JP2008146209A - Document retrieval device, document retrieval method and document retrieval program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve usability and retrieval accuracy in document retrieval processing. <P>SOLUTION: An acquisition part 201 acquires an XML document. A generation part 202 generates a node list from the XML document. An input part 203 receives input of a retrieval condition. A calculation part 204 calculates a score showing matching degree of the retrieval condition for each node shown in the node list. A decision part 205 decides whether each the node shown in the node list satisfies a prescribed matching condition or not. An addition part 206 adds the score of the node decided that the node satisfies the prescribed matching condition by the decision part 205 to the score of a parent node to which the node belongs. A determination part 207 determines the node having the high matching degree of the retrieval condition as a retrieval result from the nodes shown in the node list based on the score added by the addition part 206 and the score calculated by the calculation part 204. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a document search apparatus, a document search method, and a document search program for searching a node that matches a search condition input by a natural sentence from a hierarchically structured document set.

  Conventionally, by dividing a series of documents into sub-documents of appropriate units such as chapters and sections and making them into a hierarchical structure, there are many publishing diversity, accuracy of information retrieval, ease of division / combination, etc. It is supposed that the merit of can be obtained. For example, an XML document is a typical example of a hierarchically structured document. In the XML document, document elements and document text are marked up by tags. Therefore, it is possible to model with a tree structure composed of element nodes and text nodes. As a result, it is possible to search for partial documents in node units.

  As a general search method used for a partial document search process for an XML document, there is a method for searching all partial documents in a minimum unit including a specified search character string from XML documents (for example, (See Patent Document 1).

JP-A-11-306205

  However, in the conventional technique described in Patent Document 1, a plurality of minimum unit parts are included even when a plurality of minimum unit partial documents including a search character string are included in the same partial document. Documents are searched in pieces. For example, when a document is divided into chapters, sections, terms, and sentences, partial documents in units of sentences including a search character string are searched in fragments. In such a case, it is desirable for the user to search as a search result a partial document of a unit (for example, a chapter or a section) that includes a plurality of partial documents of the minimum unit.

  In order to deal with such problems, not only all partial documents of the minimum unit (for example, sentence) including the search string are searched, but also various units (for example, chapters, sections, terms, etc.) related to the search string ) Has also been devised, but in this case, some of the searched partial documents have a low relevance to the search character string and are far from what the user intended. Many documents are included.

  Further, in the prior art described in the above-mentioned Patent Document 1, since the search processing considering the strength of relevance is not performed, it is possible to search only a partial document having a stronger relevance to the search character string, The retrieved partial documents cannot be displayed in the order corresponding to the strength of the relationship with the search character string.

  As described above, in the above-described conventional technology, there is a problem that not only the partial documents of an appropriate unit and number cannot be searched but also the searched partial documents cannot be displayed in an appropriate order. .

  In order to solve the above-described problems caused by the prior art, the present invention retrieves partial documents of an appropriate unit and number, and displays the retrieved partial documents in an appropriate order. It is another object of the present invention to provide a document search apparatus, a document search method, and a document search program capable of improving usability.

  In order to solve the above-described problems and achieve the object, a document search apparatus according to the present invention is a document search apparatus that searches a hierarchically-structured document set for a node that matches a search condition input by a natural sentence. An acquisition unit for acquiring the document set, a generation unit for generating a node list from the document set acquired by the acquisition unit, an input unit for receiving input of the search condition, and the generation unit For each node indicated in the node list, a calculation means for calculating a score indicating the degree of match of the search condition based on the search condition input by the input means, and a node list generated by the generation means A determination means for determining whether or not a predetermined conformity condition is satisfied for each indicated node; and a predetermined conformance condition by the determination means. Based on the addition means for adding the score of the node determined to satisfy the score of the parent node to which the node belongs, the score added by the addition means, and the score calculated by the calculation means, And determining means for determining, as a search result, a node having a high degree of matching with the search condition from the node list generated by the generating means.

  According to the present invention, in a document search process, when a node having a text node satisfies a predetermined matching condition, the parent node to which this node belongs can be handled as a node having a high matching degree of the search condition.

