 104-3 manages the search units generated by the search unit generation unit 104-2.
The similarity search unit 104-4 performs a similar search in the rejection reason DB 102 using the search unit character string generated by the search unit generation unit 104-2 as a search key.

The search result management table 104-5 stores search result information for each search unit in the similarity search unit 104-4.
The indication unit estimation unit 104-6 determines a description portion that is supposed to be referred to in the expected rejection reason notification based on the information in the search result management table 104-5.
The result display unit 104-7 presents the description portion estimated by the indication unit estimation unit 104-6 and the search result when the description portion is used as a search key to the user.

The operation of the embodiment having the configuration of FIG. 1 will be sequentially described below.
The operation of the embodiment includes two phases, a data creation phase and a new patent specification evaluation phase. In the data creation phase, data stored in the rejection reason DB 102 is automatically created from the records in the patent examination DB 101. In the new patent specification evaluation phase, a new patent specification is input, a description portion that is likely to be rejected is estimated, and presented to the user together with a record in the rejection reason DB 102 that is likely to be a basis. Hereinafter, the present embodiment will be described in the order of the data creation phase and the new patent specification evaluation phase.

<Data creation phase>
FIG. 2 is an operation flowchart showing a process for automatically creating data stored in the rejection reason DB 102 from the patent examination DB 101. The operation of each processing unit included in the data creation unit 103 in FIG. 1 will be described below according to this operation flowchart.

Patent Examination DB101
The patent examination DB 101 stores, as one record, information obtained by associating a notice of reasons for refusal, a cited document, and an opinion document together with final disposal information.

The notice of reasons for refusal is a document notifying the reasons for refusal that cannot be assessed as a result of examination when there is a reason that the patent cannot be assessed for the patent application for which the request for examination has been made.
The cited document is a document cited in the notice of reasons for refusal as the basis for the reasons for refusal.

The opinion is a document that describes the amendment made by the inventor who received the notice of reasons for refusal and the contents of amendment and objection.
The patent examination DB 101 can be realized by a general database system.

The above-mentioned document information has been difficult to obtain in the past. However, since the JPO has released “organized standardized data”, it can now be easily obtained.
The information stored in the patent examination DB 101 may store each document data itself, but stores only reference information to an external database and accesses an external DB connected via the Internet as necessary. It may be configured to acquire an actual document.

  The data creation unit 103 acquires the first record from the patent examination DB 101 (step S201 in FIG. 2). If acquisition is successful, the following series of processing from step S203 to S208 is executed (YES in step S202 of FIG. 2).

Rejection Reason Notice Analysis Unit 103-1
The rejection reason notification analysis unit 103-1 analyzes the contents of the rejection reason notification and extracts information for creating information to be stored in the rejection reason DB 102 (step S 203 in FIG. 2). This analysis process will be described with reference to FIG. 4, which is a typical rejection notice.

Step 1
The rejection reason notification analysis unit 103-1 performs layout analysis on the rejection reason notification. The notice of reasons for refusal has a standard layout, and the written text often uses standard phrases. For this reason, the rejection reason notification analysis unit 103-1 uses the information on the standard wording regarding the reason for rejection notification, and the rule for grasping the description content of the sentence described in the notification of reason for rejection to be analyzed Is provided. Examples of such rules include the following.

The line starting with rule 1 1- "patent application number" and containing a blank (or tab delimiter) that matches the description pattern of the patent gazette is the number of the patent gazette that is the object of the reason for refusal Represents. For example, it is a description pattern such as “JPNNNNN-NNNNNN”, “JPNNN-NNNNNN”, and “Japanese Patent Application NNNN-NNNNNN”. Note that each N represents a number. Also in the following description, when N is used in the description of the pattern, it is assumed to represent a number in the same manner.

If there is a line containing only the character string “2-reason” in Rule 1, the paragraph following that line is a paragraph describing the reason for refusal, and the Patent Law, which is a legal provision for the reason for refusal The character string which represents the applicable rule in is included.

If there is a line containing only the character string “3-” in Rule 1, the paragraph following that line shall be a reference to the description content part of the published patent publication of the application for which the request for examination was rejected. The details of the reasons for refusal based on the description of the cited reference that became the basis for the refusal are described.

In the paragraph that describes the detailed contents of the 4-rejection reason of Rule 1, if there is an item format that begins with a character string that includes numbers such as “•” and “(N)”, one item part Describes the details of the reasons for refusal for one description of the published patent publication of the application for which the request for examination has been made.

Based on the rules as described above, the description content of each paragraph in the notice of reasons for refusal is determined by the notice of reason for refusal analyzer 103-1. In the example shown in FIG. 4, the following information is recognized.

“JPAAA-ABBBBBB” is the number of the patent publication that is the subject of examination.

The reason for refusal is described in the paragraph following “reason” from “this application” to “cannot receive a patent”.

Each bullet paragraph following the "SL", specifically the paragraph leading to "(1)""- described Ru Tei" from paragraph reaching the "(2)", "a-technology" from The details of the reasons for refusal are described respectively.

Step 2
Next, the rejection reason notice analysis unit 103-1 extracts information based on the description content determined for each paragraph based on the determination result of the layout analysis described above. When extracting the information, the following rules are applied for each description content.

For the description part 1-rejection reason of Rule 2, the partial character string beginning with “Patent Law No.” in the description sentence is subjected to morphological analysis. The reason up to the morpheme analyzed as “particle” in the morpheme analysis result is extracted as the reason for rejection. In the example of FIG. 4, “Patent Act Article 29 Paragraph 2” is extracted as the reason for refusal.

