JP6524668B2 - Document retrieval apparatus, document retrieval method, program, - Google Patents

Description

  The present invention relates to a document search apparatus and the like for searching for highly relevant documents or document information from a plurality of documents.

  BACKGROUND Conventionally, a search engine (eg, google (registered trademark) or the like) for searching for desired information from a plurality of pieces of information exists.

  Patent Document 1 discloses a search for presenting words or word information existing between input words based on a word space generated by repeating a search for other words related to the input word each time the user inputs a word. The system is described.

JP 2012-123639 A

  However, in the technology described in Patent Document 1 described above, search results presented to the user are limited between input words, and thus new information that the user does not expect at all has never been obtained.

  The present invention has been made in view of the above-described problems, and its object is to search for a document or the like highly relevant to the word selected by the user, and to perform a display according to the degree of association An apparatus etc. are provided.

In order to solve the above-mentioned problems, a first invention relates to each document, document information which is information about each document or word information appearing in each document, and a document word which is a degree of association between each document and each word information storage means for storing a relevance or document relevancy is relevance between each document, first display means for selectably displaying the word information or document information on the screen, if the word information has been selected To calculate the association score with each document by referring to the document word association degree, and when document information is selected, calculate the association score with each document by referring to the inter-document association degree. , based on the calculated relevance score, the display extracting means for extracting a high document relevant with the selected word information or document information, the information before SL document extracted by the extraction unit to the screen comprising a second display means, said extracting means Document, characterized in that generating a score list of merging related score calculated for each document to all the word information or document information that may have been selected, to retrieve documents based on the score list It is a search device.

According to a second aspect of the present invention, a computer is a document, document information that is information about each document, word information appearing in each document, or information about each document, and document word relevance or each document that is a degree of association between each document and each word information Storing the inter-document association degree, which is the inter-document degree, the first display step for displaying the word information or the document information on the screen in a selectable manner, and the document when the word information is selected The association score with each document is calculated with reference to the word association degree, and when the document information is selected, the association score with each document is calculated with reference to the inter-document association degree, and the calculated association based on the score, and a second display step for displaying an extraction step of extracting a high document relevant with the selected word information or document information, the information of the document extracted by the previous SL extraction step on the screen include the Out step, characterized in that generating a score list of merging related score calculated for each document to all the word information or document information that may have been selected, to retrieve documents based on the score list Document retrieval method.

A third aspect of the invention relates to a computer, document information that is each document, word information appearing in each document, or document information that is information about each document, and document word relevance or each document that is a degree of association between each document and each word information Storage means for storing the inter-document association degree, which is the association degree between the first and second display means for selectively displaying word information or document information on the screen, and the document when the word information is selected The association score with each document is calculated with reference to the word association degree, and when the document information is selected, the association score with each document is calculated with reference to the inter-document association degree, and the calculated association based on the scores, and second display means for displaying an extraction means for extracting a high document relevant with the selected word information or document information, the information of the document extracted by the pre-Symbol extracting means on the display screen Function as a document search device equipped with A so Help program, the extracting unit may generate a score list of merging related score for each document that has been calculated for all the word information or document information that may have been selected, to the score list It is a program characterized by extracting a document based on it .

  According to the present invention, it is possible to provide a document search device or the like for searching for a document or the like highly relevant to the word selected by the user and performing display according to the degree of association.

System configuration diagram showing a configuration example of a document search system Block diagram showing an example of the hardware configuration of the document search server and terminal Flow chart showing the flow of the document registration process A diagram showing an example of data stored in a document DB, a word DB, and a document word frequency DB Flow chart showing a flow of inter-word relevance data generation processing Flow chart showing the flow of inter-document relevance data generation processing Flow chart showing the flow of document word relevance data generation process Diagram showing an example of data stored in inter-word relevance DB, inter-document relevance DB, and document-word relevance DB Flow chart showing the flow of the document search process A diagram showing an example of the first screen Diagram showing an example of an array that holds history Figure showing an example of a score list A diagram showing an example of the second screen A diagram showing an example of the selection screen

Hereinafter, preferred embodiments of the present invention will be described in detail based on the drawings.
In the present invention, the text data includes at least one of a document, a word appearing in the document, and information on the document (such as an author name of the document).