  In the document search device according to the present invention, in the above-described invention, output control for controlling output so that nodes determined as search results by the determination unit are displayed in descending order of matching degree of the search conditions. Characterized by further providing means

  According to the present invention, it is possible to control output so that nodes having a high degree of matching of search conditions searched from a document set are displayed in an appropriate order.

  In the document search device according to the present invention as set forth in the invention described above, the determining means is configured to generate the generating means based on the score added by the adding means and the score calculated by the calculating means. The node list generated by the above is sorted in descending order of matching degree of the search condition, and a predetermined number of nodes are determined as search results from the top of the sorted node list.

  According to the present invention, a necessary number of nodes having a higher matching degree of search conditions can be searched from the document set.

  Further, in the document search device according to the present invention, in the invention described above, the calculation means uses the TF-IDF method for each node indicated in the node list generated by the generation means. A score indicating the degree of match of the search condition is calculated based on the search condition input by the input means.

  According to the present invention, in the document search process, by calculating the score using the TF-IDF method, the keyword is not simply a node in which many keywords included in the search condition appear, but the keyword is characterized by the node. Can be handled as a node having a high degree of matching with the search condition.

  The document search method according to the present invention is a document search method for searching a node that matches a search condition input by a natural sentence from a hierarchically structured document set, and acquiring the document set And a generation step of generating a node list from the document set acquired by the acquisition step, an input step of receiving input of the search condition, and for each node indicated in the node list generated by the generation step, Based on the search condition input in the input step, a calculation step for calculating a score indicating the degree of match of the search condition, and a predetermined match for each node indicated in the node list generated by the generation step A determination step for determining whether or not a condition is satisfied, and a score of a node determined to satisfy a predetermined conformity condition by the determination step The node list generated by the generation step based on the addition step of adding to the score of the parent node to which the node belongs, the score added by the addition step, and the score calculated by the calculation step A determination step of determining, as a search result, a node having a high degree of matching with the search condition is executed by a computer.

  According to the present invention, in a document search process, when a node having a text node satisfies a predetermined matching condition, the parent node to which this node belongs can be handled by the computer as a node having a high matching degree of the search condition. it can.

  The document search program according to the present invention is a document search program for searching a hierarchically-structured document set or a node that matches a search condition input by a natural sentence, and acquiring the document set And a generation step of generating a node list from the document set acquired by the acquisition step, an input step of receiving input of the search condition, and for each node indicated in the node list generated by the generation step, Based on the search condition input in the input step, a calculation step for calculating a score indicating the degree of match of the search condition, and a predetermined match for each node indicated in the node list generated by the generation step A determination step for determining whether or not a condition is satisfied, and a no determination that the predetermined determination condition is satisfied by the determination step. Is generated by the generating step based on the addition step of adding the score of the parent node to the score of the parent node to which the node belongs, the score added by the adding step, and the score calculated by the calculating step A determination step of determining, as a search result, a node having a high degree of matching with the search condition from the node list is executed by a computer.

  According to the present invention, in a document search process, when a node having a text node satisfies a predetermined matching condition, the parent node to which this node belongs can be handled by the computer as a node having a high matching degree of the search condition. it can.

  According to the document search device, the document search method, and the document search program according to the present invention, the document search processing is performed by searching for the appropriate unit and number of partial documents and displaying the searched partial documents in an appropriate order. It is possible to improve search accuracy and usability.

  Exemplary embodiments of a document search device, a document search method, and a document search program according to the present invention will be described below in detail with reference to the accompanying drawings using an XML document as an example of a hierarchically structured document set. To do.

(Hardware configuration of document search apparatus 100)
First, the hardware configuration of the document search apparatus according to this embodiment will be described. FIG. 1 is a block diagram showing an example of a hardware configuration of the document search apparatus according to this embodiment.

  In FIG. 1, a document search apparatus 100 includes a CPU (Central Processing Unit) 101, a ROM (Read Only Memory) 102, a RAM (Random Access Memory) 103, an HDD (Hard Disc Drive) 104, and an HD (Hard Disc). ) 105, FDD (Flexible Disc Drive) 106, FD (Flexible Disc) 107, CD-RW (Compact Disc ReWritable) drive 108, CD-RW 109, display 110, keyboard 111, mouse 112, A network I / F (interface) 113, a communication cable 114, and a bus 120 are provided.