For the part that details the reason for refusal 2 in Rule 2, the description part in the published patent publication to be examined and the description in the cited document that became the basis for the rejection are extracted for each paragraph. The In this extraction, extraction processing using the following processing rules is performed.

Rule 2-2 1-Extraction of description part in a public patent gazette to be examined A description pattern indicating a rejected part in a public patent gazette to be examined (for example, “Claim N”) and the contents are described Extraction rules based on typical phrases (for example, "(character string matching the descriptive pattern indicating the rejected part) is described as" ... "") The contents in the published patent publication to be examined are extracted. In the first clause paragraph in the example of FIG. 4, “claim 1” is extracted as information (pointer information) indicating the description part in the public patent gazette to be examined which has been rejected. The rules to extract do not match.

2. Extraction of description part in known example that is the basis of refusal 2 of rule 2 Citation example description pattern (for example, “citation example N”, “cited document N”, “prior application N”, etc.) and its entity Correlation processing with information is performed. In the example of FIG. 4, “Citation 1” , “JPXXXX-YYYYY Publication”, “Citation 2” and “Actual Application No. XXXXX- It is matched.

  Next, the following processing rule is applied to a typical wording, for example, a wording “(written in a character string that matches the citation example description pattern)”. That is, a character string representing an entity associated with “(character string matching the citation example description pattern) is extracted as pointer information indicating a description location of the citation example, and a portion of“... Extracted as description.

In addition, the following processing rule is applied to the phrase “... Is shown in (a character string that matches the citation example description pattern)”. That is, a character string representing an entity associated with “(character string matching the citation example description pattern)” is extracted as pointer information indicating a description location of the citation example, and “...” is the portion of the citation example. Extracted as description. In the example of FIG. 4, this processing rule is applied.
In addition, regarding the last verb of each processing rule, it may be assumed to match one of the specific verb sets defined in advance, such as “described”, “shown”, and “described”. .

  In addition, a processing rule for simultaneously extracting “description part in a public patent gazette to be examined” and “description in a publicly known example as a basis for rejection” is also applied. For example, “(a character string that matches the description pattern representing the rejected part) is a conventionally known technique (a character string that matches the cited example description pattern) [N1], [N2],. "Is as follows. First, “(a character string that matches a description pattern representing a rejected part)” is extracted as pointer information indicating the rejected part in the published patent publication of the application for which examination is requested. Furthermore, “the portion described in paragraph symbols [N1] and [N2] in the information representing the entity associated with (character string matching the cited example description pattern)” is pointer information indicating the description content of the cited document. Extracted.

   As a result of applying the above rule 2-2 1 and rule 2-2 2 to the rejection reason notice shown in FIG. The information as shown in Table 1 in FIG. 9 is acquired regarding “the description part of” and “the description in the known example that has become the basis for the rejection”. The information that has not been extracted is described as “-”. Here, with respect to information 1 in Table 1, one cited document is linked, and with respect to information 2, two cited documents are linked. In order to clarify this, for the sake of convenience, a number is added to the end of the “pointer information to the description in the cited document” item and the “description content” item. For example, “pointer information 1 to description in cited document” “pointer information 2 to description in cited document”.

Opinion analysis unit 103-4
The opinion analysis unit 103-4 analyzes the opinion written by the applicant (right holder) with respect to the reason for rejection notification, and further links the information extracted from the reason for rejection notification. Information is extracted (step S204 in FIG. 2). This analysis process will be described with reference to FIG. 5 which is a typical opinion document.

Step 1
The opinion document analysis unit 103-4 performs layout analysis on the opinion document. By this layout analysis, the paragraph describing the content of the opinion is determined. For example, ““ [Content of Opinion]
A determination rule such as “determining that the content of the opinion is described in the paragraph below the line described as“ is used ”is used.

  If the paragraph describing the opinion matches a bullet pattern that starts with a character string such as “•” or “(N)”, the detailed reason for refusal corresponding to each item in the bullet point location It is determined that it is a statement of opinion in the explanation part. In this process, it is usually considered that the opinion is clearly related to the notice of reasons for refusal that triggered the creation of the opinion. A description pattern similar to the pattern is often used. For this reason, it is relatively easy to determine the correspondence. In addition, even if the description pattern of the clauses is different, it is possible to reject from the standard description pattern that clearly indicates the corresponding numerical part and the part in the notice of reasons for refusal as the first sentence of the clause part. Correspondence with the notice of reason can be taken. For example, in the notice of reasons for refusal, if the opinion corresponding to the part listed in the form of a character string including numbers, for example (1), is listed in the form of “opinion 1”, it corresponds. Correspondence can be taken by matching the numerical part. In addition, as the first sentence of the clause part, for example, “about the description of ~ in the notice of reasons for refusal” or “the examiner is in the notice of reasons for refusal (~)” It can be determined that the item to be written is described.

  In the example of the opinion book in FIG. 5, the corresponding opinion is described in the same itemized form as the notice of reasons for refusal. Furthermore, the typical description pattern “about“ claim N ”as pointed out in the notice of reasons for refusal” can be associated with the corresponding part in the notice of reasons for refusal.

  In addition, a plurality of indications regarding the notice of reasons for refusal may be collectively described as one amendment content in the opinion document without distinction. In many cases, the amendment at this time is to replace all the descriptions in the claims described in the published patent gazette of the application for which the examination request has been made. In such a case, the above rules cannot be dealt with. However, if the notice of reasons for refusal states details of the reasons for refusal for each claim, the correspondence can be made by matching the numbers of the claims, so such a correspondence rule is also provided. Also good.

