JP2003141164A - Patent document retrieval method, device and program, and storage medium for storing patent document retrieving program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To automatically retrieve other patent documents by using a F-term given to the patent document, only by inputting a document number of the specific patent document without studying the patent classification and a keyword in advance, to judge the similarity from many technological viewpoints. SOLUTION: This method of retrieving the similar patent documents by inputting the patent document number by using the patent information database, comprises processes of (a) reading out the patent document data of the inputted patent document number from the patent information database, (b) extracting the multi-viewpoint patent classification symbol from the read-out patent document data, (c) retrieving the corresponding patent documents from the patent information database, while applying the extracted multi-viewpoint patent classification symbol as the successive retrieval items, (d) counting the number of multi-viewpoint patent document classification symbol to which the document is fitted every retrieval patent document, and (e) displaying the retrieved patent document with the fitted multi-viewpoint patent classification symbol and the counted value.

Description

Detailed Description of the Invention

[0001]

The present invention relates to a patent document search method, a patent document search device, and a patent document which efficiently search for patent documents having contents similar to a specific patent document and display the similarity from a technical point of view. The present invention relates to a storage medium storing a search program and a patent document search program.

[0002]

2. Description of the Related Art In searching for patent information, a patent document having contents similar to a specific patent document may be searched for various purposes. For example, to determine the patentability of a patent application,
When searching for a material to be invalidated or creating a patent map, it is necessary to search for patent documents having similar contents. In such a case, a similar patent document is generally searched using the patent classification assigned to the document.

However, when the International Patent Classification (IPC) and FI are used as the patent classification, a large number of documents are searched, which is noisy and inefficient. Also F
When trying to use a term, it is difficult to understand the F term, and it is difficult to find out the name (title) of the term. In particular, when searching for similar patent documents with patent documents having dozens of F-terms as specific patent documents, it is very labor-intensive to examine the F-term titles and consider their combinations, which is a reality. Not at all.

On the other hand, it has been performed to search for similar patent documents in consideration of some keywords expressing specific patent documents. For example, an attempt has been made to calculate the degree of similarity by the ratio of the number of keywords in a specific patent document to the number of keywords in the searched patent document (Japanese Patent Laid-Open No. 5-13).
5109). However, in some patent documents, the same technical content is expressed by different words, so that the keywords often leak. Further, patent documents are often extracted from different technical fields with a high degree of similarity, which makes it difficult to select keywords.

Further, a concept search for creating and searching a search sentence that briefly expresses the contents of a specific patent document, a similarity search for searching by using a description such as a summary of a specific patent document as a search sentence, etc. are announced. However, in the present situation, not only the search results are noisy, but also specific patent documents themselves are not searched in many cases.

[0006] On the other hand, the F-term is classified into one in which the FI is further re-developed and one in which it is re-divided. However, it is characterized by a system in which the FI in a specific technical range is finely divided from a plurality of different points. A large number of F-terms are given by experts by grasping the technical contents of patent documents from various viewpoints.

However, when a general user tries to use the F-term, the F-term system is difficult to understand, and it is difficult to find the classification corresponding to the intended technical point of view and check the name (title). Since it is a complicated work, it is currently not used. In particular, when searching for similar patent documents with patent documents having dozens of F-terms as specific patent documents, it is very labor-intensive to examine the F-term titles and consider their combinations, which is a reality. Not at all.

[0008]

However, since the expert understands the technical contents of patent documents and gives a large number of F terms from various technical viewpoints, the keywords in patent documents are mutually exclusive. A common F rather than a measure of similarity
It is expected that the reliability of the similarity will be higher when the term is used as a measure of the similarity between patent documents. Conventionally,
There is no patent document search method for automatically displaying the degree of similarity between a particular patent document and another patent document having similar contents with a common number of F terms. Furthermore, there is no one that displays a common F-term as data for judging the content of similarity.

