Information processing system, information processing method, and program

Abstract

To provide a system of easily obtaining a an appropriate retrieval result.SOLUTION: A retrieval query when the document is retrieved is stored in association with the document as complementary information. For a document which does not have the retrieval query, components of the document are stored in association as the complementary information. By using the retrieval query received from a user, the document stored in storage means and the complementary information stored in association with the document are retrieved as the retrieval object.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to an information processing system, an information processing method, and a program. [Background technology]

[0002] In companies, various electronic data is being created and accumulated, and there is a demand for effective use of the accumulated electronic data. However, the amount of accumulated data is huge, and a major issue is that it takes a lot of time to find useful data or data you want.

[0003] In recent years, full-text search systems, such as enterprise search, have been attracting much attention as a means of solving the problem of efficiently finding data.

[0004] In full-text search, scoring is generally used to determine the ranking of search results. Various scoring methods based on tf-idf and BM25 have been proposed, but the appropriate scoring method varies depending on the user's vocabulary and search intent / perspective.

[0005] It is not realistic for users to select the scoring method that suits them best when searching, and it is also difficult to select the most balanced scoring method in advance for all users of the search system.

[0006] Ranking learning defines click logs (browsing history) or annotations for search queries (explicit evaluations of search results by searchers) as relevant documents for a search query and uses them as learning data.

[0007] Then, by considering the score obtained by the scoring method for a search query as a single feature (ranking feature) and treating the combination of scores from multiple scoring methods as a characteristic, the system learns to rank relevant documents in the training data, making it possible to present relevant documents in user click logs and annotations at the top of search results.

[0008] Patent Document 1 describes a technology that extracts previously used search queries or linguistic features extracted from search queries from training data, stores them in a new field for registered documents (hereinafter referred to as the "search query field"), and uses them as ranking features, thereby improving the accuracy of presenting documents that have been clicked (or annotated) in response to past search queries at the top of the results. [Prior art documents] [Patent documents]

[0009] [Patent Document 1] JP 2020-71678 A DISCLOSURE OF THEINVENTION [Problem to be solved by the invention]

[0010] However, it is rare that learning data (click logs or annotations) exists for all documents registered as search targets.

[0011] In Patent Document 1, there are registered documents that have no value in the search query field, and the score for the search query field is 0. Conversely, if there is a value in the search query field, a score can be obtained if there is even a slight match in linguistic features, and even documents with little relevance can have a higher score than documents with no value, and the possibility of them ranking highly increases.

[0012] Because click logs and annotations are selected from search results, documents that appear higher in the results tend to accumulate more learning data. As a result, documents that appear at the top of the results during the early stages of learning tend to continue to appear at the top of the results afterwards, which leads to biased learning data and the inability to respond to changes in vocabulary and search intent.

[0013] Therefore, in order to obtain appropriate search results, it is desirable to reduce the influence that the presence or absence of learning data for a registered document has on the ranking of the search results.

[0014] Therefore, an object of the present invention is to provide a mechanism for obtaining appropriate search results. [Means for solving the problem]

[0015] The information processing system of the present invention is characterized by comprising: a storage means for storing a search query used when searching for a document in association with the document as complementary information, and for a document for which the search query does not exist, for storing components of the document in association with the complementary information; and a search means for using a search query received from a user to search the documents stored in the storage means and the complementary information stored in association with the document as search targets. Effect of the Invention

[0016] According to the present invention, it is possible to obtain appropriate search results. [Brief description of the drawings]