  With the association rules as described above, the paragraphs describing the correction contents are associated with each piece of information extracted from the rejection reason notice. If there is no match with the description pattern of the association rule, the description paragraphs with the same correction content are associated with all the extracted information from the rejection reason notification.

Step 2
The opinion document analysis unit 103-4 extracts the content of the objection and the correction content from the paragraph describing the opinion obtained as the layout analysis determination result. In this extraction, a processing rule based on a description pattern of a typical phrase is used.

  For example, “procedure amendment N of procedure amendment” is extracted as pointer information indicating the amendment from a description pattern “amendment of“ ... ”described in procedure amendment N of procedure amendment”. The "..." part is extracted as a description of the correction content. The example of FIG. 5 corresponds to this rule.

  In addition to the rules applied to one sentence as described above, there are also processing rules that span multiple sentences. For example, in a paragraph that follows the line that says “[contents of opinion]”, there is a sentence that includes the expression “... When a paragraph with a line break (such as “Claim N”) continues, “Claim N” is extracted as pointer information indicating the correction content, and the subsequent portion is extracted as a description of the correction content.

  If it does not match the above extraction rule, the entire description of the paragraph describing the opinion is extracted. This is an effective processing rule for extracting the content of the objection when the objection is rejected without correction.

  In the example of the written opinion of FIG. 5, the above-described processing is characterized in that “...” is filled in as a correction content for the paragraph (1) in which the details of the reason for refusal are stated in the reason for refusal. "Concrete formwork" is linked. In addition, for the paragraph (2), the correction content is “to form the molds, interpose an elastic body between the molds, and attach a removable fixing part between the molds. The concrete formwork according to claim 1 is characterized. Then, information as shown in Table 2 in FIG. 10 is obtained in which the extraction results from the opinion sheet are added to the information extracted from the reasons for refusal shown in Table 1 in FIG.

Pointer information analysis unit 103-2
In the information extracted by analyzing the notice of reasons for refusal and the opinion, the actual description content cannot be acquired, and only the pointer information may be acquired. In such a case, the pointer information analysis unit 103-2 acquires the actual content based on the pointer information (step S205 in FIG. 2).

  For example, in the information 1 shown in Table 2 of FIG. 10, the item “description in patent to be examined” is only pointer information, and the description content is not extracted. Therefore, the primary information in the patent examination DB 101 is accessed using the information “Claim 1” of “JPAAA-BBBBB” as pointer information, and the description content is acquired. As a result, description contents as shown in Table 3 of FIG. 11 are obtained.

  The pointer information used as a clue by the pointer information analysis unit 103-2 is not always the one that obtains the description of such a claim. For example, the description content stored in the rejection reason DB 102 can be acquired by a combination of information specifying the document and information indicating the description position in the specified document. Information specifying the document includes the title of the paper in addition to the publication number of the patent mentioned in the example. In the patent examination DB 101, since cited documents are already linked, it is easy to specify a document to be extracted. In addition to the character string attached to the line immediately before the paragraph to be extracted, such as “Claim 1” and “Procedure Correction 1” as information indicating the description position in the specified document, for example, patent Some are based on paragraph marks given in the gazette. In many cases, the cited paper or the like points to the corresponding part by the page number and the paragraph number, but this part can also be specified by a simple layout analysis.

Data shaping unit 103-3
The data shaping unit 103-3 processes the information extracted and linked as described above, and shapes the information into the format stored in the rejection reason DB 102 (step S206 in FIG. 2).

  First, the data shaping unit 103-3 deletes the non-terminology with respect to the description content of the extracted / linked information. Examples of the terminology include “as described in claim N” in the description of claims. Also, the part representing the itemized list is deleted as a non-term.

  In addition, when a registered patent is associated with one record of the patent examination DB 101 to be processed, the data shaping unit 103-3 associates “registration” with the extracted information as a final disposal result, and registers the registered patent. If there is not, “withdrawal”, or a state indicating an unassessed state during examination of the applicant's response waiting or response result, or a state indicating a state in which an appeal is being requested, is associated. In the example of Table 2 in FIG. 10, when a registered patent gazette exists, it is shaped into information as shown in Table 3 in FIG.

  Further, as shown in information 2 of Table 2 in FIG. 11, when a description in one examination-targeted patent is rejected by a plurality of citations, the data shaping unit 103-3 determines the description in each cited document. To determine whether the description in the examination patent can be divided. Then, if the data shaping unit 103-3 can be divided, the data shaping unit 103-3 also performs a process of creating the division result as new information. For example, the description in the patent to be examined in the information 2 of Table 2 in FIG. 11 is "Concrete formwork characterized in that the formwork is ... and the elastic body is interposed between the formwork". Suppose there was. The description 1 in the cited document is “mould between molds ...” and the description 2 is “insert an elastic body between the molds”, and the corresponding description is included in the description in the patent to be examined. Since it exists, it is determined that it can be divided. As a result, information 2-1 and information 2-2 as shown in Table 4 of FIG.

Reason for rejection DB102
The rejection reason DB 102 stores information automatically created from one record in the patent examination DB 101 (step S207 in FIG. 2).

  The information registered in the database is information 2-1 and information 2-2 newly created from information 1, information 2, and information 2 in the example of Table 3 in FIG. 11 and Table 4 in FIG. There are four. An example of data stored in the rejection reason DB 102 is shown in Table 5 of FIG. In Table 5, the ID corresponds to the number of the extracted information for simplification of explanation, but may actually be another form of value.

