JP5928673B2 - Relay server - Google Patents

Description

  The present invention relates to a relay server for patent search, a patent search method, and a program thereof. In particular, the present invention relates to a relay server, a patent search method, and a program for searching foreign patents.

  Patent Document 1 aims to search for a target document even if a keyword is input in an arbitrary language. When an English keyword is input, the input keyword is translated into Japanese by the translation processing unit. A document search system is disclosed in which a target document is searched from an English document file that is input to a search processing unit and refers to an English index file.

  Further, in Patent Document 2, a keyword of a second language corresponding to a keyword of the first language stored in the replacement table is read using the input keyword of the first language as a key, and the keyword of the second language is used. A patent information retrieval system for creating a retrieval formula is disclosed.

  As a method for calculating the importance of words in a document in order to extract important words in the document, for example, a method using residual IDF (RIDF) shown in paragraphs 0024 to 0037 of Patent Document 3 Is known.

JP-A-5-324719 Special Registration 3666005 JP 2004-145790 A

  However, according to the above prior art, for example, it is not possible to reduce the time and effort required to determine whether to apply for a Japanese application and examination request for a patent filed in a foreign country such as Germany or France. There was a first problem). That is, in such a case, the following work is required. 1) Understand the points of the description and claims written in German, 2) Select Japanese keywords from the points understood, consider synonyms for this keyword, and 3) From the points understood when necessary Confirm the relevant F-term, 4) Decide the final search formula, 5) Connect the Japanese gazette to the searchable patent DB, perform the search, and 6) Check the contents of the Japanese gazette that was hit in the search . Here, there has been a problem that it takes a lot of man-hours to grasp the points in the specification and claims described in the foreign language such as 1) German. There is also a problem that the patent classification such as IPC granted in a foreign application is not necessarily appropriate.

  In such a case, it is conceivable to select Japanese keywords by looking at the translation of the machine translation after machine translation of the text data of the patent filed in the first country. However, the accuracy of machine translation is not very high unless special measures are taken, and if Japanese keywords are selected from the translations of the machine translation, an incorrect keyword may often be selected as a result (second issue) )was there. On the other hand, when the conventional technique described in Patent Document 2 is used to select an accurate keyword, it is necessary to prepare a pair of keywords in a plurality of languages for each technical field in some cases, which takes a lot of man-hours. There was a problem that it would end up.

  On the other hand, when trying to apply for an application filed in the first country in Japan overseas, for example, when trying to apply for a patent in Russia or China, if we try to conduct an application search, we will face the following problems. I could n’t. 1) There is a problem in the accuracy of application searches and infringement searches overseas, especially in non-English speaking countries. For example, when an overseas patent search company is requested to search for a non-English-speaking application, the accuracy of the search may be questionable despite the high cost. 2) I can't verify the search results myself (I had to believe in a patent search company). Especially when engineers in R & D departments conduct patent searches themselves, non-English speaking application searches and infringement searches were extremely difficult.

In order to solve the above problems 1) and 2), even if a search is made using a paid patent DB that can search the full English translated text, the accuracy is not always sufficient due to English translation by machine translation. There is concern about using it to narrow down the number of cases by excluding noise. Furthermore, even if Japanese keywords are translated into the target language, the keywords translated into the target language are entered into the patent database of the target country, and the search results are machine-translated into Japanese, they are similarly narrowed down due to low translation accuracy. (3rd problem) For example, in a so-called statistics-based translation engine, a sufficient amount of corpus analysis improves the translation accuracy of technical terms centered on nouns, but on the other hand, it does not perform morphological analysis, so it focuses on the translation of verbs. The expression of sentences often becomes unnatural. On the other hand, in the rule-based translation engine, the expression of the sentence becomes natural because the morphological analysis is accurate, but on the other hand, it is difficult to increase the accuracy of the technical term because the technical term cannot be selected according to the context. In particular, in the case of patent documents, it is often necessary to select different translations for the same word depending on the technical field. For example, the English line is used to mean "line" in the communication field, "line" in the word processing software field, "line" in the graphic field, "power line" in the high voltage field, and "production line" in the production technology field. Although it is used to mean “production line” and “line (meaning command system)” in the business field, it is not easy to translate these by machine translation.

  The purpose of the present invention is to connect the patent research using a language other than English, in addition to the patent office and intellectual property staff doing the patent research using a language other than English. It is an object of the present invention to provide a relay server for a patent search system that can be easily performed without being aware of the destination.

In order to solve the first problem, the relay server according to claim 1 is a relay server configured to be able to communicate with a client terminal and a patent database server,
Importance calculating means for calculating the importance of each word of the examination target patent data in order to extract the important word from the examination target patent data received from the client terminal;
First patent classification determination means for determining a patent classification of the examination target patent data using an important word extracted based on the importance;
A first translation processing unit that identifies a specialized dictionary using the determined patent classification, and translates the important word into a desired language using the identified specialized dictionary;
First search expression generation means for generating a search expression using an important word translated into the desired language;
First connection request generation means for generating a connection request according to the connection destination for a patent database server which is a connection destination according to the desired language;
First search expression transmitting means for transmitting the search expression to the patent database;
First receiving means for receiving search result information matching a search formula from the patent database .

In order to solve the second problem, the relay server according to claim 2 includes:
Second translation processing means for translating the received search result information into a desired language;
First display information generating means for generating display information for displaying the translation result of the second translation processing means on the client
And further comprising .

The relay server according to 請 Motomeko 3 is a relay server configured to communicate with the client terminal and patent database server,
Second patent classification determination means for determining a patent classification using a keyword received from the client terminal;
A third translation processing unit that identifies a specialized dictionary using the determined patent classification, and translates the received keyword into a desired language using the identified specialized dictionary;
Second search expression generation means for generating a search expression using the keyword translated into the desired language;
A second connection request generating means for generating a connection request according to the connection destination for the patent database server which is a connection destination according to the desired language;
Second search expression transmission means for transmitting the search expression generated by the second search expression generation means to the patent database server;
Second receiving means for receiving search result information which is result information searched by the patent database server using the search expression generated by the transmitted second search expression generating means;
A fourth translation processing means for identifying a specialized dictionary using the determined patent classification, and translating search result information received from the patent database server using the identified specialized dictionary into a desired language;
Transmitting means for transmitting the translated search result information to the client;
It is characterized by providing.

The relay server according to 請 Motomeko 4, the fourth translation processing means, statistical-based to perform translation by performing statistical processing using a corpus analysis result of accumulating the results of corpus analysis the accumulation And a rule-based translation engine that translates the original text divided into words by applying morphological analysis by applying a predetermined translation rule, and the statistical-based translation for the search result information acquired from the patent database server Generating second display information for displaying the translation execution result by the engine and the translation execution result by the rule-based translation engine for the search result information acquired from the patent database server in comparison with the screen of the client; Display information generating means is further provided.

According to the invention described in claim 1 of the present application, the relay server of the present invention is configured to be able to communicate with the client terminal and the external existing (a plurality of) patent database servers. The operator of the client terminal is aware of the application country of the patent data to be examined, the written language, or the patent database of the connection destination by sending the patent data to be examined and data specifying the application country to this relay server. The patent search can be executed without doing so.

According to the second aspect of the present invention, the second translation processing means for translating the received search result information into a desired language, and the translation result of the second translation processing means for displaying on the client Since it further includes a first display information generating means for generating display information, the search result information can be translated into a suitable language according to the accuracy of machine translation between languages, the language ability of the operator, and the like .