  FIG. 1 is a view showing a configuration example of a document search system 200 according to the present embodiment. As shown in FIG. 1, the document search server 100 and one or more terminals 101 used by a user are mutually connected by communication via a network 102.

  The document search server 100 is a server device that stores various databases (see FIGS. 3 and 5). Details will be described later.

  The terminal 101 is a computer used by the user, and displays a search result and the like transmitted from the document search server 100 via the network 102. Further, the terminal 101 receives a selection of a word from the user from the first screen (FIG. 10) and the second screen (FIG. 13) described later, and transmits the word selection to the document search server 100. The terminal 101 may be a portable terminal, a mobile terminal or the like instead of a general-purpose computer.

  The function of the document search system 200 may be realized by integrating the functions of the document search server 100 and the terminal 101, which will be described later, into a single computer. Also, the functions of the document search server 100 do not have to be built on a single computer, but may be built on separate computers connected by the network 102.

  The document search system 200 in the present embodiment is for searching for document information highly relevant to the word and presenting it on the terminal 101 for the word selected by the user.

  FIG. 2 is a block diagram showing an example of the hardware configuration of the document search server 100 (terminal 101) according to the present embodiment. As shown in FIG. 2, the document search server 100 (terminal 101) may, for example, control unit 11, storage unit 12, media input / output unit 13, communication control unit 14, input unit 15, display unit 16, peripheral device I / The F unit 17 and the like are connected via the bus 18 and configured.

The control unit 11 has a central processing unit (CPU) and a read only (ROM).
Memory), RAM (Random Access Memory), etc.
The CPU calls a program stored in the storage unit 12, the ROM, the storage medium, etc. to a work memory area on the RAM and executes it, controls driving of each device connected via the bus 18, and controls the document search server 100 ( The terminal 101) implements the processing to be described later. The ROM is a non-volatile memory and permanently holds a computer boot program, a program such as a BIOS, data and the like. The RAM is a volatile memory, which temporarily holds loaded programs, data and the like, and includes a work area used by the control unit 11 to perform each processing.

  The storage unit 12 is an HDD (Hard Disk Drive) or the like, and stores a program executed by the control unit 11, data necessary for program execution, an OS (Operating System), and the like. These program codes are read by the control unit 11 as necessary, transferred to the RAM, read by the CPU, and executed.

The media input / output unit 13 is, for example, a media input / output device such as a CD drive, a DVD drive, an MO drive, or a floppy (registered trademark) disk drive, and performs data input / output.
The communication control unit 14 includes a communication control device, a communication port, and the like, is a communication interface that mediates communication between a computer and a network, and controls communication with other devices via the network.

The input unit 15 performs data input, and includes, for example, an input device such as a keyboard, a pointing device such as a mouse, and a numeric keypad. The input data is output to the control unit 11.
The display unit 16 includes, for example, a display device such as a CRT monitor or a liquid crystal panel, and a logic circuit (video adapter or the like) for executing display processing in cooperation with the display device. The displayed information is displayed on the display device.
The input unit 15 and the display unit 16 may be, for example, a display with a touch panel in which those functions are integrated.

A peripheral device I / F unit (interface) 17 is a port for connecting a peripheral device to a computer, and the computer transmits and receives data to and from the peripheral device via the peripheral device I / F unit 17. The peripheral device I / F unit 17 is configured by USB, IEEE 1394, RS-232C or the like, and normally has a plurality of peripheral device I / F. The connection form with the peripheral device may be wired or wireless.
The bus 18 is a path that mediates the exchange of control signals, data signals, and the like between the devices.