  In Table 5, only the “description content” of the extracted information is registered in each record of the database, but it may be one record including pointer information to each description content. In this case, when a record in the rejection reason DB 102 is presented to the user as a processing result for the new patent specification, the user clicks, for example, one item in the record. Then, the extraction source document is retrieved by the pointer information corresponding to the item and presented to the user. Such a function can be provided in the rejection reason DB search subsystem 104.

  As described above, when the data creation unit 103 finishes the storage process in the rejection reason DB 102 for one record extracted from the patent examination DB 101, the data creation unit 103 acquires the next one record from the patent examination DB 101 (step S208 in FIG. 2). The process similar to the above is repeatedly executed. When there are no more records acquired from the patent examination DB 101, the data creation unit 103 ends the process (NO in step S202 of FIG. 2).

<New patent specification evaluation phase>
FIG. 3 is an operation flowchart showing the new patent specification evaluation phase process executed by the rejection reason DB search subsystem 104. In this phase, the new patent specification is input, the description of the claim is analyzed, and the portion that is likely to receive the rejection indication in the expected notification of reasons for refusal is recorded together with the record in the reason for rejection DB 102 that is likely to be the basis. Presented to.

  The rejection reason DB search subsystem 104 first inputs a new patent specification (step S301 in FIG. 3), sets a variable N = 1 for counting claims (step S302 in FIG. 3), and claims the new patent specification. The description of N is acquired (step S303 in FIG. 3). When the rejection reason DB search subsystem 104 succeeds in the acquisition, the rejection reason DB search subsystem 104 executes a series of processes from the following steps S305 to S313 (YES in step S304 in FIG. 3).

Claim description division unit 104-1
The claim description division unit 104-1 divides the description into minimum units for performing the similarity search process for each claim in the input new patent specification (step S305 in FIG. 3).

  The division rules used in this process include punctuation marks, or “characteristic” (or similar expressions such as “having”, “having”, etc.) The division is performed by. For example, a simple rule such as “divide immediately before the occurrence of this wording and do not divide at the reading point when a reading comes immediately after this wording” may be used.

For example, when the claim of the patent specification illustrated in FIG. 6 is input, the following division is performed according to the above rule. Further, since the minimum unit including “characterized by” is a portion describing the object of the invention, special processing is performed in the subsequent processing.

A sheet metal formwork panel;
-It consists of a separator that keeps the space between opposing formwork panels,
-The above-mentioned formwork panel is ...
-With holes in parallel on one side of the front and back sides,
-Concrete formwork characterized by having a punched hole on the other side surface-

Note that the split process is not a process based on the simple rules as described above, but the description of the claim is analyzed in more detail, and the description is divided into constituent element units, or only the description part that claims patentability. It is also possible to apply an existing technique in which the minimum unit for performing the similar search process is determined by dividing the portion.

Search unit generation unit 104-2
The search unit generation unit 104-2 creates a unit for searching for a record in the rejection reason DB 102 and a description serving as a key for the unit based on the minimum search unit created by the claim description division unit 104-1 (FIG. 3). Step S306).

  More specifically, the search unit generation unit 104-2 generates the minimum search unit divided by the claim description division unit 104-1, and all combinations thereof. Next, the search unit generation unit 104-2 appropriately generates a search unit ID and a minimum search unit constituting the search unit, and stores the generated search key together with the created search key in the search unit management table 104-3.

However, since the minimum unit including “characterized by” is a part describing the object of the invention, the search unit generation unit 104-2 has the search unit management table 104-3 for this part. Is not included in the minimum search unit combination.
The search unit management table 104-3 having the contents shown in Table 6 of FIG. 14 is generated by the above processing, for example, corresponding to the claim example of FIG.

  Next, the rejection reason DB search subsystem 104 sequentially retrieves each search unit from the search unit management table 104-3 generated as described above (steps S307 and S311 in FIG. 3) until there is no search unit. (The determination in step S308 in FIG. 3 is YES), the following steps S309 and S310 in FIG. 3 are repeatedly executed.

Similarity search unit 104-4
The similarity search unit 104-4 includes a search key corresponding to the search unit extracted from the search unit management table 104-3 and the description content of the published patent gazette of the application for which examination is requested in each record of the rejection reason DB 102. Similarity is calculated, and a record with high similarity is acquired as a search result (step S309 in FIG. 3).

As the similarity calculated here, for example, a cosine measure in a vector space model, which is widely used in general, can be used. For the keyword set {Wi}, the degree of similarity is Freq1 (Wi) for the number of occurrences of the keyword Wi in the search key, and Freq2 (Wi) for the number of occurrences of the keyword Wi in the description contents to be searched. Then, it is calculated by the following formula.

In the similar search, a search key may be obtained by adding a character string describing the subject of the invention to each search unit. In this case, it is possible to obtain an effect that records including character strings such as “concrete” and “formwork” are easily searched.
Furthermore, for each keyword set, the synonym dictionary is used to eliminate the keyword notation fluctuation and calculate the similarity, or the weighting based on the importance of the keyword is performed to calculate the similarity. You may apply the existing precision improvement technique in a similar search.

  The search result in each search unit is acquired by the similar search method as described above. The search result unit ID, its constituent minimum unit ID, the maximum similarity value in the search result, and the next indication unit estimation process are used. The processing flag is 1 record. This record is stored in the search result management table 104-5. The processing flag takes one of three values such as “undecided”, “pointed unit”, and “non-pointed unit”, but when the similar search result is stored, the processing flag is set to “undecided”. . The search result in each search unit is stored on the memory (step S310 in FIG. 3).