In order to determine the patentability of a particular patent document or to search for a material for invalidation, it is known to other patent documents having similar contents to the particular patent document.・ There was no automatic display of future application. Also, although there was a cited document database that displays other patent documents having a citation relationship with a specific patent document,
There has been no automatic indication of a citation relationship with a particular patent document and other patent documents having similar contents.

[0010]

SUMMARY OF THE INVENTION The present invention has been made in view of such circumstances, and is given to a patent document by simply inputting the document number of a particular patent document without considering the patent classification or the keyword in advance. By automatically searching for other patent documents using the F-term and displaying the F-terms that hit the searched patent documents as well as the number of F-terms that have been hit, the similarity is calculated. It is a first object to provide a patent document search method that can be judged from many technical points of view.

Further, according to the present invention, by simply inputting the reference number of a specific patent document without considering the patent classification or the keyword in advance, other patent documents can be automatically identified by using the F-term assigned to the patent document. A patent document search method is performed in which the degree of similarity can be determined from many technical points of view by displaying the F terms that hit the searched patent documents and the F terms that have hit the searched patent documents. A second object of the present invention is to provide a patent document search device that is directly used for the purpose.

Further, according to the present invention, by simply inputting the reference number of a specific patent document without considering the patent classification or the keyword in advance, other patent documents can be automatically identified by using the F-term assigned to the patent document. A computer-based patent document search method capable of determining the degree of similarity from many technical points of view by displaying the F-terms that hit the searched patent documents and the F-terms that hit the searched patent documents A third object is to provide a computer program to be used directly by the computer.

Further, according to the present invention, by simply inputting the reference number of a particular patent document without considering the patent classification or the keyword in advance, other patent documents can be automatically identified by using the F-term assigned to the patent document. A computer-based patent document search method capable of determining the degree of similarity from many technical points of view by displaying the F-terms that hit the searched patent documents and the F-terms that hit the searched patent documents A fourth object is to provide a storage medium that stores a computer program that is directly used to implement the above.

[0014]

According to the present invention, a first object is a method of searching for a similar patent document by inputting a patent document number using a patent information database, and (a) the input patent document The patent document data of the number is read from the patent information database; (b) The multi-view patent classification symbols are extracted from the read patent document data; (c) The patent information with the extracted multi-view patent classification symbols as search items in order Retrieving the relevant patent document from the database; (d) Counting the number of multi-view patent classification symbols that hit that document for each searched patent document; (e) the multi-view patent classification symbol that has been hit and The searched patent document is displayed with the count value added thereto. A method of searching a patent document characterized by comprising the above steps.

According to the present invention, a second object is an apparatus for retrieving similar patent documents by inputting a patent document number using a patent information database, and (a) patent information including patent document data. Database; (b) Input means for inputting the patent document number and commanding various operations; (c) Search means for searching the patent information database of the input patent document number from the patent information database; (d) Input (E) A memory for storing the search item; (f) A memory for storing the multi-view patent classification symbol of the input patent document of the patent document number; (g) A memory Retrieval means for retrieving relevant patent documents from the patent information database using the retrieved multi-view patent classification symbols as retrieval items in sequence; (h) Memory for storing patent document data of retrieved patent documents; (i) Retrieval Counting means for counting the number of multi-viewpoint patent classification symbols that hit the patent document for each patent document; (j) the searched multi-view patent classification symbol and the searched patent document with the count value (K) control means for controlling the operation of each of the above means; and a patent document search device characterized by the above.

According to the present invention, the third object is achieved by a patent document search program for causing a computer to execute the method according to any one of claims 1 to 4.

According to the present invention, the fourth object is achieved by a storage medium storing a patent document search program for causing a computer to execute the method according to any one of claims 1 to 4.

The patent information database used here stores various bibliographical item data of patent documents. Bibliographic items include various numbers and dates, various patent classifications such as multi-point patent classifications, titles of inventions, applicants, inventors, etc., but also summary, claims, drawings and full text of specification. Or if some of them are accumulated, you can use them conveniently. Further, when displaying a classification title such as a multi-view patent classification, it is necessary to accumulate a patent classification table. Furthermore, when displaying the citation relationship between patent documents, it is necessary that the cited document data is accumulated.