According to the invention described in claim 3 of the present application, the relay server of the present invention is configured to be able to communicate with the client terminal and the external existing (a plurality of) patent database servers. The operator of the client terminal transmits information specifying the keyword and the country of search destination to this relay server, so that the patent of the connection destination (and the language of patent data in the database) is not conscious . A search can be performed.

1 is a system configuration diagram showing an entire patent search system including a relay server of the present invention. It is a hardware block diagram which showed the hardware configuration of the relay server 15, each client, and each patent DB server of this invention. It is the functional block diagram which showed the function structure of the relay server of this invention. It is a flowchart which shows the first half part (search process) of the process performed with the client of this invention, a relay server, and patent DB. It is a flowchart which shows the latter half part (except search processing) of the process performed with the client and relay server of this invention. It is a figure which shows the detailed flow of S430 of FIG. 4 of this invention. It is a figure which shows the detailed flow of S433 of FIG. 4 of this invention. It is a figure which shows the detailed flow of the retranslation processing (S528 of FIG. 5) of this invention. It is a figure which shows the detailed flow of a search request production | generation / transmission process (S428 of FIG. 4) of this invention. It is a figure which shows the initial screen of the foreign patent search system of this invention. FIG. 6 is a diagram showing a client display screen (S404 in FIG. 4) for domestic application search using foreign gazette data of the present invention. FIG. 5 is a diagram showing a client display screen (S404 in FIG. 4) for foreign patent search using Japanese keywords of the present invention. 11 shows a client display screen that displays a list of unexecuted batch jobs displayed when the job creation button 149 is pressed on the screen of FIG. 11 or when the job creation button 159 is pressed on the screen of FIG. S404). It is a figure which shows the client display screen at the time of the search result display of this invention (S407 of FIG. 4). FIG. 6 is a diagram showing a client display screen when referring to a job execution history according to the present invention (S530 in FIG. 5). It is a figure which shows an example of the RIDF table of this invention. It is a figure which shows an example of the prior probability table of this invention. It is a figure which shows an example of the conditional probability table of this invention. It is a figure which shows an example of the log | history table of this invention. It is a figure which shows an example of the connection destination information table of this invention. It is a figure which shows an example of the dictionary master of this invention. It is a figure which shows an example (dictionary ID = D1002) of the dictionary table of this invention. FIG. 4 is a functional block diagram showing a functional configuration different from that of FIG. 3 of the relay server of the present invention. The

  Hereinafter, an embodiment of a patent search system centering on a relay server of the present invention will be described with reference to the drawings.

  First, FIG. 1 is a diagram showing a system configuration of a patent search system. The relay server 15, the client A11, the client B12, the client C13, the patent DB server P16, the patent DB server Q17, and the patent DB server R18 are configured to be able to communicate via the network 10. The client A11, the client B12, and the client C13 are not limited to these three units, and may be configured by a larger number of clients. The patent DB server P16, the patent DB server Q17, and the patent DB server R18 are not limited to these three units, and may be configured by a larger number of patent DBs.

The network 10 may be a so-called Internet, a local area network (LAN), or a dedicated line. The protocol is typically TCP / IP, but is not limited thereto. The patent DB server P16, the patent DB server Q17, and the patent DB server R18 each store a patent database of a different country. Even if a plurality of patent DB servers store a patent DB of the same country. A single patent DB server may store a plurality of patent DBs.

  For example, the patent DB server P16 stores a DB for Chinese patents (Chinese) and a DB for international patents (English), and the patent DB server Q17 stores a DB for Russian patents (Russian). The DB server R18 stores a DB of Japanese patents. In addition, other patent DB servers (not shown) store DBs of US patents and EP patents, and the relay server 15 is configured to be able to communicate with these patent DB servers (not shown). The client A11, the client B12, and the client C13 are used by different companies and individuals, but it is of course possible to connect a plurality of clients to the patent search system in the same company.

  The client A11, client B12, and client C13 are typically so-called personal computers, but may be portable terminals as long as they can communicate via the network 10. The relay server 15 relays communication between the client terminals such as the client A11, the client B12, and the client C13, and the patent DB server side such as the patent DB server P16, the patent DB server Q17, and the patent DB server R18. In addition, translation, search expression generation, creation of a search result display layout, and the like are executed. The processing of the relay server 15 will be described in detail with reference to FIG. This is the end of the description of FIG.

  Next, FIG. 2 is a diagram showing a hardware configuration of the relay server 15, the client A11, the client B12, the client C13, the patent DB server P16, the patent DB server Q17, and the patent DB server R18. The CPU 21, ROM 22, RAM 23, video graphic adapter (VGA) 24, LAN adapter 25, keyboard 26, mouse 27, HDD 28, and DVD-ROM drive 29 are respectively connected via the system bus 20.

  An interface unit may exist between these, each component, and the system bus 20, but is omitted in the drawing. The CPU 21 controls the entire relay server. Specifically, at startup, the CPU 21 reads a startup program (boot loader) from the ROM 22 to check and initialize hardware, and then reads a system program from the HDD 28 and loads it into the RAM 23. Start up the system (OS). Further, the CPU 21 loads an application program from the HDD 28 to the RAM 23 and executes it in accordance with a user operation.

  The ROM 22 stores the startup program as described above, and the RAM 23 is used as a so-called main memory as an operating system, application program, and work area. The VGA 24 is an adapter for connecting to a liquid crystal display or a CRT to display a screen, and the LAN adapter 25 is connected to an external network to transmit / receive data. The keyboard 26 and mouse 27 function as input means for inputting user operations. In the HDD 28, various programs (operating system, application programs for executing the flowcharts shown in FIGS. 4 to 12 of the present invention, and HTML screen layout information to be displayed on the client screen), various tables, and search history information are stored. (Search condition information and detailed information of search results) are stored. The DVD-ROM drive 29 is a drive for reading a DVD-ROM or DVD-R storing various programs, and can write various data to the DVD-R. This is the end of the description of FIG.

Next, the functional configuration of the relay server 15 of the present invention and the outline of the processing in the relay server 15 will be described with reference to FIG. FIG. 3 is a functional block diagram of the relay server of the present invention. Receiver Transferring unit 305 stores the appropriate RAM23 in session memory or the HDD28 based on the session ID included operating information and search result information received from the outside thereto, the operation information received from the external operation information It transmits to the determination part 310. The operation information determination unit 310 determines whether the operation information is a search request (or a history reference request. If the operation information is a search request, the operation information including the search condition is sent to the importance calculation unit 330. When the operation information is a history reference request, the history reference condition is sent to the history reference unit 315. The history reference unit 315 uses the received history reference condition to store history information that matches the condition from the history information storage unit 320. The data is transmitted from the reception unit / transmission unit 305 to the client via 310.

  On the other hand, when the operation information is a search request, the importance calculation unit 330 receives this search request, and when a non-Japanese file is specified in the search condition included in the search request, the importance calculation unit 330 stores the file as a file. The importance level of each included word is calculated, and the calculated high-importance word is extracted as an important word and stored in the work area of the RAM 23. The IPC determination unit 335 determines the patent classification by using the important word stored in the work area of the RAM 23 and the IPC read from the KW / IPC storage unit 355 using this important word as a key, and stores it in the work area of the RAM 23. When searching for a Japanese patent (when searching for the patent DB server R18), the patent classification may of course be FI instead of IPC.