  An example of the search result management table 104-5 created as a result of applying the above processing to the search unit management table 104-3 shown in Table 6 of FIG. 14 is shown in Table 7 of FIG.

  When the rejection reason DB search subsystem 104 completes the above-described similar search processing for all search units in the search unit management table 104-3 (NO in step 308 of FIG. 3), the following indication unit estimation is performed. The process of the unit 104-6 is executed.

Pointed unit estimation unit 104-6
The indication unit estimation unit 104-6 estimates the range indicated in the expected rejection reason notification based on the created search result management table 104-5 (step S312 in FIG. 3). Here, the processing of the following substeps is executed.

Substep 1
In the set of search units satisfying the following conditions 1 and 2, those having the maximum similarity of No. 1 and No. 2 are selected (search unit IDs of the selection results are X and Y, respectively).

Condition 1: “Undecided” flag is assigned.
Condition 2: There is no search unit with a search unit finer than that and with an “undecided” flag.

Substep 2
Let Z be a search unit that includes both of the two selected search units and has a processing flag of “undecided”. If the maximum similarity in each search unit is R _X , R _Y , R _Z , the following processing flag change rules are applied in order according to these three values, and processing is performed based on the rules at the time when the conditions are matched. The flag is changed.

(A) If R _Z ≧ R _X ≧ R _Y , change the processing flags of X and Y to “unpointed unit”.
(B) If R _X ≧ R _Z ≧ R _Y , change the processing flag of X to “pointed unit”.
(C) If R _X ≧ R _Y ≧ R _Z , change the processing flag of X and Y to “pointed unit”.

Substep 3
As a result of the processing flag change processing described above, the processing flag is set to “not pointed out” for all the search units that include the minimum search unit constituting the search unit that becomes the “pointed unit” and the processing flag is “undecided”. Changed to “Unit”.

Substep 4
The above processing from sub-steps 1 to 3 is repeated until no further processing can be performed.

Substep 5
If a search unit that remains the “undecided” flag remains, the processing flag of the search unit is changed to “pointed unit” and the process ends.

When the above sub-step process is applied to the search result management table 104-5 illustrated as Table 6 in FIG. 14, the following process is performed.

1. Since all the search units in the table are “undecided”, the set of search units that satisfy the conditions is that whose search unit ID is {1, 2, 3, 4, 5}. Among these, 4 and 5 having the largest maximum similarity are selected (substep 1).

The minimum search unit including both 2.4 (configured from the minimum search unit D) and 5 (configured from the minimum search unit E) is 15 (configured from the minimum search units D and E). Comparing the respective maximum similarities, since R ₁₅ > R ₄ and R ₁₅ > R ₅ , the processing flag at 4 and 5 is changed to “unpointed unit” (substep 2).

3. In the processing in the sub-step 2, there is no processing flag with “pointed unit”, so the processing in the sub-step 3 is not performed.

4). Returning from sub-step 4 to sub-step 1, search candidates are selected again. In this case, the set of search units satisfying the condition is the search unit ID of {1, 2, 3, 15}. In this set, 15 and 3 which are the top two of the maximum similarity are selected.

5. The search unit including both of the selected search units is 25, in which the minimum structural unit is composed of C, D, and E, and this processing flag is “undecided”, so this search unit is selected. When the maximum similarity values in the three selected search units are compared, since R ₁₅ > R ₂₅ and R ₃ > R ₂₅ , the processing flag of the search units 15 and 3 is changed to “pointed unit” ( Sub-step 2).

6). In the above sub-step 2, the processing flags of the search units 15 and 3 are changed to “pointed unit”. For this reason, all the processing flags of the search units including the minimum search units D and E constituting the search unit 15 and having the processing flag “undecided” are changed to “non-pointed unit” (sub-step 3). For example, a search unit composed of minimum search units A and D is applicable. Similarly, the processing flag of the search unit including the minimum search unit C constituting the search unit 3 and having the processing flag “undecided” is changed to “non-pointed unit”. For example, a search unit composed of minimum search units B and C is applicable.

7). Returning from sub-step 4 to sub-step 1, the search unit to be processed next is selected. Since there are only 1 and 2 search units that can be selected at this time, these two are selected.

8). There are only 6 undetermined search units including the selected search unit. When these three search units are compared, R ₂ > R ₆ > R ₁ , so that the processing flag of the search unit 2 is “pointed out” Unit "(substep 2).

9. Since 2 is the “pointed unit” in the substep 2, the processing flag 6 is set as the “non-pointed unit” (substep 3).

10. Even if the sub-step 4 returns to the sub-step 1, the search unit to which the “undecided” flag is assigned is only 1. Therefore, no further processing can be performed. Accordingly, one processing flag is set as the “pointing unit” and the processing ends (sub-step 5).

Through the above processing, the search unit 1 (composed from the minimum search unit A), the search unit 2 (composed from the minimum search unit B), the search unit 3 (composed from the minimum unit C), and the search unit 15 (from the minimum units D and E). Composition) was estimated as the indicated unit.

   After the above-described processing, the rejection reason DB search subsystem 104 increments the value of the variable N for counting claims by 1 (step S313 in FIG. 3), and executes the same processing for the next claim N ( (Repeat after step S303 in FIG. 3).

When there is no claim to be processed (the determination in step S304 in FIG. 3 is NO), rejection reason DB search subsystem 104 executes the following result display process.
Result display section 104-7
The result display unit 104-7 processes the description of the claim item in the input specification based on the search unit estimated by the indication unit estimation unit 104-6, the search result, and the maximum similarity value. Presented to the user (step S314 in FIG. 3).