  The CPU 101 controls the entire document search apparatus 100. The ROM 102 stores various control programs and the like. The RAM 103 stores variable data in a rewritable manner and functions as a work area for the CPU 101. The HDD 104 controls reading / writing of data with respect to the HD 105 according to the control of the CPU 101. The HD 105 stores data written according to the control of the HDD 104.

  The FDD 106 controls reading / writing of data with respect to the FD 107 according to the control of the CPU 101. The FD 107 is detachable and stores data written according to the control of the FDD 106. The CD-RW drive 108 controls reading / writing of data with respect to the CD-RW (or CD-R, CD-ROM) 109 according to the control of the CPU 101. The CD-RW 109 is detachable and stores data written according to the control of the CD-RW drive 108.

  The display 110 displays a cursor, a menu, a window, or various data such as characters and images. The keyboard 111 includes a plurality of keys for inputting characters, numerical values, various instructions, and the like. The mouse 112 selects and executes various instructions, selects a processing target, moves a mouse pointer, and the like. A network I / F 113 is connected to a network such as a LAN, WAN, or the Internet via a communication cable 114, and functions as an interface between the network and the CPU 101. A bus 120 connects the above-described units.

(Functional configuration of document search apparatus 100)
Next, a functional configuration of the document search apparatus 100 according to this embodiment will be described. FIG. 2 is a block diagram showing a functional configuration of the document search apparatus 100 according to this embodiment.

  As shown in FIG. 2, the document search apparatus 100 includes an acquisition unit 201, a generation unit 202, an input unit 203, a calculation unit 204, a determination unit 205, an addition unit 206, a determination unit 207, and output control. A unit 208 and a display unit 209 are provided.

  The acquisition unit 201 acquires an XML document. For example, the acquisition unit 201 acquires an XML document by reading an XML document file designated by the user. In this case, the XML document file is not limited to the one stored in the document search device 100, and may be stored in another device connected to the document search device 100, for example. Specifically, the acquisition unit 201 realizes its function by the CPU 101 executing a program stored in the ROM 102, RAM 103, HD 105, and FD 107 shown in FIG.

  The generation unit 202 generates a node list from the XML document acquired by the acquisition unit 201. The node list here is a list of all element nodes existing in the XML document based on the XML document modeled in a tree structure. For example, for each element node, an index, a path, etc. Contains information. When a text node belongs to an element node, information such as an index and text of the text node is associated. The node list generated by the generation unit 202 is temporarily stored in the memory in a list format in which the path of each node is shown, for example.

  The generation unit 202 not only generates a node list for all nodes existing in the XML document, but also generates a node list for nodes within a predetermined range and a node list for nodes within a range specified by the user. You may make it do. A specific procedure for generating the node list will be described later with reference to FIG. Specifically, the generation unit 202 realizes its function by the CPU 101 executing programs stored in the ROM 102, RAM 103, HD 105, and FD 107 shown in FIG.

  The input unit 203 receives an input of a search condition (search query sentence). For example, the search condition is input as “J company I Taro” or “(J company I Taro)”, the former means that both “J company” and “I Taro” are included, and the latter is It means that either “Company J” or “I Taro” is included. Here, the search conditions are not limited to those directly input to the document search apparatus 100 by the user, but may be those transmitted from other apparatuses connected to the document search apparatus 100, for example. Specifically, the input unit 203 realizes its function by, for example, the keyboard 111, the mouse 112, the network I / F 113, and the like shown in FIG.

  The calculation unit 204 calculates a score indicating the matching degree of the search condition based on the search condition input by the input unit 203 for each node indicated in the node list generated by the generation unit 202. The score (TF-IDF) can be obtained by the following calculation formula (1).

  TFIDF = TF × log (N / DF) (1)

  In the calculation formula (1), TF indicates the number of appearances of the search character string in the text node. N indicates the total number of text nodes. DF indicates the number of text nodes including the search character string.

  In the present embodiment, the score is calculated using the TF-IDF method. However, the present invention is not limited to this, and the score may be calculated using another method. Specifically, the calculation unit 204 realizes its function when the CPU 101 executes programs stored in the ROM 102, the RAM 103, the HD 105, and the FD 107 shown in FIG.