  Subsequently, a document registration process executed when the document search server 100 receives registration of a new document will be described using FIGS. 3 and 4. FIG. 3 is a flowchart showing the flow of the document registration process. FIG. 4 is a diagram showing an example of data stored in the document DB, the word DB, and the document word frequency DB.

  When the control unit 11 of the document search server 100 receives an input of a document from the input unit 15 or the like (step S101), the control unit 11 assigns a new document ID 21 and registers it in the document DB (step S102).

  FIG. 4A shows an example of the document DB. The document DB is a DB that holds document information, and data is added each time registration of a new document is received. The document DB stores information such as a document ID 21, a data type 22, a heading 23, an author 24, a theme name 25, and a text 26. The document ID 21 is a value that uniquely identifies a document, and is assigned each time a document is accepted.

  In the present embodiment, the documents are news, commentary on the subject, books, magazines, articles, articles, other documents, and the like. The data type 22 is for storing the type of document as described above.

  The control unit 11 of the document search server 100 executes morphological analysis on the newly registered document (document headline 23, author 24, theme name 25, text 26 etc.) (step S103), and extracts the key word. (Step S104). Morphological analysis and keyword extraction can use, for example, general-purpose software.

  The control unit 11 of the document search server 100 determines whether or not the extracted important word is already registered in the word DB, and when it is not registered, the new word ID 31 is assigned and the notation 32 thereof is displayed. The word is registered in the word DB (step S105).

  FIG. 4B shows an example of the word DB. The word DB is a DB that holds word information, and stores word ID 31 and notation 32 in association with each other. If the key word extracted in step S104 of the document registration process is not registered in the word DB, data is newly added. The word ID 31 is a value that uniquely identifies a word, and is assigned each time data is added.

  The control unit 11 of the document search server 100 counts, for each important word, the number of times the important word extracted in step S103 appears in the newly registered document. The control unit 11 of the document search server 100 registers the document ID 21 of the document, the word ID 32 of the important word, and the number of appearances in the document word frequency DB (step S107), and ends the process.

  FIG. 4C shows an example of the word document word frequency DB. The word document word frequency DB is a value representing the frequency (frequency of appearance 43) of the appearance of the key word of the word ID 42 in the document of the document ID 41. The document ID 41 is associated with the document ID 21, and the word ID 42 is associated with the word ID 31.

  As described above, when a new document registration is received, the document search server 100 registers document information such as data type, headline, author, and text in the document DB, and extracts important words of the document. The document search server 100 registers the extracted important words in the word DB, counts the number of appearances in the document, and registers the number in the document word frequency DB. Each time the document registration process is performed, data as many as the number of extracted important words is newly added to the document word frequency DB.

  Next, referring to FIG. 5 to FIG. 8, the document search server 100 calculates the degree of association between words, the degree of association between documents, and the degree of association between word documents using the DB shown in FIG. The process of storing in advance will be described.

FIG. 5 is a flow chart showing the flow of inter-word relevance data generation processing.
The control unit 11 of the document search server 100 selects and inputs the word A and the word B from the word DB (FIG. 4B) stored in the storage unit 12 (step S201). The control unit 11 of the document search server 100 calculates the degree of inter-word association between the word A and the word B (step S202).

  As the degree of association between words, for example, a mutual information amount which is a measure of the degree of co-occurrence of the word A and the word B in the document is calculated. Specifically, the number of documents in which word A appears, the number of documents in which word B appears, and the number of documents in which word A and word B both appear using the document word frequency DB (FIG. 4C). Count. Using these values, the mutual information amount of the word A and the word B is calculated.

  The control unit 11 of the document search server 100 adds the word ID of word A and the word ID of word B, and the calculated inter-word relevance to the inter-word relevance DB (step S203), and ends the process. .

  As described above, the document search server 100 calculates the degree of association between two different words registered in the word DB, and stores the degree of association in the inter-word association degree DB.

  FIG. 8 is a diagram showing an example of data stored in the inter-word relevance degree DB, the inter-document relevance degree DB, and the document word relevance degree DB.