  The following can be adopted as a method for presenting the processing result. For example, a range of indication units having a maximum similarity exceeding a predetermined threshold value is highlighted. When the user selects and clicks the highlighted portion with a pointing device such as a mouse, the top candidates of the similar search results in the rejection reason DB 102 when the highlighted portion is used as a search key are displayed.

  Further, depending on the estimation result of the indication unit, a plurality of non-consecutive description parts may be one search unit. Even in such a case, when one range is selected, it can be shown to the user by means such as changing the color that the range in which the description part constituting the same search unit is also selected is selected. .

  FIG. 7 shows an initial state of an example of a processing result for the claim example of the new patent specification of FIG. In this example, 0.5 (see FIG. 15) is set as the threshold value to be displayed, but this value may be changeable by the user. Further, as shown in FIG. 7, the description of the input specification may be devised so as to make it easier to see by inserting line breaks in units of indications. Further, by changing the brightness of the highlight based on the maximum similarity of the search results, an idea may be made to make the description part with the search results having a high similarity more conspicuous.

  FIG. 8 is a display example in which the search result of the rejection reason DB 102 is displayed as a balloon by the user's action. In this example, the search result record is displayed in accordance with the similarity. Further, only the records that exceed the specified similarity threshold may be displayed.

FIG. 16 is a diagram illustrating an example of a hardware configuration of a computer that can realize the system of the embodiment illustrated in FIG. 1.
The computer shown in FIG. 16 includes a CPU 1601, a memory 1602, an input device 1603, an output device 1604, an external storage device 1605, a portable recording medium driving device 1606 into which a portable recording medium 1609 is inserted, and a network connection device 1607. However, they are connected to each other by a bus 1608. The configuration shown in the figure is an example of a computer that can implement the above system, and such a computer is not limited to this configuration.

  The CPU 1601 controls the entire computer. The memory 1602 is a memory such as a RAM that temporarily stores a program or data stored in the external storage device 1605 (or the portable recording medium 1609) when executing a program or updating data. The CUP 1601 performs overall control by reading the program into the memory 1602 and executing it.

  The input device 1603 includes, for example, a keyboard, a mouse, and their interface control devices. The input device 1603 detects an input operation by a user using a keyboard, a mouse, or the like, and notifies the CPU 1601 of the detection result.

  The output device 1604 includes a display device, a printing device, etc. and their interface control devices. The output device 1604 outputs data sent under the control of the CPU 1601 to a display device or a printing device.

The external storage device 1605 is, for example, a hard disk storage device. Mainly used for storing various data and programs.
The portable recording medium driving device 1606 accommodates a portable recording medium 1609 such as an optical disk, SDRAM, or Compact Flash (registered trademark), and has an auxiliary role for the external storage device 1605.

The network connection device 1607 is a device for connecting, for example, a LAN (local area network) or WAN (wide area network) communication line.
The system according to the embodiment shown in FIG. 1 is realized by the CPU 1601 executing a program equipped with the functions shown in the operation flowcharts of FIGS. 2 and 3 necessary for the system. The program may be distributed by being recorded in, for example, the external storage device 1605 or the portable recording medium 1609, or may be acquired from the network by the network connection device 1607.

In the above-described embodiment, in the operation flowchart of FIG. 3 executed by the rejection reason DB search subsystem 104, processing is performed for each claim. Therefore, even in the case where one indication unit is the widest, it is within one claim. stay. However, this limitation is added for the sake of clarity in describing one embodiment, and is not an essential limitation of the disclosed technology. In other words, when there are a plurality of claims whose entirety is a single indication unit, the description of those claims is regarded as a minimum search unit, and the same processing as described above is executed. It is also possible to obtain a result in which a plurality of claims are combined into one indication unit.
In addition, various forms can be applied to the processing for realizing each processing unit.