Next, the translation processing unit I 340 uses the patent classification stored in the work area to identify which specialized dictionary of the dictionary information storage unit 360 is to be used, and uses the identified specialized dictionary to identify important words included in the file as Japanese. The translated important words are stored in the work area of the RAM 23. Further, the search expression generation unit I 345 generates a search expression using the important words translated into Japanese stored in the work area of the RAM 23 and the patent classification determined by the IPC determination unit 335 and stores the search expression in the work area of the RAM 23. . The connection request generation unit I 350 reads the connection script and connection destination URL to the patent DB (patent DB server R18) of the Japanese patent from the connection destination information storage unit 325, stores it in the work area of the RAM 23, and the reception unit / transmission unit. 305 transmits the connection script stored in the work area of the RAM 23 to the connection destination URL, then transmits the search formula stored in the work area of the RAM 23 to the patent DB server R18, and the search result received from the return patent DB server R18. Information is stored in the work area of the RAM 23.

When the translation process II 365 determines that the search result information stored in the work area of the RAM 23 is in Japanese, the process proceeds to the sort processing unit 370. If the sort information is specified in the operation information, the sort processing unit 370 returns the search result. The information is sorted by the designated sort key and stored in the work area of the RAM 23. Next, the display information generation unit I 375 generates display information (for example, HTML information) for displaying the result information on the client screen from the sorted search result information according to the layout designation included in the operation information, and the work of the RAM 23 Remember in the area. The reception unit / transmission unit 305 transfers this display information to the requesting client.

On the other hand, when a Japanese keyword is specified in the search condition, the importance calculation unit 330 calculates the importance of each keyword, sorts the keywords in descending order of importance, and stores them in the work area of the RAM 23. To do. Next, the IPC determination unit 335 determines the patent classification by using the IPC read from the KW / IPC storage unit 355 by using the keyword having high importance stored in the work area of the RAM 23 as a key, and stores it in the work area of the RAM 23. Next, the translation processing unit I 340 uses the patent classification stored in the work area to identify which specialized dictionary of the dictionary information storage unit 360 is to be used, and the keyword is included in the operation information using the identified specialized dictionary. The translated important word is stored in the work area of the RAM 23. Further, the search expression generation unit I 345 generates a search expression using the keyword translated into the target language stored in the work area of the RAM 23 and the patent classification determined by the IPC determination unit 335 and stores the search expression in the work area of the RAM 23.

The connection request generation unit I 350 reads the connection script and the connection destination URL to the patent DB (patent DB server P16) in the target language from the connection destination information storage unit 325, stores it in the work area of the RAM 23, and receives / transmits the reception unit / transmission unit. 305 transmits the connection script stored in the work area of the RAM 23 to the connection destination URL, then transmits the search formula stored in the work area of the RAM 23 to the patent DB server P16, and the search result received from the return patent DB server P16. Information is stored in the work area of the RAM 23. The translation processing unit II 365 is a dictionary of the target language included in the operation information, and translates the search result information into Japanese using the dictionary corresponding to the patent classification determined by the IPC determination unit. Store in work area. When the sort information is specified in the operation information, the sort processing unit 370 sorts the search result information translated into Japanese with the specified sort key and stores it in the work area of the RAM 23. Next, the display information generation unit I 375 generates display information (for example, HTML information) for displaying the result information on the client screen from the sorted search result information according to the layout designation included in the operation information, and the work of the RAM 23 Remember in the area. The reception unit / transmission unit 305 transfers this display information to the requesting client. This is the end of the description of FIG. Details of the overall processing will be described in detail with reference to the flowcharts and screens in FIG.

  Next, an outline of the patent search processing and the like centering on the relay server 15 of the present invention will be described with reference to the flowcharts of FIGS. In this flow, steps S421 to S433 are processes executed under the control of the CPU 21 of the relay server 15, and steps S401 to S408 are processes executed under the control of the CPU 21 of the client A11. S446 is a process executed under the control of the CPU 21 of the patent DB server P16.

  First, in step S401, the client A11 transmits a connection request to the relay server 15, in step S421, the relay server 15 receives the connection request, transmits a return authentication request screen to the client A11, and the client A11 displays an authentication screen (not shown). When the CPU 21 detects that the user ID and password are input and the OK button is pressed, the CPU 21 transmits the input user ID and password to the relay server 15. The relay server 15 receives the user ID and password and stores them in the work area of the RAM 23, and the CPU 21 advances the process to step S422. In step S422, the CPU 21 of the relay server 15 searches an unillustrated authentication table (a table in which the user ID and password are encrypted and stored) stored in the HDD 28 using the user ID as a key, and the password for the user ID is determined. If they match, the process proceeds to step S423. If they do not match (if the user ID matches but the password does not match, or there is no user ID match), The process proceeds to step S424.

  In step S424, the CPU 21 of the relay server 15 transmits an error message to the effect that the user ID or password does not match to the client A11. In step S403, the client A11 receives the error message, and displays the error message on a dialog screen (not shown) (a screen displaying a message and an “OK” button). When the CPU 21 detects that the “OK” button has been pressed on this dialog screen, the process returns to step S401. On the other hand, in step S423, the CPU 21 of the relay server 15 transmits initial screen information (for example, HTML, the same applies hereinafter) for displaying the initial screen to the client stored in the HDD 28 to the client A11. In step S402, the client A11 receives the initial screen information, displays the initial screen shown in FIG. 13 on the display, and advances the process to step S404. In step S404, the initial screen 130 shown in FIG. 1. Domestic patent search by foreign publication data 2. Foreign patent search using Japanese keywords Search result query, 4. By pressing the selection buttons 131 to 135 for the management information setting / changing menus, request information for requesting the relay server 15 for screen information corresponding to each menu is transmitted.

In step S425, the relay server 15 transmits screen information corresponding to the received request information to the client A11. For example, when the selection button “1” 131 is pressed on the screen of FIG. 10, the screen of the domestic patent search based on the foreign gazette data shown in FIG. 11 is displayed on the screen of the client A11. When the selection button “2” 132 is pressed on the screen of FIG. 10, the foreign patent search screen using Japanese keywords shown in FIG. 12 is displayed on the screen of the client A11. When the selection button “4” 134 is pressed on the screen of FIG. 10, the management information setting / change screen is displayed on the screen of the client A11, and the selection of the default search engine and the display of the bibliographic information of the search results are displayed. It is possible to select items, add new or change connection destination information tables (FIG. 20), maintain the dictionary master (FIG. 21), etc. These translation options, display item information and management information are stored in the relay server 15. Stored in the HDD 28. When a predetermined time (for example, 0.2 seconds) elapses in step S404, the process proceeds to step S405.

  In step S405, the CPU 21 of the client A11 determines whether there is a processing request. Specifically, for example, in the screen of FIG. 11, at least one of the check boxes for countries such as US and WO, for example, the check box 141 of CN is checked, and the text is pasted in the original data paste field 143. Or a file reference button 144 is pressed to select a specific file on an explorer-like screen (not shown), and the search button 148 or the button 147 of, for example, “100” is selected with the maximum number of radio buttons. This is a case where the CPU 21 detects that the job creation button 149 has been pressed. The same applies to the screen of FIG. 12, and the difference from the screen of FIG. 11 is that text is pasted in the original data paste field 143 or a file reference button 144 is pressed and a specific file is displayed on a screen similar to Explorer (not shown). In FIG. 12, a search expression (an expression combining a keyword and “+” meaning logical sum or “*” meaning logical product and parentheses) is input to the search expression input field 153 in FIG. This is a point that at least one of the fields (for example, “automobile”) needs to be specified in the IPC field 154 or the field specification field 155 (an error occurs if both are not specified). The sort option can be designated by pressing the designation button 142 or the sort designation button 152 on the screen of Fig. 12. Selection of these buttons and input to each input column are possible. Power is stored in the work area of the RAM23 as the operation information.