For the construction of the patent information database, various organized and standardized SGML data provided by the Patent Office can be used. "Organized and standardized SG, which is regularly provided
The above-mentioned various bibliographic items can be accumulated from "ML data".
Various bibliographic items of documents that are not included in the "sorted / standardized SGML data" can be accumulated from the "sorted / standardized SGML data (back)".

The data of cited documents and F-terms for the documents that are not included in the "organized / standardized SGML data" are "cited document master organized / standardized SGML data (back)" and "F term search master organized / standardized". SGML data (for back) ”can be stored. The organized and standardized SGML data includes progress information of examinations, trials, and registrations, so if you accumulate this information, you can read the progress information of the input documents or retrieved documents as necessary. It can be used for output such as display and printing.

The patent document number means an application number, a publication number,
There is no limitation as long as it is a publication number, a patent number, etc. and is a valid number in the database of patent documents. The F-term is a typical multi-view patent classification.

The publicly known / predecessor application relationship generally refers to the following four relationships by comparing the filing date of each patent document and the publication date (the publication date of the publication when there is no publication). There is. The searched patent document is known to the input patent document. It is a prior application for the patent document in which the searched patent document is input. It is a later application for the patent document in which the searched patent document is input. The input patent document is known to the searched patent document (here,
It is said that the searched patent document is reversely known to the input patent document). In addition, there are cases where the same application date is applied.

The citation relationship is a relationship whether the input patent document is cited by the retrieved patent document or the retrieved patent document is cited by the input patent document. Here, the citation refers to the one cited by the examiner at the time of examination, but if the patent document number described as a prior art example in the patent application specification exists as data, it is also included as one of the cited documents. preferable. In that case,
The two types of citations are displayed separately. Such information can be extracted from the organized and standardized SGML data.

[0024]

DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below with reference to FIGS. FIG. 1 is a block diagram showing a configuration example of an apparatus which is an embodiment of the present invention, FIG. 2 is an operation flow chart thereof, and FIG.
Is a display example of similar patent documents retrieved by this embodiment.

This device, as shown in FIG.
0, input means 12, display means 14, and patent information database 16 are provided. The CPU 10 is connected to a patent information database 16 that stores bibliographic information and the like. The patent information database 16 is stored in a recording medium such as a hard disk or a CD-ROM. The output unit 18 can output the search result and the like.

Here, the CPU 10 is composed of a personal computer (also called a personal computer, PC) or the like for convenience,
The input / output device may be connected by a communication circuit network which is a network using LAN or the Internet.
The patent information database 16 is a CP via the communication network.
It may be connected to U10. Needless to say, the method of the present invention can be implemented even if the user side is provided with a CPU and an input / output device and directly connected to the patent information database 16.

The patent document search method of the present invention will be described with reference to FIG. The program of the patent document search method of the present invention is in an external storage means built in or connected to the CPU 10, and the CPU 10 performs the following various operations by the control means 10A according to the program. It is assumed that this program has been launched at the start.

First, the operator designates a search item using the input means 12 (step S100 in FIG. 2). The designation of search items means F-term, FI, IP as the patent classification.
Which category of C is selected as a search item, whether to determine the known / previous application relationship between the searched document and the input document patent, and the citation relationship between the searched document and the input document Whether or not to make a determination is designated as a search item.

These search items are stored in the search item memory 10C by the control means 10A. If the number of search items is fixed, the search items may be automatically set in the search item memory 10C by the program, or may be included in the program, so this designation is omitted.

Next, a document number, for example, a public number, is input (step S102). The public number searched in another database may be transferred as a method of inputting the public number. By the instruction to complete the input of the document number, the control means 10A
Using the input document number, the patent information database 16 is searched for the patent document of the corresponding number through the searching means 10B, and the data of the document is held in the input document memory 10D (step S104). Furthermore, if there is citation-related data, it is also retained. Step 100 and Step 1
02 may be in the reverse order.