  The determination unit 205 determines whether a predetermined matching condition is satisfied based on the score calculated by the calculation unit 204 for each node indicated in the node list generated by the generation unit 202. Examples of the matching conditions include “whether the score is greater than or equal to a predetermined value or less than a predetermined value”, “whether the total score with the scores of other peer nodes is equal to or greater than a predetermined value”, “other Whether or not the total number of nodes with the peer node is equal to or less than a predetermined value ”is not limited to this, but other matching conditions may be used. In addition, the matching condition may be set in advance or may be specified by the user. Specifically, the determination unit 205 realizes its function when the CPU 101 executes a program stored in the ROM 102, the RAM 103, the HD 105, and the FD 107 shown in FIG.

  The adding unit 206 adds the score of the node determined by the determining unit 205 as satisfying a predetermined matching condition to the score of the parent node to which this node belongs. For example, when the score of the node “A / B / D” determined to satisfy the predetermined conformity condition is “5” and the score of the parent node “A / B” to which this node belongs is “5” The score of the node “A / B” becomes “10” by the addition processing by the adding unit 206. The addition unit 206 performs the above addition processing on all the nodes indicated in the node list generated by the generation unit 202 in order from the root node (that is, a deep node) to the root node. Specifically, the adding unit 206 realizes its function by the CPU 101 executing a program stored in the ROM 102, the RAM 103, the HD 105, and the FD 107 shown in FIG.

  The determination unit 207 sorts the node list generated by the generation unit 202 based on the score added by the addition unit 206 and the score calculated by the calculation unit 204 in descending order of matching degree of the search condition, A predetermined number of nodes from the top of the sorted node list are determined as search results.

  Note that the number of node searches determined as a search result by the determination unit 207 is not limited to that set in advance, and may be specified by the user, for example. Specifically, the determination unit 207 realizes its function by the CPU 101 executing a program stored in the ROM 102, RAM 103, HD 105, and FD 107 shown in FIG.

  The output control unit 208 controls the output so that the nodes determined as the search results by the determination unit 207 are displayed on the display unit 209 in descending order of matching degree of the search conditions. Note that the output control unit 208 not only controls to display the node determined as the search result by the determination unit 207, but also controls to output to a file or connected to the document search device 100, for example. You may control to transmit to another apparatus. Specifically, the output control unit 208 realizes its function when the CPU 101 executes a program stored in the ROM 102, RAM 103, HD 105, and FD 107 shown in FIG.

  The display unit 209 displays the nodes determined as the search results by the determination unit 207 under the control of the output control unit 208 in descending order of the matching degree of the search conditions. Specifically, the display unit 209 realizes its function by the display 110 shown in FIG. 1, for example.

(Example of XML document)
Next, an example of an XML document used in the document search apparatus 100 according to the embodiment of the present invention will be described. FIG. 3 is an explanatory diagram showing an example of an XML document used in the document search apparatus 100 according to the embodiment of the present invention.

  FIG. 3 shows an XML document “c: ¥ documents ¥ 0123.xml” modeled in a tree structure. In FIG. 3, nodes 1 to 11 indicate element nodes, and numerals “1” to “11” indicate indexes. Nodes A to E indicate text nodes, and the letters “A” to “E” indicate indexes.

  In FIG. 3, for example, the text node A belongs to the element node 4. The text node A has the text “XML, scheme”. For example, when this text node A is indicated using a tag, it is indicated as “<article> <body> <sec> <p1> XML, scheme </ p1> </ sec> </ body> </ article>”. be able to.

(Node list generation procedure by the generation unit 202)
Next, a node list generation procedure by the generation unit 202 will be described. FIG. 4 is a flowchart illustrating an example of a node list generation procedure by the generation unit 202.

  First, one element node is selected from an XML document modeled in a tree structure (step S401). First, the highest element node is selected. For example, in the case of the XML document shown in FIG. 3, element node 1 is selected.

  Next, the element node selected in step S401 is added to the node list (step S402). Here, the information added to the node list includes the index and path of the element node. For example, in the case of the element node 1 in the XML document shown in FIG. 3, the index is “1”, the path is “/ article”, and the like.