  When the search button 148 is pressed in FIG. 11 or the search button 158 is pressed in FIG. 12 in the above-described state, the process proceeds to step S406, and the operation information or search request stored in the work area of the RAM 23 is used. Process request information including a request category, a user ID, and a process request time indicating that the request is present is transmitted to the relay server 15.

  In the above state, when the job creation button 149 is pressed in FIG. 11 or the job creation button 159 is pressed in FIG. 12, the CPU 21 advances the process to step S406. Indicates that the operation information and job storage request stored in the work area of the RAM 23 (request to store the search specified by the operation information as one job information in the work area of the RAM 23 of the relay server 15). Processing request information including the request category is transmitted to the relay server 15, and information on the batch search job that has been created and has not been executed is received from the relay server 15, and the screen of FIG. 13 is displayed. When the batch job execution button 162 is pressed on the screen of FIG. 13, the CPU 21 transmits processing request information including a batch job execution request to the relay server 15.

  The relay server 15 receives the processing request information in step S426, stores it in the work area of the RAM 23, and advances the processing to step S427. In step S427, the processing request information stored in the work area of the RAM 23 is referred to and it is determined whether the processing request is a search request. If it is a search request, the process proceeds to step S428. If it is not a search request, the process proceeds to step S521 in FIG.

  In step S428, referring to the processing request information stored in the work area of the RAM 23, the patent DB server to be connected (patent DB stored in the HDD 28 of the patent DB server P16) is specified and specified. A connection request to P16 is generated, the generated connection request is transmitted to the requested patent DB server P16, the authentication screen information received from the return patent DB server P16 is stored in the work area of the RAM 23, and the corresponding authentication information is stored. Is transmitted to the connection request destination patent DB server P16, and the process proceeds to step S429. Step S428 will be described in more detail later with reference to the flowchart of FIG.

  In step S429, in response to the screen information received from the patent DB server P16, the screen information is requested until a transition is made to the search condition input screen, the next screen information is received from the return patent DB server P16, and the search condition input screen is displayed. If transition is made, the process proceeds to step S430. In step S430, a search request including a search formula is generated based on the search condition included in the processing request information stored in the work area of the RAM 23, and is transmitted to the patent DB server P16 as search request information. At this time, if the search request is a search in the foreign patent DB, a search expression is generated by translating a Japanese search keyword into a search destination language. Step S430 will be described in detail later with reference to the flowchart of FIG.

In step S4 4 3, Patent DB server P16 stores the received RAM23 work area a search request information, search and perform a search based on the search expression contained in the request information, patents and utility models that match the search expression Are stored in the work area of the RAM 23, and the process proceeds to step S444. In step S444, the list information stored in the work area of the RAM 23 (information that can display one item including the publication number, the applicant, the name of the invention, the filing date, etc. in one line) is transmitted to the relay server 15. In step S431, the relay server 15 receives the list information, stores it in the work area of the RAM 23, and advances the process to step S432. In step S432, details indicating that the patent DB server P16 is requested to receive detailed information (other bibliographic information, issues, claims 1 and the like) corresponding to the maximum number included in the operation information stored in the work area of the RAM 23. Send an information request.

  In step S445, the patent DB server P16 receives the detailed information request, stores it in the work area of the RAM 23, and advances the process to step S446. In step S446, the number of detailed information corresponding to the maximum number is read from the patent DB and transmitted to the relay server 15.

  In step S433, if the received detailed information includes a foreign language text, it is translated into Japanese, a display layout is generated, and transmitted to the client A11 as display information. Further, the history table (FIG. 19) and the detail file (FIG. 19) of the history information storage unit 320 of the HDD 28 are used as the history request information including the detailed information as the received search result information and the operation information as the search condition information. In the history table result / condition information). Step S433 will be described later in detail with reference to the flowchart of FIG. When a predetermined period (for example, two months) elapses, the detailed information is deleted from the HDD 28, and the search condition information is continuously stored in the HDD 28.

  In step S407, the client A11 receives the display information, stores it in the work area of the RAM 23, and displays the search result on the screen based on this display information. An example of the search result display screen is shown in FIG. In the screen of FIG. 14, since the number of cases matching the condition (645) exceeds the maximum number of 100 cases, a message indicating that the first 100 cases are displayed is displayed. Can be sorted and displayed by a desired item that can be set as a sort key, such as the newest issue date or the oldest issue date. If a predetermined time (for example, 0.2 seconds) has elapsed, the process proceeds to step S408. In step S408, if there is a screen transition request (for example, pressing of the transition request button 177 to the main screen on the screen of FIG. 14), the process returns to step S402. Instead of returning to S402, it is of course possible to display the “Previous Screen” button on the screen of FIG. 14 and return the processing to Step S404 in response to pressing of the “Previous Screen” button. This is the end of the description of FIG.

  Next, an outline of processing on the relay server 15 side when there is a processing request other than a search request will be described using the flowchart of FIG. In the flowchart of FIG. 5, steps S501 to S504 are processes executed under the control of the CPU 21 of the client A11, and steps S521 to S532 are processes executed under the control of the CPU 21 of the relay server 15.

  In step S521, the relay server 15 refers to the operation information included in the processing request information stored in the work area of the RAM 23, and whether there is a reference request (the button “3” for requesting a search result query on the screen of FIG. 10). 133 is pressed and the search history reference button 176 is pressed on the screen of FIG. 14), if there is a reference request, the process proceeds to step S522, and if there is no reference request, the process proceeds to step S531. Proceed to In step S522 when there is a reference request, the search job execution history list shown in FIG. 15 is displayed with reference to the result / condition information of the history table shown in FIG. 19 stored in the HDD 28, and the process goes to step S523. Proceed.

  In step S523, if the search condition remains but there is no detailed information, the process proceeds to step S428 in FIG. 4 in order to perform a search based on this search condition. If there is detailed information, the process proceeds to step S524. In step S524, with reference to the operation information stored in the work area of the RAM 23 while the screen of FIG. 15 is displayed, whether the sort has been designated, that is, the result / condition information file link to the file shown in FIG. The sort designation button 182 is pressed before 181 is selected, and it is determined whether a sort item and ascending or descending order are designated on a screen (not shown). If there is a sort designation, the process proceeds to step S525. If there is no sort designation and the result / condition information link is selected, the process proceeds to step S526. Further, when a link to a file with a specific result / search information is selected after the retranslation button 184 is pressed on the screen of FIG. 15, an unillustrated for specifying the retranslation condition of the selected file. A dialog will appear, allowing you to select a different translation engine from the original search (for example, only rule-based translation, but re-translation, both rule-based and statistical-based translation engines), and English data ( The English translation information of the patent or the English patent information of the corresponding patent) can be obtained from other patent DBs and displayed using the publication number as a key.