  The inter-word association degree DB shown in FIG. 8A stores the degree of association 53 between different words (word ID 51 and word ID 52) registered in the word DB. Every time the document search server 100 executes the inter-word association degree data generation process shown in FIG. 5, new data is added to the word association degree DB.

FIG. 6 is a flowchart showing the flow of inter-document relevance data generation processing.
The control unit 11 of the document search server 100 selects and inputs the document A and the document B from the document DB (FIG. 4A) stored in the storage unit 12 (step S301). The control unit 11 of the document search server 100 calculates the degree of inter-document relevance between the document A and the document B (step S302).

  Here, as the inter-document relevance, for example, the cosine similarity of the document A and the document B is calculated. Specifically, a word appearing in the document A and its appearance frequency, and a word appearing in the document B and its appearance frequency are acquired using the document word frequency DB (FIG. 4C). Using these, a document vector A and a document vector B expressing features of the document A and the document B are generated, and cosine similarity is calculated based on the document vector A and the document vector B.

  The control unit 11 of the document search server 100 newly adds the document ID of the document A and the document ID of the document B and the calculated inter-document relevance to the inter-document relevance DB (step S303), and ends the process. .

  As described above, the document search server 100 calculates the degree of association between two different documents registered in the document DB, and stores the degree of association in the inter-document association degree DB.

  The inter-document association degree DB shown in FIG. 8B stores the association degree 63 between different documents (document ID 61 and document ID 62) registered in the document DB. Every time the document search server 100 executes the inter-document relevance degree data generation process shown in FIG. 6, new data is added to the word relevance degree DB.

FIG. 7 is a flowchart showing the flow of the document word relevance data generation process.
The control unit 11 of the document search server 100 selects the document A from the document DB (FIG. 4 (a)) stored in the storage unit 12, selects the word B from the word DB (FIG. 4 (b)), Input is made (step S401). The control unit 11 of the document search server 100 calculates the degree of document-word association between the document A and the word B (step S402).

  Here, for example, the cosine similarity of the document A and the word B is calculated as the document word relevance. Specifically, the appearance frequency of the word B in the document A is acquired using the document word frequency DB (FIG. 4C) to generate a document vector of the document A, and the cosine similarity of the document A and the word B is generated. Calculate

  The control unit 11 of the document search server 100 adds the document ID of the document A, the word ID of the word B, and the calculated document word relevance to the document word relevance DB (step S403), and ends the process. .

  As described above, the document search server 100 calculates the degree of association between the document registered in the document DB and the word registered in the word DB, and stores the degree of association in the document word association DB.

  The document word degree of association DB shown in (c) of FIG. 8 stores the degree of association 73 between the document (document ID 71) registered in the document DB and the word (word DB 72) registered in the word DB. Each time the document search server 100 executes the document word relevancy data generation process shown in FIG. 7, new data is added to the word relevancy DB.

  Next, the document search process performed by the document search system 200 will be described with reference to FIGS. 9 to 14. The document search server 100 executes the document registration process of FIG. 2 and the association degree data generation process of FIGS. 5 to 7, and the DB shown in FIG. 4 and the DB shown in FIG. Is assumed to be stored.

  In the screen before transition displayed on the terminal 101, the terminal 101 accepts the selection of a word from the user, and the document search server 100 searches for a document and a word highly relevant to the selected word and searches the terminal 101 Send the result. In the screen after transition on which the search result is displayed, the terminal 101 receives a new selection of words from the user, and the document search server 100 searches for documents and words highly relevant to the selected words again. As described above, the document search process is a process that is repeatedly executed.

FIG. 9 is a flowchart showing the flow of the document search process.
The control unit 11 of the document search server 100 transmits a first screen (screen before transition) for displaying the search key in which the word is arranged to the terminal 101, and the control unit 11 of the terminal 101 receives the first screen The first screen is displayed (step S501).