Regarding the above embodiment, the following additional notes are disclosed.
(Appendix 1)
In an apparatus that supports the evaluation and creation of the patent specification by searching a patent examination database that accumulates text information related to past patent applications to be examined for which an examination request has been made with respect to the appendix of the patent specification. ,
Searching and analyzing the notice of reasons for refusal from the patent examination database, analyzing or referring to the public patent gazette subject to examination rejected in the notice of reasons for refusal, A data creation unit that creates information associated with the description or reference information in the cited reference cited for the reason for refusal;
A rejection reason database for storing information created by the data creation unit;
A new patent specification is input, the similar reason is searched in the rejection reason database using the description in the appendix of the new patent specification as a key, and the information in the new patent specification similar to the record information searched from the rejection reason database A refusal reason database search unit that estimates the description range of the information and presents the description range together with the information of the retrieved record;
Patent specification evaluation / creation work support device characterized by including:
(Appendix 2)
The data creation unit searches the patent examination database for an opinion document linked to the notice of reason for refusal and analyzes it, and provides a description or reference information regarding the objection or the contents of amendment to the reason for refusal. Create the information by further linking to the information linked to the book,
The patent specification evaluation / creation work support apparatus according to Supplementary Note 1, wherein the apparatus includes:
(Appendix 3)
The data creation unit obtains a description corresponding to the reference information from the patent examination database or a patent related database connected to the outside, and creates the information as the information.
3. The patent specification evaluation / creation work support apparatus according to any one of appendices 1 and 2, characterized in that:
(Appendix 4)
The rejection reason database search unit includes:
A supplementary description dividing unit that divides the supplementary description of the new patent specification into minimum search units for performing the similarity search processing;
Based on the minimum search unit created by the supplementary description dividing unit, a search unit for searching for a record in the rejection reason database and a search unit generating unit for creating a key description thereof;
The similarity between the search key corresponding to the search unit created by the search unit generator and the description content of the published patent gazette of the application for which examination is requested in each record of the rejection reason database is calculated, and the similarity is high A similar search unit that acquires records as search results;
The patent specification evaluation / creation work support device according to any one of appendices 1 to 3, further comprising:
(Appendix 5)
In a method for supporting the evaluation and creation of the patent specification by searching a patent examination database that accumulates text information relating to past patent applications to be examined for which an examination request has been made with respect to the appendix of the patent specification. ,
Searching and analyzing the notice of reasons for refusal from the patent examination database, analyzing or referring to the public patent gazette subject to examination rejected in the notice of reasons for refusal, A data creation step of creating information associated with the description or reference information in the cited document cited for the reason for refusal, and storing the information in the reason for refusal database;
A new patent specification is input, the similar reason is searched in the rejection reason database using the description in the appendix of the new patent specification as a key, and the information in the new patent specification similar to the record information searched from the rejection reason database Refusal reason database search step for estimating the description range of the information and presenting the description range together with the information of the retrieved record;
Patent specification evaluation / creation work support method characterized by including:
(Appendix 6)
The data creation step searches and analyzes the opinion document linked to the notification of reasons for refusal from the patent examination database, and provides a description or reference information relating to the objection to the reason for refusal or amendments, or the reference information. Create the information by further linking to the information linked to the book,
The patent specification evaluation / creation work support method according to appendix 5, characterized in that:
(Appendix 7)
In the data creation step, a description corresponding to the reference information is acquired from the patent examination database or a patent-related database connected to the outside and created as the information.
The patent specification evaluation / creation work support method according to any one of appendix 5 or 6, characterized in that:
(Appendix 8)
The rejection reason database search step includes:
A supplementary description dividing step of dividing the supplementary description of the new patent specification into a minimum search unit for performing the similarity search processing;
Based on the minimum search unit created by the additional description division step, a search unit for searching for a record in the rejection reason database and a search unit generating step for creating a key description thereof;
The similarity between the search key corresponding to the search unit created by the search unit generation step and the description content of the published patent gazette of the application for which examination is requested in each record of the rejection reason database is calculated, and the similarity is high A similar search step that retrieves records as search results;
The patent specification evaluation / creation work support method according to any one of appendices 5 to 7, further comprising:
(Appendix 9)
By searching a patent examination database in which text information relating to past patent applications subject to examination for which examination requests have been made with respect to the appendix of the patent specification, a computer supporting the evaluation and creation of the patent specification is obtained. ,
Searching and analyzing the notice of reasons for refusal from the patent examination database, analyzing or referring to the public patent gazette subject to examination rejected in the notice of reasons for refusal, A data creation step of creating information associated with the description or reference information in the cited document cited for the reason for refusal, and storing the information in the reason for refusal database;
A new patent specification is input, the similar reason is searched in the rejection reason database using the description in the appendix of the new patent specification as a key, and the information in the new patent specification similar to the record information searched from the rejection reason database Refusal reason database search step for estimating the description range of the information and presenting the description range together with the information of the retrieved record;
A program for running
(Appendix 10)
The data creation step searches and analyzes the opinion document linked to the notification of reasons for refusal from the patent examination database, and provides a description or reference information relating to the objection to the reason for refusal or amendments, or the reference information. Create the information by further linking to the information linked to the book,
The program according to appendix 9, characterized by:
(Appendix 11)
In the data creation step, a description corresponding to the reference information is acquired from the patent examination database or a patent-related database connected to the outside and created as the information.
The program according to any one of appendixes 9 and 10, characterized in that:
(Appendix 12)
The rejection reason database search step includes:
A supplementary description dividing step of dividing the supplementary description of the new patent specification into a minimum search unit for performing the similarity search processing;
Based on the minimum search unit created by the additional description division step, a search unit for searching for a record in the rejection reason database and a search unit generating step for creating a key description thereof;
The similarity between the search key corresponding to the search unit created by the search unit generation step and the description content of the published patent gazette of the application for which examination is requested in each record of the rejection reason database is calculated, and the similarity is high A similar search step that retrieves records as search results;
The program according to any one of appendices 9 to 11, further comprising:

  The disclosed technology is a patent specification management software system. More specifically, a new patent specification is evaluated by evaluating claims of a new patent specification based on descriptions of reasons for refusal, opinions, amendments, etc. It can be used for software systems that support the creation of

101 Patent Examination DB
102 Reason for rejection DB
DESCRIPTION OF SYMBOLS 103 Data preparation part 103-1 Reason for rejection notification analysis part 103-2 Pointer information analysis part 103-3 Data shaping part 103-4 Opinion book analysis part 104 Rejection reason DB search subsystem 104-1 Claim description division part 104- 2 Search unit generation unit 104-3 Search unit management table 104-4 Similarity search unit 104-5 Search result management table 104-6 Result display unit 1601 CPU
1602 Memory 1603 Input device 1604 Output device 1605 External storage device 1606 Portable recording medium drive device 1607 Network connection device 1608 Bus 1609 Portable recording medium

Claims (5)