  In step S525, the designated sort items are sorted in the designated ascending or descending order and stored in the work area of the RAM 23, and the process proceeds to step S526. In step S526, the display information (for example, HTML) to be displayed on the client screen is generated using the detailed information stored in the work area of the RAM 23 or read from the HDD 28, and the client A11 is generated. Send to.

  In step S501, the client A11 receives the display information, stores it in the work area of the RAM 23, displays a screen similar to FIG. 14 based on this display information, and advances the process to step S502. In step S502, if the main screen button 177 in FIG. 14 is pressed, the process returns to step S402 in FIG. 4, and if the re-search button 178 is pressed, the process returns to step S404 in FIG. Not shown in the flowchart).

  On the other hand, if it is determined in step S521 that there is no reference request, the process proceeds to step S531. In step S531, with reference to the operation information stored in the work area of the RAM 23, it is determined whether a job creation request has been made in step S404 of FIG. That is, it is determined whether the job creation button 149 is pressed on the screen of FIG. 11 or the job creation button 159 is pressed on the screen of FIG. If it is determined that there is a job creation request, the process proceeds to step S532. If it is determined that there is no job creation request, the process proceeds to step S527. In the job storage step S532, the condition information of the newly created job is added to the history table (FIG. 19) of the history information storage unit 320 stored in the HDD 28 of the relay server 15, and the process proceeds to step S527. At the time of addition, the search condition is stored in the result / condition information file, and the result information is not stored. When the job is one of a plurality of (for example, five) batch jobs, batch number information such as 1-5 (the first of five cases) is also stored as a part of the search condition.

  In step S527, with reference to the operation information stored in the work area of the RAM 23, it is determined whether the retranslation button 184 is pressed on the screen of FIG. 15 in step S404 of FIG. If it is determined that the retranslation button 184 has been pressed, the process proceeds to step S528. If it is determined that the retranslation button 184 has not been pressed, the process proceeds to step S530. In step S530, for example, the management information setting / change button “4” 134 is pressed on the screen of FIG. 10 to change the URL of the connection destination information shown in FIG. This means processing for maintaining the dictionary table shown in FIG. 22, a case where the display switching request button 179 is pressed on the search result screen of FIG.

  In step S528, the operation information stored in the work area of the RAM 23 is referred to perform retranslation with the specified option, the retranslation result is stored in the work area of the RAM 23, and the process proceeds to step S529. The option here means, for example, W translation (execution of both translation by a rule-based translation engine and translation by a statistics-based translation engine) and acquisition of English data (acquisition of English translation data or acquisition of corresponding English gazette data). In the case of W translation, a plurality of different rule-based translation engines may be used, or a plurality of different statistical-based translation engines may be used. It is also possible to translate using different specialized dictionary sets using the same translation engine. Step S528 will be described in detail later with reference to the flowchart of FIG.

  In step S529, display information for laying out the retranslation result stored in the work area of the RAM 23 and displaying it on the client A11 is generated and transmitted to the client A11. In step S503, in step S504, the client A11 receives the display information, stores it in the work area of the RAM 23, displays a screen similar to FIG. 14 based on this display information, and advances the process to step S504. In step S504, if the main screen button 177 in FIG. 14 is pressed, the process returns to step S402 in FIG. 4, and if the re-search button 178 is pressed, the process returns to step S404 in FIG. Not shown in the flowchart). This is the end of the description of FIG.

  Next, details of step S430 (search request generation / transmission processing) in FIG. 4 will be described with reference to the flowchart in FIG. Steps S621 to S630 and steps S641 to S649 in FIG. 6 are processes executed under the control of the CPU 21 of the relay server 15.

  First, in step S621, referring to the operation information stored in the work area of the RAM 23, whether the requested search is a domestic patent search using foreign gazette data in FIG. 11 or a foreign patent search using Japanese keywords in FIG. Is determined, the determination result is stored in the work area of the RAM 23, and the process proceeds to step S622. In step S622, if it is determined that the foreign patent search is determined in step S621, the process proceeds to step S623. If it is determined that the foreign patent search is not performed, the process proceeds to step S641.

  First, a case where it is determined that the search is not a foreign patent search (and therefore a Japanese patent search is performed based on foreign gazette data) will be described (steps S641 to S649). In the present embodiment, as shown in FIG. 11, the IPC is estimated based on Chinese gazette data, and a patent DB of Japanese patents is searched. The following processing is performed as preprocessing before this processing is executed. First, a corpus is created by collecting several n patent documents sufficient to cover all fields of IPC. Actually, n = about 60,000, and n is a sufficient number to cover all fields of IPC. All the documents in the corpus are analyzed. First, the prior probability that the first patent classification (first IPC) of each of the n documents is a specific patent classification is calculated, and a prior probability table shown in FIG. 17 is created. Here, the category ID has a one-to-one correspondence with the patent classification (IPC), and the total number is n n n n n. In addition, it is desirable that n n n n n be several hundred to several thousand by aggregating IPC with an appropriate number of digits.

Furthermore, m words are extracted from all the documents of the corpus. For example, m is assumed to be close to 1 million words at the maximum, and therefore has a 6-digit number, which corresponds to the word IDs in FIGS. 16 and 18 being 6 digits. Further, in the relay server of the present invention, correspondence to about 100 languages at the maximum is assumed, the number of dictionaries per language is assumed to be about 100 dictionaries, and the dictionary ID of the dictionary master of FIG. It has a total of 4 digits, 2 digits for the dictionary number. For example, the language code for Chinese is 05, 01 for English, and 11 for Russian. Furthermore, the word ID in the dictionary table of FIG. 22 is a total of 8 digits including 2 digits of the language code + 6 digits of the word code of FIGS. Next, the probability of appearance of each of m words in each IPC is calculated, and a conditional probability table for each word for each IPC shown in FIG. 18 is created. Further, RIDF = log (n / ni) + log (1-exp (−Fi / n)) for each of m words included in the patent corpus is calculated and stored in the importance degree table of FIG. RIDF is a so-called residual IDF, and RIDF applies to general words whose Poisson distribution is a number or a general word, but on the other hand, it does not apply to content words with special contents. The difference between the document frequency estimated by applying the Poisson distribution to the frequency of occurrence and the measured value is used as the importance of the word. In the above formula, n is the number of documents included in the corpus, ni is the frequency of documents including the word Wi, and Fi is the frequency of the word Wi in the patent corpus. 16, 17, and 18 are stored in the HDD 28 of the relay server 15.

Next, the processing from step S641 to step S649 will be described. First, foreign gazette data (a file designated by pressing the original data input field 143 or the file reference button 144 in FIG. 11) stored in the work area of the RAM 23 is stored in another work area of the RAM 23. Then, the process proceeds to step S642. In step S642, using the appearance frequency of each word included in the document d that is the publication data to be processed and the appearance frequency of each word in the corpus, the importance i of the word Wi, which is the following expression, is defined as follows: Calculate with
Importance i = fi x RIDF
= Fi x (log (n / ni) + log (1-exp (-Fi / n)))
Here, fi is the frequency of the word Wi in the document d (publication data to be processed), n is the number of documents included in the corpus, ni is the document frequency including the word Wi, and Fi is the frequency of the word Wi in the patent corpus. . FIG. 16 shows a RIDF table for each of m words included in the patent corpus. Thereby, all the importance levels i of the words included in the document d, which is the publication data to be processed, can be calculated. The importance i of each word is stored in the work area of the RAM 23, and the process proceeds to step S643.