The control means 10A is an input document memory 10D.
The specified patent classification (only the F term in this embodiment) is extracted from the data stored in the search classification memory 10
It is held at E (step S106).

The control means 10A has a search classification memory 10E.
Search means 10B using the F-term symbol stored in
The patent information database 16 is searched through (step S108). Therefore, if there is a patent document to which the F-term symbol is added, the patent document data is loaded into the search document memory 10F together with the F-term symbol (step S11).
0). Further, if there is citation-related data, it is retrieved and stored in the retrieved document memory 10F. Patent information database 1 with F-term symbols stored in the sequential search classification memory 10E
No. 6 is searched, and the F-term symbol is added to the patent documents that are redundantly searched with different F-term symbols. The searched patent document is processed so as not to include the input document.

When the search of the patent document is completed for all the F-term symbols held in the search classification memory 10E, the counting means 10G makes the F-term symbol (input) for each searched patent document. F-term symbols common to the patent document and the searched patent document are counted (step S112), and the count value is added to each patent document stored in the searched document memory 10F.

In the following, if the search item stored in the search item memory 10C has a public relation / predecessor / poster application relation other than the F term, a route a, and a public / predecessor / poster application relation, route b, a quotation. If there is a relationship, go to route c, and if neither, go to route d.

First, the route a will be described. Step S1
The process proceeds from step 12 to step S114 (route a), and the known / previous / late application relationship of each patent document held in the search document memory 10F is determined. That is, publicly known / earlier application determination means 1
The filing date and publication date of the patent document searched at 0H are extracted from the retrieval document memory 10F, the filing date and publication date of the input patent document are extracted from the input document memory 10D, and they are compared and publicly known Judge the relationship.

Specifically, if the publication date of the searched patent document is before the filing date of the input patent document, the searched patent document is "known" to the input patent document.
To determine. Further, if the filing date of the searched patent document is before the filing date of the entered patent document and the publication date of the searched patent document is after the filing date of the entered patent document, “first application” To determine. In addition, if the filing date of the searched patent document is not after the filing date of the input patent document and after the publication date of the input patent document, it is determined as “later application”.

Further, if the filing date of the searched patent document is earlier than the published date of the input patent document, it is judged as "reverse public knowledge". Furthermore, if the patent documents retrieved and the input patent documents have the same filing date, it is determined to be “application on the same day”. If there is a priority claim date as the application date, it will be adopted.

The determination result such as "publicly known" relating to the known / earlier application obtained in this way is added to each patent document stored in the searched document memory 10F. Next, the process proceeds from step S114 to step S116 (route a). Here, the citation relationship determination means 10J cites the cited document data stored in the input document memory 10D and the citation relation data of each patent document stored in the patent document memory 10F and the citation of the patent document input and the retrieved patent document. Determine the relationship.

If the patent document input to the retrieved patent document is citation data, the retrieved patent document is referred to as "cited". On the contrary, if the input patent document has the searched patent document as the citation data, the searched patent document is “cited” by the input patent document.

“Citation” of the quoting relationship thus obtained
The determination result such as "is added to each patent document stored in the search document memory 10F. Next, the process proceeds from step S116 to step S118 (route a). Number of F terms that hit the patent documents searched here (common number of F terms)
Arrange in descending order.

Next, the document number, the common F-term number, the common F-term symbol, the known / previous / subsequent application relationship, and the citation relationship are displayed in the descending order of the common F-term for each searched patent document.
(Step 120). A display example of the displayed search result will be described later with reference to FIG.

A description will be given of the route b in the case where the search items are only known / earlier application relationships other than the F term. Since the operation of each step is the same below, only the route will be described. The process proceeds from step S112 to step S114 (route b), the publicly known / earlier application relationship is determined, and the determination result is added to each patent document. Further, from step S114 to step S
Proceed to 118 (route b) and proceed to step 120.