  FIG. 10 shows an example of the first screen 80. As shown in FIG. As illustrated, the first screen 80 displays a plurality of search buttons 81a, 81b, 81c, and 81d in which words are arranged. The words arranged in the search button 81 are word notation, an author of a document, and the like. Selection of the word displayed on the initial screen may be appropriately performed by the document search server 100 in accordance with the tendency of the user who uses the document search system 200.

  The size of the search button 81 represents the number of appearances or importance of a word or document associated with a word, and the document search server 100 may calculate the size of the search button 81 using the document word frequency DB. it can. This allows the user to recognize at a glance the frequency or importance of a document or the like indicated by the displayed word. Also, the color, shape, and the like of the search button 81 may be changed depending on whether the word placed on the search button 81 is associated with the word or the document.

  When the search button 81 is selected by the user on the first screen 80, the document search server 100 uses a word or a document that is highly related to the word (hereinafter referred to as a selected word) placed on the selected search button 81. It changes to the 2nd screen (FIG. 13) which displays the search button which was searched and arranged the word which shows them.

  It returns to the explanation of FIG. The control unit 11 of the terminal 101 determines whether selection of a word has been detected (step S502). When the selection of the word is not detected (NO in step S502), the control unit 11 of the terminal 101 returns to step S502.

  When the selection of a word is detected (YES in step S502), the control unit 11 of the terminal 101 transmits the selected word to the document search server 100, and the control unit 11 of the document search server 100 receives the selected word It is stored in an array as a search history and is stored in the storage unit 12 or the RAM of the control unit 11 (step S503).

  FIG. 11 is a diagram showing an example of an array 78 which holds a history. As shown in FIG. 11, in the array 78, words (hereinafter referred to as history words) previously selected by the user are stored. The document search server 1 can store the newly received selected word at the head of the array 78 so as to distinguish it from the history word and store it.

  The control unit 11 of the document search server 100 extracts words from the array (step S504), searches for words and documents associated with the extracted words, and acquires the degree of association (step S505).

  Specifically, the document search server 100 determines whether the word extracted from the array (hereinafter, referred to as an extracted word) is a word related to a word or a word related to a document. If the word is a word, the document search server 100 searches the top N (N is an arbitrary natural number) words having a high degree of association with the extracted word using the inter-word association degree DB, and the association degree To get Also, using the document word degree of association DB, the top N documents having a high degree of association with the extracted word are retrieved and their degree of association is acquired.

  Similarly, when the extracted word is a word related to a document, the document search server 100 uses the inter-document relevance DB and the document word relevance DB to determine the top N documents having a high degree of association with the extracted word. Search top N words and get their relevance.

  The control unit 11 of the document search server 100 determines whether the extracted word is a selected word (step S506). If the extracted word is a selected word (YES in step S506), the control unit 11 of the document search server 100 The score of the retrieved document or word is set by multiplying the acquired degree of association by n (n is an arbitrary value) (step S507).

  Since the selected word is the currently selected word, it is considered to be a word closer to the user's search purpose than the other words (history words) in the array. Therefore, in order to present the user with words and documents highly relevant to the selected word, the degree of relevance of the words and documents highly relevant to the selected word is weighted to calculate a score.

  If the selected word is not selected (NO in step S506), the control unit 11 of the document search server 100 sets the acquired degree of association as the score of the document or word searched as it is (step S507).

  The control unit 11 of the document search server 100 determines whether there is the next word in the array (step S509). If there is the next word (YES in step S509), the control unit 11 of the document search server 100 returns to step S504. That is, the processing of step S504 to step S508 is repeatedly executed by the number of words stored in the array. The document search server 100 searches for a highly relevant document or word for each of the words stored in the array, and obtains its score.

  If there is no next word (NO in step S509), the control unit 11 of the document search server 100 generates screen information after transition (a document score list, a word score list, etc.) based on the score, and the terminal It transmits to 101, and the control part 11 of the terminal 101 displays the screen (2nd screen) after transition on the display part 16 based on the received information (step S510). The control unit 11 of the terminal 101 returns to step S502, and accepts the selection of the word again on the second screen.