[0017] [Figure 1] 1 is a diagram illustrating an example of a system configuration of a full-text search device according to an embodiment of the present invention. [Diagram 2] 2 is a block diagram showing an example of a hardware configuration of a full-text search device and a client terminal according to an embodiment of the present invention. FIG. [Diagram 3] 1 is a diagram illustrating an example of a functional configuration of a full-text search device according to an embodiment of the present invention. [Figure 4]11 is a flowchart showing an example of a learning process in the full-text search device according to the embodiment of the present invention. [Diagram 5] FIG. 2 is a diagram showing an example of a configuration of a learning data storage area in the embodiment of the present invention. [Figure 6] FIG. 2 is a diagram showing an example of a configuration of a search target storage area in the embodiment of the present invention. [Figure 7] 11 is a flowchart illustrating an example of a process for adding a search query field in the embodiment of the present invention. [Figure 8] 1 is a flowchart illustrating an example of a learning data generation process in an embodiment of the present invention. [Figure 9] 11 is a flowchart illustrating an example of a feature vector generation process for a search query in an embodiment of the present invention. [Figure 10] FIG. 4 is a diagram showing an example of a browsing history in the embodiment of the present invention. [Figure 11] FIG. 4 is a diagram showing an example of a document storage table in the embodiment of the present invention. [Figure 12] FIG. 2 is a diagram showing an example of a browsing history in which search queries are aggregated according to an embodiment of the present invention. [Figure 13] FIG. 13 is a diagram showing an example of a document storage table after a search query field is completed in the embodiment of the present invention. [Figure 14] FIG. 10 is a diagram showing an example of a feature vector for a search query in an embodiment of the present invention. [Figure 15] FIG. 11 is a diagram illustrating an example of learning data for ranking learning in the embodiment of the present invention. [Figure 16] FIG. 2 is a diagram showing an example of a learning model acquired as a result of ranking learning in an embodiment of the present invention. [Figure 17] FIG. 2 is a diagram showing an outline of a feature vector generation process in the embodiment of the present invention. [Figure 18] 11 is a flowchart illustrating an example of a search process in the embodiment of the present invention. [Figure 19]FIG. 10 is a diagram showing an example of a feature vector for a search query in an embodiment of the present invention. [Figure 20] FIG. 11 is a diagram showing an example of a search result in the embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0018] Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

[0019] FIG. 1 is a diagram showing an example of the configuration of an information processing system according to the present invention.

[0020] FIG. 1 is a diagram showing an example of a system configuration of a full-text search device according to an embodiment of the present invention.

[0021] The full-text search device 100 is configured such that a client terminal 110 is connected via a LAN 120 .

[0022] The full-text search device 100 has the functionality of a typical full-text search system, and provides the functionality of returning a list of documents that match a search request as search results for pre-registered documents, and of checking the contents of any document from the search results.

[0023] Furthermore, the full-text search device 100 acquires the browsing history or evaluation of the search results by the searcher, and stores them in association with the search query.

[0024] The client terminal 110 provides the functions of the full-text search device 100 to the user via a browser.

[0025] FIG. 2 is a block diagram showing an example of a hardware configuration of the information processing device of the present invention.

[0026] As shown in FIG. 2, the information processing device has a CPU (Central Processing Unit) 201, a ROM (Read Only Memory) 202, a RAM (Random Access Memory) 203, a storage device 204, an input controller 205, an audio controller 206, a video controller 207, a memory controller 208, and a communication I / F controller 209 connected via a system bus 200.

[0027] The CPU 201 comprehensively controls each device and controller connected to the system bus 200 .

[0028] ROM 202 or external memory 213 holds the BIOS (Basic Input / Output System) and OS (Operating System), which are control programs executed by CPU 201, computer-readable and executable programs for realizing this information processing method, and various necessary data (including data tables).

[0029] The RAM 203 functions as a main memory, a work area, etc., for the CPU 201. The CPU 201 loads programs and the like required for executing processes from the ROM 202 or the external memory 213 into the RAM 203, and executes the loaded programs to realize various operations.

[0030] The input controller 205 controls input from an input device such as a keyboard 210 or a pointing device such as a mouse (not shown). If the input device is a touch panel, the user can give various instructions by pressing (touching with a finger or the like) an icon, cursor, or button displayed on the touch panel.

[0031] The touch panel may be a touch panel capable of detecting positions touched by multiple fingers, such as a multi-touch screen.

[0032] The video controller 207 controls display on an external output device such as a display 212. The display also includes the display of a notebook computer integrated with the main body. Note that the external output device is not limited to a display, and may be, for example, a projector. In addition, for the device capable of receiving the above-mentioned touch operation, an input device is also provided.