  Next, in step S643, the appearance frequency of each word included in the document d (the publication data to be processed) is calculated and stored in the work area of the RAM 23, and the process proceeds to step S644. In step S644, the appearance frequency of each word included in the publication data to be processed calculated in step S643 and the conditional probability and category ID of each corresponding word in the conditional probability table shown in FIG. 18 are used as a key in FIG. With reference to the prior probability table shown, the probability (a posteriori probability) that the document d belongs to the category ID is calculated for all categories, the calculated probability for each category ID is stored in the work area of the RAM 23, and the process proceeds to step S645. . In step S645, the work area of the RAM 23 is referred to, the category ID giving the highest probability is specified, and the patent classification (IPC) corresponding to this category ID is estimated as the patent classification of the publication data to be processed. The estimated patent classification is stored in the work area of the RAM 23, and the process proceeds to step S646.

The above calculation is expressed as follows.
If the conditional probability that the document d belongs to a certain class c (category ID) is P (c | d), then Bayes' theorem
Call it. Here, tk is the importance of the word wk. The class Cmap that gives the maximum posterior probability is

Is a document class. Where P ^ (c) is the prior probability of class c estimated from the corpus, P ^ (tk | c) is the conditional probability of word wk estimated from the corpus,

Is calculated. T _ck is the sum of the importance of the word wk in the document to which the class c belongs.

  In step S646, the dictionary shown in FIG. 21 using the predetermined number of IPC digits stored in the work area of the RAM 23 (three digits in the example of the dictionary master in FIG. 21 and G05 in the first three digits of IPC) and language (CN) as keys. With reference to the master, a dictionary ID to be used for translation is specified, the dictionary ID is stored in the work area of the RAM 23, and the process proceeds to step S647. In this embodiment, dictionary ID = D0502. In step S647, the translation is executed with reference to the dictionary table specified by the dictionary ID. In this embodiment, since CN is selected on the screen of FIG. 11, several words (usually about 5 words) with high importance included in the Chinese gazette data are selected and translated into Japanese. The word keyword is stored in the work area of the RAM 23, and the process proceeds to step S648.

  In step S648, the synonym of the Japanese keyword stored in the work area of the RAM 23 is read from a synonym dictionary (not shown) stored in the HDD 28 of the relay server 15, and stored in the work area of the RAM 23 together with the original keyword. A search expression is generated using the keywords and synonyms and the estimated IPC stored in the work area of the RAM 23, stored in the work area of the RAM 23, and the process proceeds to step S649. In step S649, the connection destination URL and connection parameters to the patent DB of the Japanese patent that is the connection destination (in this embodiment, the JPO electronic library) and the connection destination ID (in this embodiment, PD0001) as a key are shown in FIG. Is read out from the connection destination information table and the connection parameter file (stored in the connection destination information storage unit 325 of the HDD 28) and transmitted to the connection destination patent DB (patent DB server R18), and the return search condition input screen information is displayed. After receiving from the patent DB server R18, the search formula stored in the work area of the RAM 23 is transmitted and the process is terminated. This is the end of the description of the processing in the case of searching Japanese patents using foreign gazette data in FIG.

  Next, the case of conducting a foreign patent search will be described with reference to FIG. 6 (steps S623 to S631). In step S622, if it is determined in step S621 that a foreign patent search is made, the process proceeds to step S623. In step S623, the operation information stored in the work area of the RAM 23 is read out and stored in the variable area in order to determine whether IPC is specified in the IPC input field 154 on the screen of FIG. The process proceeds to S624. In step S624, it is determined whether or not IPC is specified based on the read operation information. If it is determined that IPC is not specified, the process proceeds to step S625. If it is determined that IPC is designated, the process proceeds to step S627.

In step S625, it is determined by reading the operation information stored in the work area of the RAM 23 whether the field has been selected in the field designation field 155 of FIG. If it is determined that there is a field designation, the process proceeds to step S626. In step S626, as shown in FIG. 6 , the IPC as the patent classification is determined using the field designation, and the process proceeds to step S627. Other hand, when it is determined that there is no field specifying the process proceeds to step S631. In step S631, an error message “Please enter IPC or specify one of the fields.” Is sent to client A11, the process returns to step S425 in FIG. 4 and the process ends (step S631). In S425, the screen information of FIG. 12 is transmitted to the client A11 again).

On the other hand, in step S627, the language (CN) is identified from the country name by referring to the connection destination table in FIG. 20 using the country name as a key based on the designation of the target country in FIG. 12 (CN in FIG. 12). And the field designated in the field designation field 155 on the screen in FIG. 12 or the designation of the IPC determined in step S624 as a key, the dictionary master in FIG. 21 is referred to, the dictionary ID is specified, and this dictionary ID ( In the embodiment, D0502) is stored in the work area of the RAM 23, and the process proceeds to step S628. In step S628, the Japanese keyword input in the search expression field 153 is translated into the target language (Chinese in the embodiment) using the dictionary of the dictionary ID stored in the work area of the RAM 23, and the work area of the RAM 23 is used. And the process proceeds to step S629.

  In step S629, the synonym of the keyword translated into the target language is read from a synonym dictionary (not shown) stored in the HDD 28 of the relay server 15, and stored in the work area of the RAM 23 together with the original keyword. A search expression is generated using the word and the specified IPC stored in the work area of the RAM 23, stored in the work area of the RAM 23, and the process proceeds to step S630. In step S630, the connection destination URL and connection parameters to the patent DB of the Chinese patent that is the connection destination (in this embodiment, SIPO), and the connection destination shown in FIG. 20 using the connection destination ID (in this embodiment, PD0003) as a key. A search that is read from the information table and the connection parameter file, transmitted to the patent DB (patent DB server P16) of the connection destination, and the search condition input screen information is received from the patent DB server R16 and then stored in the work area of the RAM 23 The expression is transmitted and the process is terminated. This is the end of the description of the processing when searching for foreign patents using Japanese keywords, and the description of FIG. 6 ends.

  Next, details of step S433 (display information generation / transmission processing) in FIG. 4 will be described with reference to the flowchart in FIG. Steps S720 to S738 in FIG. 7 are processes executed under the control of the CPU 21 of the relay server 15. In step S720, the detailed information of the patent that matches the search condition received from the patent DB server is stored in the HDD 28, and the process proceeds to step S721. In step S721, a search option is read from the operation information stored in the work area of the RAM 23, and the process proceeds to step S722. In step S722, if the retrieved search option is a foreign patent search (a search condition is input and a search is executed on the screen of FIG. 12), the process proceeds to step S723. In the case of Japanese patent search using gazette data), the process proceeds to step S727. In step S727, the detailed information of the Japanese patent that matches the search conditions stored in the HDD 28 is read and stored in the work area of the RAM 23, and the process proceeds to step S728.

  In step S723, the dictionary master of FIG. 21 is referred to using the designated IPC or field and target country of the operation information stored in the work area of the RAM 23 as a key, and the corresponding dictionary ID is stored in the work area of the RAM 23. Then, the dictionary used for translation is specified, and the process proceeds to step S724. In step S724, with reference to the operation information stored in the work area of the RAM 23, it is determined whether the operation information includes both rule-based translation and statistical-based translation, and both translation specifications are included. If YES in step S725, the process advances to step S725; otherwise, the process advances to step S726. When the default translation engine designation is a statistics-based translation engine, the execution order of steps S725 and S726 is reversed.