Search the patent examination database that stores the past applications for which examination requests have been filed and the notification of reasons for refusal issued for the applications subject to examination, linked to the description of the claims in the application documents. In the apparatus for supporting the evaluation and creation work of the application documents,
From the result of morphological analysis of the notice of reasons for refusal taken out from the patent examination database and the rules determined in advance, the application to be examined described in the notice of reasons for refusal and the reasons for refusal in the application The description of the claim in which is indicated, the text that is the basis for the reason for refusal, the cited document indicated by the reason for refusal and the wording described in the cited document, and creates information that links the extracted information, A data creation unit to be stored in the rejection reason database;
A claim description dividing unit that reads the application documents and divides the description of the claims of the application documents into a minimum search unit for performing a similar search;
Based on the minimum search unit created by the claim description division unit, a search unit for searching for a record in the rejection reason database and a search unit generation unit for creating a key description thereof,
Calculates a similarity between the description contents of unexamined patent publication application request for examination in each record of the search key and the rejection reason database corresponding to the search unit the search unit generation unit created has been performed, resulting A similarity search unit for associating with a search unit corresponding to the search key used for calculating the similarity,
And it pointed out the unit estimating unit to estimate a pointed content is in the range to be noted objection to the previous SL application documents,
Only contains the results display unit that displays the indicated range,
The indication unit estimation unit includes:
A search unit having a relatively high similarity to the description content among search units for which an estimation result of the indication range is not determined is selected from the first search unit and the second search unit;
A first similarity that is a similarity with respect to the first search unit, a second similarity that is a similarity with respect to the second search unit, and the first search unit and the second search unit. Obtaining a third similarity that includes both and is related to a third search unit for which an estimation result has not been determined;
If the first similarity is greater than the third similarity, the first search unit is estimated as the indication range;
Patent specification evaluation / creation work characterized in that when the first and second similarities are greater than the third similarity, the first and second search units are estimated in the indicated range. Support device. The data creation unit searches and analyzes the opinion document linked to the notice of reasons for refusal from the patent examination database, describes an objection to the reason for refusal or a description of amendments, or an application in which the reason for refusal is pointed out The final disposition information is further linked to information linked to the notice of reasons for refusal to create the information,
The patent specification evaluation / creation work support apparatus according to claim 1, wherein: The data creation unit obtains a description corresponding to the final disposal information from the patent examination database or a patent related database connected to the outside, and creates the information as the information.
The patent specification evaluation / creation work support apparatus according to claim 2, wherein: Search the patent examination database that stores the past applications for which examination requests have been filed and the notification of reasons for refusal issued for the applications subject to examination, linked to the description of the claims in the application documents. In a method for supporting the evaluation and creation of the application documents,
From the result of morphological analysis of the notice of reasons for refusal taken out from the patent examination database and the rules determined in advance, the application to be examined described in the notice of reasons for refusal and the reasons for refusal in the application The description of the claim in which is indicated, the text that is the basis for the reason for refusal, the cited document indicated by the reason for refusal and the wording described in the cited document, and creates information that links the extracted information, the reasons for refusal database to store,
Read the application documents, and divide the description of the claims of the application documents into search minimum units for performing a similarity search,
Based on the minimum search unit, create a search unit for searching for a record in the rejection reason database and a description as a key thereof,
The similarity between the search key corresponding to the search unit and the description content of the published patent publication of the application for which examination was requested in each record of the rejection reason database was calculated , and the obtained similarity was used for the calculation. Corresponds to the search unit corresponding to the search key ,
Before Symbol similarity is relatively high retrieval units of the description contents among the search unit of estimation result corresponding to point out content is in the range is not determined to be pointed rejection reason for the application documents, the Select 1 search unit and 2nd search unit,
A first similarity that is a similarity with respect to the first search unit, a second similarity that is a similarity with respect to the second search unit, and the first search unit and the second search unit. Obtaining a third similarity that includes both and is related to a third search unit for which an estimation result has not been determined;
If the first similarity is greater than the third similarity, the first search unit is estimated as the indication range;
If the first and second similarities are greater than the third similarity, the first and second search units are estimated as the indication range;
A patent specification evaluation / creation work support method comprising a process of displaying the indication range. Search the patent examination database that stores the past applications for which examination requests have been filed and the notification of reasons for refusal issued for the applications subject to examination, linked to the description of the claims in the application documents. To a computer that supports the evaluation and creation of the application documents,
From the result of morphological analysis of the notice of reasons for refusal taken out from the patent examination database and the rules determined in advance, the application to be examined described in the notice of reasons for refusal and the reasons for refusal in the application The description of the claim in which is indicated, the text that is the basis for the reason for refusal, the cited document indicated by the reason for refusal and the wording described in the cited document, and creates information that links the extracted information, the reasons for refusal database to store,
Read the application documents,
The description of the claim of the application document is divided into minimum search units for performing a similar search,
Based on the minimum search unit, create a search unit for searching for a record in the rejection reason database and a description as a key thereof,
The similarity between the search key corresponding to the search unit and the description content of the published patent publication of the application for which examination was requested in each record of the rejection reason database was calculated , and the obtained similarity was used for the calculation. Corresponds to the search unit corresponding to the search key ,
A search unit having a relatively high similarity to the description content among search units for which an estimation result as to whether it falls within the indicated range , which is a range in which the reason for refusal of the application document is pointed out, is firstly selected. Select a search unit and a second search unit for
A first similarity that is a similarity with respect to the first search unit, a second similarity that is a similarity with respect to the second search unit, and the first search unit and the second search unit. Obtaining a third similarity that includes both and is related to a third search unit for which an estimation result has not been determined;
If the first similarity is greater than the third similarity, the first search unit is estimated as the indication range;
If the first and second similarities are greater than the third similarity, the first and second search units are estimated as the indication range;
A program characterized by causing the processing for displaying the indicated range.