[0033] The video controller 207 is capable of controlling a video memory (VRAM) for display control, and can use part of the RAM 203 as the video memory area, or can provide a separate dedicated video memory.

[0034] The memory controller 208 controls access to the external memory 213. As the external memory, an external storage device (hard disk) for storing a boot program, various applications, font data, user files, edit files, and various data, a flexible disk (FD), or a compact flash (registered trademark) memory connected to a PCMCIA card slot via an adapter, etc., can be used.

[0035] The communication I / F controller 209 connects and communicates with external devices via a network, and executes communication control processing on the network. For example, communication using TCP / IP, telephone lines such as ISDN, and communication using 4G and 5G lines for mobile phones are possible.

[0036] The CPU 201 enables display on the display 212 by, for example, executing a process of expanding (rasterizing) an outline font in a display information area in the RAM 203. The CPU 201 also enables a user to give instructions using a mouse cursor (not shown) or the like on the display 212.

[0037] Next, the function of the information processing system in this embodiment will be described with reference to FIG.

[0038] The client terminal 110 provides the functions of the full-text search device 100 to the user via a Web browser 310. The Web browser 310 has a function of displaying a screen for accepting input of a search query from the user and a screen for displaying the results of a search performed by the full-text search device 100.

[0039] The learning processing unit 301 of the full-text search device 100 has a function of generating a learning model using data stored in the learning data storage area 302 and the search target storage area 303 .

[0040] The learning data storage area 302 stores a browsing history storage table that records browsing history that associates documents viewed by a user with the search queries used at the time of said viewing, as well as a learning model that is the learning result of the learning processing unit 301.

[0041] The search target storage area 303 stores a document storage table in which indexes of documents to be searched are registered.

[0042] The search processing unit 304 has a function of generating feature vectors for search queries of the browsing history stored in the learning data storage area 302, searching a document storage table stored in the search target storage area 303 based on a search query received from a user via the Web browser 310, and returning the search results to the Web browser 310.

[0043] (Learning process) Next, the learning process executed by the full-text search device according to the embodiment of the present invention will be described with reference to the flowchart of FIG.

[0044] FIG. 4 is a flow chart showing a process of learning a past browsing history, which is executed at predetermined intervals by the CPU 201 of the full-text search device 100 by reading and executing a predetermined control program.

[0045] In step S401, the learning processing unit 301 acquires a past browsing history from the browsing history storage table 501 in the learning data storage area 302 shown in FIG.

[0046] In step S402, the learning processing unit 301 adds the browsing history acquired in step S401 to the search query field in the document storage table 601 in the search target storage area 303 shown in Fig. 6. Details will be described later.

[0047] The document storage table 601 has a management structure equivalent to that of a general full-text search system, and has a number of fields for storing each part of a document, and an index for high-speed full-text search of the fields. Each field can also have multiple indexes based on search terms, word identification methods for registered documents (morphological analysis, n-gram, etc.), and scoring methods (tf·idf, BM25, etc.).

[0048] In step S403, the learning processing unit 301 generates learning data from the browsing history acquired in step S401. Details will be described later.

[0049] In step S404, the learning processing unit 301 performs ranking learning based on the learning data generated in step S403, and acquires a learning model as a learning result. In step S405, the learning processing unit 301 stores the learning model, which is the learning result of step S404, in the learning model storage table 502 in the learning data storage area 302.

[0050] (Added search query field) Next, the above-mentioned search query field addition process in step S402 will be described with reference to the flowchart in FIG.

[0051] In step S 701 , the learning processing unit 301 adds a search query field to the document storage table 601 .

[0052] In step S702, the learning processing unit 301 aggregates search queries in the browsing history to be processed into one text for each document ID of a related document.

[0053] In step S703, the learning processing unit 301 starts the repetitive process up to step S705 for each of the aggregated related documents.

[0054] In step S704, the learning processing unit 301 associates the search query with the document ID of the related document to be processed and adds the search query to the search query field of the document storage table 601 for the aggregated related documents to be processed.