The route c in the case where the search items are only quotation relationships other than the F term will be described. The process proceeds from step S112 to step S116 (route c), the quotation relationship is determined, and the determination result is added to each patent document. Further, the process proceeds from step S116 to step S118 (route c), and further proceeds to step 120.

The route d in the case where there is no search item other than the F term will be described. The process proceeds from step S112 to step S118 (route d), and further proceeds to step 120.

Next, FIG. 3 shows a display example of the search results displayed in step 120 in this way. This is a display example in the case of route a. The documents searched here are displayed as similar documents in descending order of the common F-term. All the reference numbers are tentative and the "invention title" is also omitted.

Document number and common F for each searched patent document
The number of terms, the common F-term symbol, the known / previous application relationship and the citation relationship are displayed. Thus, by simply inputting the number of a specific patent document, a patent document having a high degree of similarity to the input patent document is displayed with the number of F term symbols indicating the degree of similarity as a guide.

Here, both the input document and the similar document are underlined in the applicant, the F term, the filing date / publication date, and the abstract, and when clicked, the data of the document is read from the input document memory 10D and opened in another window. Is displayed in. Therefore, if you want to get an overview of the literature, you can click the summary and it will be displayed immediately. In the case of the F term, all the F term symbols given to the document are displayed.

When the common F-term symbol of the patent document retrieved on the display screen is clicked (step S122),
The F term clicked is the control means 10A and the search means 1
The patent classification table in the patent information database 18 is searched via 0B (step S124), and the corresponding F-term and the titles of the upper classes are read and displayed in another window (step S126). Therefore, the viewpoint of the similarity between the input document and the similar document can be immediately confirmed.

For example, the retrieved similar document 1 is shown to have eight common F terms (that is, the F terms that hit the retrieved patent documents) with the input document, and the symbols are listed. When these common F terms are clicked, their titles are displayed in another window at the lower right of FIG. Here, an example in which the F-term code 5F041CA04 is clicked is shown. 5F041CA04
Has three dots and is therefore displayed with the F term of the upper hierarchy.

Further, by clicking the remaining seven F-term symbols, it is possible to confirm the similarity between the retrieved similar document 1 and the input document from the technical point of view. Similarly, for the documents after the similar document 2, the similarity from the technical point of view can be confirmed by clicking the F term symbol. If FI or IPC is also added to the search items, the operations in and after step S106 are performed for them, and the common F
The I number and the FI symbol, and the common IPC number and the IPC symbol may be added and displayed.

Further, since publicly known / earlier applications are shown, for example, if the input document is an unexamined published patent, the patentability of the input document can be judged from "publicly known" or "earlier application" similar documents. It becomes a document. Also, since the citation relationship is also shown,
It serves as a reference as a similar document or a document closely related to it. Also, if you click each applicant, F term, filing date and publication date, the corresponding data will be displayed so that you can refer to them.

Further, the data of the documents held in the input document memory or the retrieved document memory can be separately displayed in a table or the like, and the required data can be selected to be efficiently used for creating the patent map. Note that the publication number, registration number, or the like may be added to the document number of the similar document, or they may be displayed if the progress information is accumulated in the patent information database.

[0053]

As described above, according to the invention of claim 1, the patent searched with the common F-term number can be obtained only by inputting a specific patent document number without considering the patent classification or the keyword in advance. Since the documents are displayed, patent documents with high similarity can be obtained. At this time, it is convenient to judge the similarity by displaying the items in descending order of the common F-term.

If the title of the common F-term can be displayed, it is convenient for technically examining the content of the similarity (Claim 2). In addition, if it is possible to display whether the retrieved patent document is known to the specific patent document, whether it is a prior application, a later application, or a reverse disclosure, it is possible to select the judgment material such as patentability on the spot. It is convenient (Claim 3). If the retrieved patent document is shown to have a citation relationship with the specific patent document, it can be considered as important information as a measure of similarity (claim 4). Further, in this case, various data of the retrieved documents can be used, which is convenient because it can be used as a material for patent map creation.