  The score list transmitted to the terminal 101 by the document search server 100 in step S510 will be described. FIG. 12 is a diagram showing an example of a score list of words. As shown in FIG. 12, the word score list stores words (relevant words) with high degree of relevance and their scores. Similarly, a document score list (not shown) stores highly relevant documents and their scores.

  The document search server 100 merges the scores of the searched words or documents for each word stored in the array to generate a score list of the words or documents. That is, the document search server 100 generates not only words or documents highly relevant to the currently selected selection word but also words or documents highly relevant to the history word to generate a score list.

  Therefore, the history of words selected by the user in the past can be used to display related documents or words, and search results closer to the purpose of the user's search can be presented. Also, this can be expected to be effective in that new ideas can be proposed to the user.

  FIG. 13 is a screen example of the second screen 82. The second screen 82 shown in the figure is an example of the screen after the transition displayed when the user selects "Microorganisms (81a)" on the first screen 80 shown in FIG. The second screen 82 is divided into two areas (left pane 83 and right pane 84). Similar to the first screen 80, the left pane 83 displays a plurality of search buttons 86a, 86b, 86c and 86e in which words are arranged. "Health and energy expenses (86b)" and "energy saving (86c)", which have high relevance (score) and "energy saving (86c)" with the latest selection word "microbe (86a)", remain on the screen, but "health" with low relevance (score) The care (86d) fades out from the screen, and the "△ (86e)" fades in as a new word having high relevance (score).

  The distance between the selected word search button 86a and the other search buttons 86b, 86c, and 86e reflects the value of the score. In the illustrated example, the score values of “utility expenses (86 b)”, “energy saving (86 c)”, and “Δ business unit (86 e)” are “0.9”, “0.3”, “n”, respectively. In accordance with these, the position of the search button 86 disposed on the left pane 83 is determined. Thereby, the user can recognize at a glance the relevance between the selected word and the word displayed on the search button 86.

  The word or document indicated by the word displayed in the left pane 83 is appropriately selected by the terminal 101 or the document search server 100 based on the word or document score list.

  In the second screen 82, when the user selects the search button 86, the document search server 100 again searches for words or documents highly relevant to the word placed on the selected search button 86, and indicates the words It changes to the screen which displays the search button which arranged. Therefore, the screen changes each time a word is selected, and the displayed search button 86 is replaced.

  In the right pane 84, headlines or theme names 88a, 88b, 88c, 88d, 88e of highly relevant (scored) documents are displayed in accordance with the score list of the document received from the document search server 100. As illustrated, the specific portion or background of the display area of the document may be color-coded according to the data type of the document. The right pane 84 may also display news and the like on the Internet related to the selected word retrieved via the network 102.

  When one of the documents 88 displayed in the right pane 84 is selected by the user, the control unit 11 of the terminal 101 accesses the document search server 100 and stores detailed information of the selected document (stored in the document DB) The selection screen for arranging the information is output to the display unit 16.

  FIG. 14 is a screen example of the selection screen 90. As shown in FIG. As shown in FIG. 14, on the selection screen 90, detailed information 91 of the selected document and a back button 92 are arranged. When the back button 92 is selected by the user, the second screen 82 is returned. This allows the user to be presented with detailed information of the retrieved relevant documents.

  As described above, when the document search system 200 receives the selection of the word indicating the word or the document displayed on the terminal 101, the document search server 100 stores the selected word in the array, and the word stored in the array (selection A word or history word is extracted, and a word or document highly associated with the extracted word is retrieved from the inter-word relevance database DB or inter-document relevance database DB, and the document word relevance database DB to obtain the relevance ratio. If the extracted word is a selected word, it is weighted to the acquired degree of association to obtain the score of the word or document, and if the extracted word is a history word, the acquired degree of association is stored in the array as the word or document score as it is Merge the scores of the words or documents obtained for all the words to create a score list of words or documents. The terminal 101 arranges the distance between the selected word and the word indicating the word or document of the score list based on the score based on the word or document score list, and outputs the distance to the display unit 16.