[0055] In step S705, if there are still aggregated related documents to be processed, the learning processing unit 301 repeats the process of step S703. If there are no more aggregated related documents to be processed, the process proceeds to step S706.

[0056] In step S706, the learning processing unit 301 starts repeating the process up to step S708 for each registered document in the document storage table 601.

[0057] In step S707, if the search query field for the registered document to be processed is empty, the learning processing unit 301 proceeds to step S708, or if the search query field for the registered document is not empty, the learning processing unit 301 proceeds to step S709.

[0058] In step S708, the learning processing unit 301 substitutes a complementary string based on a value of another field of the registered document to be processed into the search query field. In this embodiment, a string combining "title" and "body" is used.

[0059] The values ​​assigned to the browsing history may be limited to certain fields, or may be selected for each registered document based on statistical characteristics such as similar string lengths, or even characteristic words (characteristic words) or sentences may be extracted and used in comparison with other documents.

[0060] In step S709, if there is still a registered document to be processed, the learning processing unit 301 executes the process from step S706, If there is no registered document to be processed, the process proceeds to step S710.

[0061] In step S710, the learning processing unit 301 creates an index for the search query field. In this embodiment, it is assumed that BM25 scoring is performed for two word extraction methods, "morphological analysis" and "n-gram".

[0062] In this way, by storing the search query when the document was viewed in the past in association with the document, and executing subsequent search processing using the search query as a search target, it becomes possible to appropriately search for documents desired by the user. Also, for documents for which there is no search query when the document was viewed in the past, elements related to the document are regarded as search queries and stored in association with them, so that an appropriate search score can be calculated even for documents that have no browsing history in the past, and search results can be displayed with an appropriate ranking. (Learning data generation process) Next, the learning data generation process in step S403 described above will be described with reference to the flowchart in FIG.

[0063] In step S801, the learning processing unit 301 initializes the learning data in the temporary area.

[0064] In step S802, the learning processing unit 301 starts the repeated process up to step S806 for each browsing history acquired in step S401.

[0065] In step S803, the learning processing unit 301 requests the search processing unit 304 to generate a feature vector for the search query of the browsing history to be processed, and acquires the feature vector. The generation of the feature vector in the search processing unit 304 will be described later.

[0066] In step S804, if the feature vector acquired in step S803 includes the document in the browsing history to be processed, the learning processing unit 301 proceeds to step S805, otherwise, the learning processing unit 301 proceeds to step S806.

[0067] In step S805, the learning processing unit 301 adds, from among the feature vectors for the search query acquired in step S803, a feature vector corresponding to a document in the browsing history to be processed, to the learning data in the temporary area in association with the document.

[0068] In step S806, if there is still browsing history to be processed, the learning processing unit 301 repeats the process from step S802. If there is no browsing history to be processed, the process ends.

[0069] (Feature vector generation process for search queries) Next, the process of generating a feature vector for a search query by the search processing unit 304 in step S803 described above will be described with reference to the flowchart in FIG.

[0070] In step S901, the search processing unit 304 creates and initializes a feature vector for the search query to be processed in a temporary area.

[0071] In step S902, the search processing unit 304 starts a repeat process up to step S911 for each field that is a full-text search target specified for the registered document in the document storage table 601 of the search target storage area 303, including the search query field.

[0072] In step S903, the search processor 304 starts repeating the process up to step S910 for each index assigned to the field to be processed.

[0073] In step S904, the search processor 304 searches the index to be processed using the search query to be processed, and obtains documents and scores as search results.

[0074] In step S905, the search processor 304 starts the repetitive process up to step S909 for each document in the search results acquired in step S904.

[0075] In step S906, if the feature vector for the document exists in the temporary area, the search processing unit 304 moves the process to step S908, or if the feature vector for the document does not exist in the temporary area, the process moves to step S907.

[0076] In step S907, the search processor 304 adds the initialized feature vector to the temporary area in association with the document. The feature vector for the document is initialized to 0 with the score for each index of each search field as an element.

[0077] In step S908, the search processing unit 304 assigns the score of the document in the search results as an element for the field and index to be processed in the feature vector for the document.