According to the inventions of claims 5 to 8, there is obtained a patent document search device which is directly used for carrying out the method according to any one of claims 1 to 4.

According to the invention of claim 9, claim 1
A storage medium storing a computer program for carrying out any one of the methods 4 is obtained.

According to the invention of claim 10, claim 1
A computer program for causing a computer to execute any one of the methods 1 to 4 is obtained.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration example of an apparatus according to an embodiment of the present invention.

FIG. 2 is an operation flowchart of the embodiment.

FIG. 3 is a diagram showing an example of a display screen of search results according to the present invention.

[Explanation of symbols]

10 CPU 12 Input means 14 Display means 16 Patent Information Database 18 Output means 10A control means 10B Search method 10C search item memory 10D input document memory 10E Search classification memory 10F Search literature memory 10G counting means 10H Known / earlier application determination means 10J Cited document determination means

Claims (10)

[Claims] 1. A method for searching for similar patent documents by inputting a patent document number using a patent information database, comprising: (a) reading the patent document data of the input patent document number from the patent information database. (B) Extract multi-view patent classification symbols from the read patent document data; (c) Search corresponding patent documents from the patent information database with the extracted multi-view patent classification symbols as sequential search items; d) Count the number of multi-view patent classification symbols that hit the document for each searched patent document; (e) Add the hit multi-view patent classification symbol and the count value Display patent documents; Patent document search method characterized by comprising the above steps. 2. The patent information database includes a patent classification table, and when the displayed multi-view patent classification symbol is designated,
The patent document search method according to claim 1, wherein the corresponding multi-view patent classification is read from the patent classification table and the title is displayed. 3. A known / previous application relationship between the input patent document and the retrieved patent document is determined from the filing date and publication date, and the determination result is attached to the retrieved patent document. The patent document search method according to claim 1, which is displayed. 4. The patent information database includes cited document data, determines a citation relationship between the input patent document and the retrieved patent document, and attaches the determination result to the retrieved patent document. The patent document search method according to claim 1, which is displayed. 5. An apparatus for searching for similar patent documents by inputting a patent document number using a patent information database, comprising: (a) a patent information database containing patent document data; (b) a patent document number. Input means for inputting and instructing various operations; (c) Search means for searching the patent information database of the input patent document number from the patent information database; (d) Patent document data of the input patent document number Memory for storing; (e) Memory for storing search items; (f) Memory for storing multi-view patent classification symbols of the patent documents having the input patent document numbers; (g) Sequential stored multi-view patent classification symbols Retrieval means for retrieving the relevant patent document from the patent information database as a retrieval item; (h) Memory for storing patent document data of the retrieved patent document; (i) Retrieving the document for each retrieved patent document. (J) Display means for displaying the retrieved patent document with the hit multipoint patent classification symbol and the count value; k) A patent document search device comprising: a control unit that controls the operation of each unit. 6. An apparatus for retrieving similar patent documents by inputting a patent document number using a patent information database, comprising: (a) the patent information database containing a patent classification table; (b) the displayed information. And a display means for displaying the title of the searched multi-view patent classification, when the multi-view patent classification symbol is designated. The patent document search device according to claim 5. 7. An apparatus for retrieving similar patent documents by entering a patent document number using a patent information database, comprising: (a) application of the input patent document and the retrieved patent document. 6. The patent document search device according to claim 5, further comprising: a publicly known / earlier / latest application relationship determining means for determining a publicly / earlier / latest application relationship between the date and the publication date; (b) a display means for displaying the determination result. . 8. An apparatus for retrieving similar patent documents by inputting a patent document number using a patent information database, comprising: (a) the patent information database containing cited document data; (b) the input. 6. The patent document search device according to claim 5, further comprising: a citation relationship determination unit that determines a citation relationship between the patent document and the searched patent document; (c) a display unit that displays the determination result. 9. A patent document search program for causing a computer to execute the method according to any one of claims 1 to 4. 10. A storage medium storing a patent document search program for causing a computer to execute the method according to any one of claims 1 to 4.