  Thereby, a highly relevant document or word is retrieved from a plurality of documents, a score that is an index of relevance to the retrieved document or word is calculated, and the relevance to the selected word is expressed by the distance. Display the searched words. In addition to the degree of association with the latest selected word, the score is calculated using the degree of association with the history of the word selected by the user in the past. Therefore, the effect that a word closer to the user's search purpose can be presented can be obtained.

  The document search system 200 of the present embodiment can be utilized, for example, as a communication tool in a company. For example, the company can search for various types of documents from different sources, such as internal and external company news, and in-company development themes, etc. according to the purpose and display documents or words highly relevant to the user in an easy-to-understand manner. can do. The user can easily continue the search operation by touch panel operation or the like through the screen on which the presented search result is displayed.

  The preferred embodiments of the document search system 200 and the like according to the present invention have been described above with reference to the accompanying drawings, but the present invention is not limited to such examples. It is apparent that those skilled in the art can conceive of various modifications or alterations within the scope of the technical idea disclosed in the present application, and of course these also fall within the technical scope of the present invention. Be understood

100 ... document search server 101 ... terminal 102 ... network 200 ... document search system 11 ... control unit 12 ... storage unit 13 ... media input / output unit 14 ... communication control Section 15: Input section 16: Display section 17: Peripheral device I / F section 80: First screen 82: Second screen 90: Selection screen

Claims (3)

Document, word information appearing in each document, or document information which is information about each document, and document-word relevance, which is the degree of association between each document and each word information, or inter-document association, which is the degree of association between each document Storage means for storing the degree of
First display means for selectably displaying word information or document information on the screen;
When word information is selected, association score with each document is calculated with reference to the document word association degree, and when document information is selected, each document is referenced with reference to the inter-document association degree Extracting means for calculating a score of relevance to the document, and extracting a document highly relevant to the selected word information or document information based on the calculated relevance score ;
Comprising a second display means for displaying the information of the document extracted by the pre-Symbol extraction unit to the screen, and
The extraction means generates, for each document, a score list obtained by merging related scores calculated for all selected word information or document information, and extracts a document based on the score list. A document search device characterized by The computer is
Document, word information appearing in each document, or document information which is information about each document, and document-word relevance, which is the degree of association between each document and each word information, or inter-document association, which is the degree of association between each document A memory step for storing the degree of
A first display step of selectably displaying word information or document information on the screen;
When word information is selected, association score with each document is calculated with reference to the document word association degree, and when document information is selected, each document is referenced with reference to the inter-document association degree Calculating an association score with the document, and extracting a document highly relevant to the selected word information or document information based on the calculated association score ;
The information of the document extracted by the previous SL extraction step comprises a second display step of displaying on the screen,
The extraction step generates a score list in which the relevance score calculated for all the selected word information or document information is merged for each document, and the document is extracted based on the score list. The document search method characterized by Computer,
Document, word information appearing in each document, or document information which is information about each document, and document-word relevance, which is the degree of association between each document and each word information, or inter-document association, which is the degree of association between each document Storage means for storing the degree of
First display means for selectably displaying word information or document information on the screen;
When word information is selected, association score with each document is calculated with reference to the document word association degree, and when document information is selected, each document is referenced with reference to the inter-document association degree Extracting means for calculating a score of relevance to the document, and extracting a document highly relevant to the selected word information or document information based on the calculated relevance score ;
A Help program to function as a pre-SL document retrieval apparatus comprising a second display means for displaying information of a document extracted on the screen, a by extraction means,
The extraction means generates a score list obtained by merging, for each document, related scores calculated for all selected word information or document information, and extracts the document based on the score list.
A program characterized by

Images