[0078] In step S909, if the search result still contains the document to be processed, the search processing unit 304 repeats the process from step S905, otherwise the process proceeds to step S910.

[0079] In step S910, if there is still an index to be processed, the search processor 304 repeats the process from step S903, If there is no index to be processed, the process proceeds to step S911.

[0080] In step S911, if there are still fields to be processed, the search processing unit 304 repeats the process from step S902. If there are no more fields to be processed, the process ends. Fig. 17 is a diagram conceptually showing the generation process of a feature vector.

[0081] (Specific example of learning process) Next, as a specific example of the learning process, a case will be described in which the learning process shown in FIG. 4 is performed on the browsing history storage table 501 shown in FIG. 10 and the document storage table 601 shown in FIG.

[0082] In step S401, the learning processing unit 301 acquires a past browsing history from the browsing history storage table 501 shown in FIG.

[0083] 10, the document ID for uniquely identifying a viewed document and the search query when the document was viewed are associated and stored as the viewing history in the viewing history storage table 501. For example, the record 1001 in Fig. 10 indicates that a search process was executed using the search query "to execute a script", and a document with document ID "d3" was viewed.

[0084] In step S402, the learning processing unit 301 starts the process of adding the search query field shown in the flowchart of FIG. 7 to the document storage table 601 shown in FIG. 11, in which the browsing history acquired in step S401 is added. In step S 701 , the learning processing unit 301 adds a search query field 1101 to the document storage table 601 .

[0085] In step S702, the learning processing unit 301 acquires the browsing history 1201 summarized by the document IDs of related documents shown in FIG. 12 from the browsing history table 501 shown in FIG.

[0086] As shown in FIG. 12, the process in step S702 aggregates search queries associated with the same document ID.

[0087] In step S703, the learning processing unit 301 starts a repeat process up to step S705 for the related documents.

[0088] In step S704, the learning processing unit 301 adds the aggregated search query 1202 to the search query field of the registered document 1102 having the same document ID (d1) in the document storage table 601.

[0089] Similarly, the learning processing unit 301 repeats the process from step S703 to step S705 for all the aggregated related documents, and the document storage table 601 becomes in the state shown in FIG.

[0090] As shown in FIG. 11, the document storage table 601 includes a document ID for uniquely identifying a document, a document title, a body text which is the content of the document, and an aggregated search query. In step S706, the learning processing unit 301 starts repeating the process up to step S708 for the registered document 1102 in the document storage table 601.

[0091] In step S707, the learning processing unit 301 determines that the search query field for the registered document 1102 is not empty, and therefore the process proceeds to step S709.

[0092] In step S709, since there is still a registered document 1103 to be processed, the learning processing unit 301 executes the process from step S706.

[0093] In step S706, the learning processing unit 301 starts repeating the process up to step S708 for the registered document 1103 in the document storage table 601.

[0094] In step S707, the learning processing unit 301 advances the process to step S708 because the search query field for the registered document 1103 to be processed is empty.

[0095] In step S708, the learning processing unit 301 combines the values ​​of the “title” and “body” fields of the registered document 1103 to be processed and assigns the combined value to the search query field of the registered document 1103.

[0096] Similarly, the learning processing unit 301 repeats the process from step S706 to step S709 for all the aggregated related documents, and the document storage table 601 becomes in the state shown in FIG.

[0097] As shown in FIG. 13, for records in which the search query field was empty in the document storage table (FIG. 11) created by the processing up to step S705, the contents of the search query (in this embodiment, the combined value of the title and body text) have been added.

[0098] In step S710, the learning processing unit 301 creates an index for the search query field 1101.

[0099] In step S403, the learning processing unit 301 performs the learning data generation process shown in FIG.

[0100] In step S801, the learning processing unit 301 initializes the learning data in the temporary area.

[0101] In step S802, the learning processing unit 301 starts a repeat process up to step S806 for the search query of the browsing history 1001 in the browsing history table 501.

[0102] In step S803, the learning processing unit 301 requests the search processing unit 304 to perform the feature vector generation process shown in FIG. 9 for the search query of the browsing history 1001 to be processed, and acquires a feature vector 1401 as shown in FIG.