  In step S725, the detailed information of the foreign patents that match the search conditions received from the foreign patent DB server P and stored in the HDD 28 is translated into Japanese using a statistically-based known translation engine (for example, Moses), and the translation result. Is stored in the work area of the RAM 23, and the process proceeds to step S726. In step S726, detailed information on foreign patents that match the search conditions received from the foreign patent DB server P is transferred to a known rule-based translation engine (for example, a conventional machine translation engine disclosed in Japanese Patent Laid-Open No. 2000-330988 (text). The number of words included in the character string that represents the grammar rules that match the part-of-speech arrangement pattern for the part-of-speech sequence attached to a series of words from the beginning to the end of the sentence. A machine translation engine using a morphological analysis method that makes the optimal solution), a cross-language translation engine), or a known statistical-based translation engine (for example, disclosed in paragraph 0009 of JP-A-2007-141133). A bilingual corpus with a large-scale preparation of bilingual sentences in the target language and target language This is a method for obtaining a translation rule for translation and its probability value, and translating the translation result into Japanese using a translation engine that obtains a translation result to which the translation rule having the highest probability value is applied. The information is stored in the area, and the process proceeds to step S728.

  In step S728, referring to the operation information stored in the work area of the RAM 23 in step S724, the operation information is sorted (the sort designation button 142 on the screen in FIG. 11 or the sort designation button 152 on the screen in FIG. 12 is pressed). It is determined whether the sort item and the ascending / descending order are designated). If the sort is designated, the process proceeds to step S729. If the sort is not designated, the process proceeds to step S730. In step S729, the detailed information sorting process is executed in accordance with the designation, and the process proceeds to step S730.

  In step S730, a new file (result / condition information file) specified by the result / condition information file in the history table is created for each job, stored in the history information storage unit 320 of the HDD 28, and operation information is stored. In addition to the job ID, start date, time, and user ID included in the processing request information included, the operation information such as keyword, IPC, sort designation presence / absence, and the job is one of a plurality of (for example, five) batch jobs. In such cases, the number of batches such as 1-5 (meaning the first of 5 cases), the detailed information received from the patent DB server, and its translation data, including the result / condition information file (for example, tagged with XML) The various types of information are distinguished), and the process proceeds to step S731. In step S731, when the process is normally executed, the “status” in the history table shown in FIG. 19 is updated from 0 (initial value) indicating that the process is being executed to 1 that indicates normal end. If there is no search result, or if the process ends abnormally, such as when the connection to the patent DB server fails, the process is updated to 2 meaning abnormal end, and the process proceeds to step S732.

  In step S732, the operation information stored in the work area of the RAM 23 is referenced to determine whether the operation information includes information instructing execution of the batch job. If the batch job is instructed, the process is performed. The process advances to step S733. If a batch job is not instructed, the process advances to step S734. In step S733, the batch number information in the history table is referred to. If the batch job is not completed, the process returns to step S430 in FIG. When the batch job is completed (for example, when the job of batch number information 5-5 is completed), the process proceeds to step S734.

  In step S734, the process proceeds to step S735 with reference to the operation information stored in the work area of the RAM 23. In step S735, a specification related to the result display (for example, whether W translation is specified) is read, and display information for displaying the search result on the screen of the client A11 according to this specification is generated including the layout. The data is stored in the work area of the RAM 23, and the process proceeds to step S736. In step S736, the display information stored in the work area of the RAM 23 is transmitted to the client A11, and the process proceeds to step S737. In step S737, it is determined whether the current job is a batch job by referring to the batch number information in the history table. If the current job is a batch job, the process proceeds to step S738. If the current job is not a batch job, the process ends. In step S738, the user master stored in the HDD 28 or other unillustrated personnel server is inquired using the user ID as a key to read out an e-mail address corresponding to the user ID, and a series of addresses is sent to the e-mail address of the user ID. An e-mail notifying the completion of the job is transmitted and the process is terminated. This is the end of the description of FIG.

  Next, details of the retranslation processing in step S528 of FIG. 5 will be described with reference to the flowchart of FIG. Steps S1121 to S1130 in FIG. 8 are processes executed under the control of the CPU 21 of the relay server 15, and are executed when the retranslate button 184 is pressed on the screen in FIG. Steps S1141 to S1146 are processes executed under the control of the CPU 21 of the patent DB server.

  In step S1121, with reference to the operation information stored in the work area of the RAM 23, it is determined whether a W translation (both translation by the rule-based translation engine and translation by the statistics-based translation engine) is requested by a re-translation request. If the W translation is requested, the process proceeds to step S1122. If the W translation is not requested, the process proceeds to step S1124. In step S1122, the result / condition information in the history information table shown in FIG. 19 is referred to, the translation designation at the previous execution is referred to, it is determined whether the translation is insufficient, and the process proceeds to step S1123. In step S1123, if there is a shortage, additional translation is executed using the detailed information received from the patent DB server of the result / search information and the IPC information, and the process proceeds to step S1124.

  In step S1124, in step S1121, referring to the operation information stored in the work area of the RAM 23, English data (the corresponding WIPO or USPTO English gazette of the gazette or the English translation data of the gazette itself) is requested as a retranslation request. If English data is requested, the process proceeds to step S1125. If English data is not requested, the process proceeds to step S1127. In step S1125, the relay server 15 refers to the connection destination information table shown in FIG. 20, and transmits a connection request to the patent DB server storing English data. In step S1141, the patent DB server storing English data receives the connection request, and the process proceeds to step S1142. In step S1142, the authentication screen information is transmitted to the relay server 15. In step S1126, the relay server 15 stores the received authentication screen information in the work area of the RAM 23, and the process proceeds to step S1127.

  In step S1127, the relay server 15 refers to the connection parameter file in the connection destination information table shown in FIG. 20 according to the received authentication screen information, and transmits the authentication information to the patent DB server storing the English data. . In step S1143, patent DB server authentication information storing English data is received, authentication processing is performed to determine whether the user ID and password match, and the process proceeds to step S1144. In step S1144, screen information for DB search is transmitted to the relay server 15. In step S1128, the relay server 15 receives the screen information, refers to the connection parameter file of the connection destination information table shown in FIG. 20 according to the received screen information, and sets the publication number requesting the connection parameter and the English data. The inquiry information including is generated, and the process proceeds to step S1129. In step S1129, this inquiry information is transmitted to a patent DB server storing English data.

In step S1145, the patent DB server storing the English data receives the inquiry information, reads the corresponding English patent data using the publication number as a key, stores it in the work area of the RAM 23, and advances the process to step S1146. In step S1146, the English patent data is transmitted to the relay server 15. In step S1130, the relay server stores the English patent data in the RAM23 of the work area, and the process ends. This is the end of the description of FIG.

  In step S1221, the patent DB server to be connected is identified by referring to the operation information stored in the work area of the RAM 23, and then the session memory in the RAM 23 (to be established by connecting to the patent DB server) The session with the patent DB server is maintained with reference to the memory area allocated to the RAM 23 for communication with the patent DB server and storage of various parameters and secured during the session maintenance). If it is already connected and the session is maintained, the process returns to step S430 (patent search request generation / transmission) in FIG. If not connected or the session is not maintained, the process proceeds to step S1222.

  In step S1222, the URL and connection parameters are stored in the work area of the RAM 23 with reference to the connection destination information table in FIG. 20, and the process proceeds to step S1223. In step S1223, a connection request is generated based on the URL and connection parameters stored in the work area of the RAM 23, stored in the work area of the RAM 23, and the process proceeds to step S1224. In step S1224, a connection request is transmitted to the patent DB server to be connected, and the process proceeds to step S1225.