[0103] Regarding the generation of the feature vector, a schematic diagram is shown in FIG. 17, and a detailed explanation is omitted here.

[0104] In step S804, the learning processing unit 301 advances the process to step S805 because the feature vector 1401 generated in step S803 includes the document (ID=d3) in the browsing history to be processed.

[0105] In step S805, the learning processing unit 301 associates the document in the browsing history 1001 to be processed with the corresponding feature vector 1402, and adds the associated feature vector to the learning data 1501 in the temporary area (1502).

[0106] In step S806, since there is still browsing history 1002 to be processed, the learning processing unit 301 repeats the process from step S803.

[0107] Thereafter, the same process is repeated to generate learning data.

[0108] In step S404, the learning processing unit 301 performs ranking learning based on the learning data generated in step S403, and acquires a learning model as shown in FIG.

[0109] In step S405, the learning processing unit 301 stores the learning model, which is the learning result of step S404, in the learning model storage table 502 in the learning data storage area 302.

[0110] (Search processing) Next, the search process using the learning model acquired in step S404 will be described with reference to the flowchart in FIG.

[0111] FIG. 18 is a flowchart showing a process in which the CPU 201 of the full-text search device 100 reads and executes a predetermined control program, and which searches for matching documents from among registered documents in response to a search request from a user via the browser 310 of the client terminal 110.

[0112] In step S1801, the search processor 304 acquires a search query as a search request from the user.

[0113] In step S1802, the search processing unit 304 generates a feature vector for the search query acquired in step S1801 by the process shown in FIG.

[0114] In step S1803, the search processing unit 304 starts the repetitive process up to step S1805 for each feature vector included in the feature vector for the acquired search query.

[0115] In step S1804, the search processing unit 304 refers to the learning model in the learning model storage table 502 in the learning data storage area 302 for the feature vector to be processed, and calculates a score for the document associated with the feature vector to be processed.

[0116] In step S1805, if there is still a feature vector to be processed, the search processing unit 304 repeats the process from step S1803, otherwise the process proceeds to step S1806.

[0117] In step S1806, the search processing unit 304 sorts the documents associated with the feature vectors according to the scores obtained in S1804, and generates the search results.

[0118] (Example of search processing) Next, as a specific example of search processing, we will explain the case where the search processing shown in Figure 18 is performed using the learning model shown in Figure 16 for the search query ``Regarding editing scripts'' targeting document d2 that is not included in the browsing history.

[0119] In step S1801, the search processor 304 acquires a search query "regarding editing a script" as a search request from a user.

[0120] In step S1802, the search processing unit 304 generates, by the process shown in FIG. 9, a feature vector 1901 shown in FIG. 19 as a feature vector for the search query "regarding editing of scripts" acquired in step S1801.

[0121] In step S1803, the search processing unit 304 starts the repetitive process up to step S1805 for the feature vector 1902 included in the feature vector 1901 for the obtained search query "regarding editing of scripts."

[0122] In step S1804, the search processing unit 304 refers to the learning model shown in FIG. 16 for the feature vector 1902 to be processed, and calculates 167.705 as the score for document d2 associated with the feature vector 1902.

[0123] In step S1805, since there are still feature vectors 1903 to be processed, the search processing unit 304 repeats the process from step S1803.

[0124] The same process is repeated thereafter to obtain the calculation result 2001 shown in FIG.

[0125] In step S1806, the search processing unit 304 rearranges the feature vectors according to the scores calculated in S1804, and generates the search results.

[0126] Document d2 is not included in the browsing history. However, because completion has been performed on the search query field, the value (2004, 2005) for the search query in the feature vector is not 0, and document d2 scores higher than documents d1 and d7.

[0127] In this way, by complementing the search query field, it is possible to obtain an appropriate score even for documents that are not included in the browsing history, thereby reducing the negative impact on the display ranking.

[0128] As described above, in the present invention, a search query when a document was viewed in the past is stored in association with the document, and the search query is also used as a search target in subsequent search processing, thereby making it possible to appropriately search for documents desired by a user. In addition, for documents for which a search query when the document was viewed in the past does not exist, elements related to the document are regarded as search queries and stored in association with the search queries, making it possible to calculate an appropriate search score even for documents that have no browsing history in the past, and to display search results in an appropriate ranking. The present invention can be embodied, for example, as a system, an apparatus, a method, a program, or a recording medium, etc. Specifically, the present invention may be applied to a system composed of a plurality of devices, or may be applied to an apparatus composed of a single device.

[0129] The program of the present invention is a program that enables a computer to execute the processing methods of the flowcharts shown in Figures 4, 7 to 9, and 18, and the storage medium of the present invention stores a program that enables a computer to execute the processing methods of Figures 4, 7 to 9, and 18. The program of the present invention may be a program for each processing method of each device in Figures 4, 7 to 9, and 18.

[0130] As described above, it goes without saying that the object of the present invention can be achieved by supplying a recording medium on which a program that realizes the functions of the above-mentioned embodiments is recorded to a system or device, and having the computer (or CPU or MPU) of that system or device read and execute the program stored on the recording medium.

[0131] In this case, the program itself read out from the recording medium will realize the novel functions of the present invention, and the recording medium on which the program is recorded will constitute the present invention.

[0132] Examples of recording media for supplying the program include flexible disks, hard disks, optical disks, magneto-optical disks, CD-ROMs, CD-Rs, DVD-ROMs, magnetic tapes, non-volatile memory cards, ROMs, EEPROMs, silicon disks, etc.

[0133] Furthermore, it goes without saying that the functions of the above-mentioned embodiments are not only realized by the computer executing a program it has read, but also includes cases where an OS (operating system) running on the computer performs some or all of the actual processing based on the instructions of the program, and the functions of the above-mentioned embodiments are realized through that processing.

[0134] Furthermore, it goes without saying that this also includes cases where a program read from a recording medium is written into a memory provided on a function expansion board inserted into a computer or a function expansion unit connected to a computer, and then a CPU or the like provided on the function expansion board or function expansion unit performs some or all of the actual processing based on the instructions of the program code, thereby realizing the functions of the above-mentioned embodiments.

[0135] The present invention may be applied to a system consisting of multiple devices, or to an apparatus consisting of a single device. Needless to say, the present invention can also be applied to a case where the object is achieved by supplying a program to a system or apparatus. In this case, the effect of the present invention can be enjoyed by the system or apparatus by reading a recording medium storing a program for achieving the present invention into the system or apparatus.

[0136] Furthermore, by downloading and reading a program for achieving the present invention from a server, database, etc. on a network using a communication program, the system or device can enjoy the effects of the present invention. Note that the present invention also includes configurations that combine the above-mentioned embodiments and their modified examples. [Explanation of symbols]

[0137] 100 Full-text search device 110 Client terminal 120 LAN

Claims

1. a storage means for storing a search query used when the document was searched for in association with the document as complementary information, and for a document for which the search query does not exist, storing information based on the contents of the document in association with the complementary information; a search means for performing a search using a search query received from a user, the search being performed on the contents of the document stored in the storage means and complementary information stored in association with the document; An information processing system comprising:

2. The information processing system according to claim 1, characterized in that, for a document for which the search query does not exist, the storage means stores information combining contents related to multiple items in the document as complementary information in association with the document.

3. 3. The information processing system according to claim 1, wherein the storage means stores, for a document for which the search query does not exist, feature words extracted from the document as complementary information in association with the document.

4. An information processing system as described in any one of claims 1 to 3, characterized in that when the document is searched and viewed, the storage means stores the search query used when the document was searched in correspondence with the search query as complementary information.

5. a storage step in which a storage means of the information processing system stores a search query used when the document is searched for in association with the document as complementary information, and stores information based on the content of the document in association with the document for which the search query does not exist as complementary information; a search step in which a search means of the information processing system searches the contents of the documents stored in the storage means and the complementary information stored in association with the documents by using a search query received from a user; An information processing method comprising:

6. A program for causing a computer to function as each of the means according to any one of claims 1 to 4.

Images