In step S1225, authentication screen information is received from the patent DB server to be connected and stored in the work area of the RAM 23, and the process proceeds to step S1226. In step S1226, the authentication information is transmitted to the patent DB server to be connected by referring to the connection destination information table according to the received authentication screen information. This is the end of the description of FIG. The present invention described in the above embodiment includes the case where the control program described in the flowcharts of FIGS. 4 to 9 for controlling the relay server 15 is provided by a storage medium or a communication medium via a communication line. A control method in which each control step executed in these flowcharts is executed by the relay server is also included in the present invention. Finally, FIG. 23 will be described. Figure 23 is a functional block diagram of FIG. 3 corresponding to claims 1 and 2 (0027 paragraph ～0033 paragraph) to base are those described respective constituent features of claim 3. The same functional blocks as those in FIG. 3 are denoted by the same reference numerals, and description of the same functional blocks is omitted. Keyword received from the client terminal end via the operation information determining unit 310 from the receiving unit and transmitting unit 2305, it is determined keywords Patent Classification in patent classification decision II2335, the determined patent classification in the translation processing unit III234 0 In response, the received keyword is translated into a desired language. Against further desired search expression in the search expression generating unit II2345 using keywords translated into the language is generated and patent database server at a connection request generator II2350 a connection destination corresponding to the desired language, connection request according to the connection destination is generated and transmitted to the patent database server is connected ahead from the receiving unit and transmitting unit 2305 specifies a specialized dictionary with the determined patent classification, the identified specialized using a dictionary to translate into the desired language search data obtained from the patent database server at the translation processing unit IV, translated search result information generated by the display information generation unit II2375 display information order to client representation , it executes a process of transmitting to the receiving unit and transmitting unit 2305 client from.

  The relay server of the present invention is suitable as a relay apparatus that provides a multilingual patent search service by connecting to a plurality of existing patent databases and a plurality of client terminals.

10 Network 15 Relay Server 11 Client A
12 Client B
13 Client C
15 Relay server 16 Patent DB server P
17 Patent DB Server Q
18 Patent DB Server R
20 System bus 21 CPU
22 ROM
23 RAM
28 HDD
305 reception unit / transmission unit 310 operation information determination unit 315 history reference unit 320 history information storage unit 325 connection destination information storage unit 330 importance calculation unit 335 IPC determination unit 340 translation processing unit I
345 Search Expression Generation Unit I
350 Connection request generator I
355 KW / IPC storage unit 360 Dictionary information storage unit 365 Translation processing unit II
370 Sort Processing Unit 375 Display Information Generation Unit I
2305 Receiver / Transmitter
2335 Patent Classification Judgment Unit II
2340 Translation Processing Unit III
2345 Search Expression Generation Unit II
2350 Connection request generator I
2365 Translation Processing Unit IV
2375 Display Information Generation Unit II

Claims (7)

A relay server configured to be able to communicate with a client terminal and a patent database server,
Importance calculating means for calculating the importance of each word of the examination target patent data in order to extract the important word from the examination target patent data received from the client terminal;
First patent classification determination means for determining a patent classification of the examination target patent data using an important word extracted based on the importance;
A first translation processing unit that identifies a specialized dictionary using the determined patent classification, and translates the important word into a desired language using the identified specialized dictionary;
First search expression generation means for generating a search expression using an important word translated into the desired language;
First connection request generation means for generating a connection request according to the connection destination for a patent database server which is a connection destination according to the desired language;
First search expression transmitting means for transmitting the search expression to the patent database;
A relay server, comprising: first receiving means for receiving search result information that matches a search formula from the patent database. Second translation processing means for translating the received search result information into a desired language;
2. The relay server according to claim 1, further comprising first display information generation means for generating display information for displaying a translation result of the second translation processing means on the client. A relay server configured to be able to communicate with a client terminal and a patent database server,
Second patent classification determination means for determining a patent classification using a keyword received from the client terminal;
A third translation processing unit that identifies a specialized dictionary using the determined patent classification , and translates the received keyword into a desired language using the identified specialized dictionary ;
Second search expression generation means for generating a search expression using the keyword translated into the desired language;
A second connection request generating means for generating a connection request according to the connection destination for the patent database server which is a connection destination according to the desired language ;
Second search expression transmission means for transmitting the search expression generated by the second search expression generation means to the patent database server;
Second receiving means for receiving search result information which is result information searched by the patent database server using the search expression generated by the transmitted second search expression generating means;
A fourth translation processing means for identifying a specialized dictionary using the determined patent classification, and translating search result information received from the patent database server using the identified specialized dictionary into a desired language;
Transmitting means for transmitting the translated search result information to the client;
A relay server comprising:   The fourth translation processing unit applies a statistical translation engine that accumulates the results of corpus analysis and performs translation by performing statistical processing using the accumulated corpus analysis results, and a predetermined translation rule And a rule-based translation engine that translates the original text divided into words by performing morphological analysis, and a translation execution result by the statistical-based translation engine for the search result information acquired from the patent database server and the patent database server A second display information generating unit configured to generate display information for displaying the translation execution result by the rule-based translation engine for the acquired search result information in comparison with the screen of the client; The relay server according to claim 3. The second display information generating means, further comprising: English acquisition means for acquiring, from the patent database, English data corresponding to the patent-targeted translation data of the statistics-based translation engine or the patent-targeted translation data of the rule-based translation engine Is a display for displaying at least one of the translation execution result of the statistics-based translation engine, the translation execution result of the rule-based translation engine, and the acquired English data on the screen of the client. The relay server according to claim 4 , wherein information is generated. A relay server control program configured to be communicable with a client terminal and a patent database server,
The relay server,
Importance calculating means for calculating the importance of each word of the examination target patent data in order to extract the important word from the examination target patent data received from the client terminal;
First patent classification determination means for determining a patent classification of the examination target patent data using an important word extracted based on the importance;
A first translation processing unit that identifies a specialized dictionary using the determined patent classification, and translates the important word into a desired language using the identified specialized dictionary;
First search expression generation means for generating a search expression using an important word translated into the desired language;
First connection request generation means for generating a connection request according to the connection destination for a patent database server which is a connection destination according to the desired language;
First search expression transmitting means for transmitting the search expression to the patent database;
A control program for functioning as first receiving means for receiving search result information that matches a search formula from the patent database. A relay server control program configured to be communicable with a client terminal and a patent database server,
The relay server,
Second patent classification determination means for determining a patent classification using a keyword received from the client terminal;
A third translation processing unit that identifies a specialized dictionary using the determined patent classification , and translates the received keyword into a desired language using the identified specialized dictionary ;
Second search expression generation means for generating a search expression using the keyword translated into the desired language;
A second connection request generating means for generating a connection request according to the connection destination for the patent database server which is a connection destination according to the desired language;
Second search expression transmission means for transmitting the search expression generated by the second search expression generation means to the patent database server;
Second receiving means for receiving search result information which is result information searched by the patent database server using the search expression generated by the transmitted second search expression generating means;
A fourth translation processing means for identifying a specialized dictionary using the determined patent classification, and translating search result information received from the patent database server using the identified specialized dictionary into a desired language;
A control program for causing the translated search result information to function as transmission means for transmitting to the client.