  Next, it is determined whether or not a text node belongs to the element node selected in step S401 (step S403). For example, in the case of element node 1 in the XML document shown in FIG. 3, it is determined that the text node does not belong, and in the case of element node 4, it is determined that the text node belongs.

  If it is determined in step S403 that the text node belongs (step S403: Yes), the element node selected in step S401 is associated with the text node belonging to this element node (step S404). ), The process proceeds to step S405.

  Here, the information associated with the element node is a text node index or text. For example, in the case of the text node A in the XML document shown in FIG. 3, the index “A”, the text “XML, scheme”, and the like. On the other hand, if it is determined in step S403 that the text node does not belong (step S403: No), step S404 is skipped and the process proceeds to step S405.

  Next, it is determined whether or not all element nodes included in the XML document have been selected (step S405). If it is determined in step S405 that all element nodes have been selected (step S405: Yes), the series of processing ends. On the other hand, if it is determined in step S405 that all element nodes have not been selected (step S405: No), the next element node is selected based on the element node selected in step S401 in the XML document. (Step S406). At this time, if a lower node and a peer node exist, the lower node is selected with priority. For example, in the XML document shown in FIG. 3, the order in which element nodes are selected is the order of the index numbers.

  Then, the process returns to step S402, and steps S402 to S406 are repeated until it is determined in step S405 that all nodes have been selected. Thereby, all the element nodes included in the XML document can be added to the node list. Further, all text nodes included in the XML document can be associated with any one of the element nodes shown in the node list.

(Example of node list generated by the generation unit 202)
Next, an example of the node list generated by the generation unit 202 will be described. FIG. 5 is an explanatory diagram illustrating an example of a node list generated by the generation unit 202.

  The node list shown in FIG. 5 is a node list generated from the XML document shown in FIG. 3 by the procedure described above with reference to FIG. 4, and columns “index1”, “pass”, “index2”, “text”. It is constituted by.

  Among these, the index of the element node is set in the column “index1”. In the column “pass”, an element node path is set. In the column “index2”, the index of the text node associated with the element node is set. Further, the text of the text node associated with the element node is set in the column “text”.

  For example, from the node list shown in FIG. 5, the element node indicated by the path “/ article / body / sec / p1” to which the index “4” is assigned is assigned the index “A” and the text “XML, It can be determined that a text node including “scheme” is associated.

(Procedure for document search processing by the document search apparatus 100)
Next, a procedure for document search processing by the document search apparatus 100 according to the embodiment of the present invention will be described. FIG. 6 is a flowchart showing an example of a procedure of document search processing by the document search apparatus 100 according to the embodiment of the present invention.

  First, the acquisition unit 201 acquires an XML document (step S601), and the generation unit 202 generates a node list from the XML document acquired in step S601 (step S602). The specific procedure for generating the node list is as described above with reference to FIG.

  Next, input of a search condition is accepted by the input unit 203 (step S603), and the search condition input in step S603 for each node indicated in the node list generated in step S602 by the calculation unit 204. Based on the above, a score indicating the degree of match of the search condition is calculated (step S604).

  Subsequently, the determination unit 205 selects one node shown in the node list generated in step S602 (step S605), and the node selected in step S605 has the score calculated in step S604. Based on this, it is determined whether or not a predetermined conformity condition is satisfied (step S606).

  If it is determined in step S606 that the predetermined conformity condition is satisfied (step S606: Yes), the adder 206 adds the score of the node selected in step S605 to the score of the parent node to which this node belongs. (Step S607), the process proceeds to Step S608. On the other hand, if it is determined in step S606 that the predetermined matching condition is not satisfied (step S606: No), step S607 is skipped and the process proceeds to step S608.

  Subsequently, the determination unit 205 determines whether or not all the nodes shown in the node list generated in step S602 have been selected (step S608). If it is determined in step S608 that all nodes have not been selected (step S608: No), steps S605 to S608 are repeated until it is determined in step S608 that all nodes have been selected.

  On the other hand, if it is determined in step S608 that all the nodes have been selected (step S608: Yes), based on the score added in step S607 by the determination unit 207 and the score calculated in step S604, The node list generated in step S602 is sorted in descending order of matching degree of the search condition (step S609), and a predetermined number of nodes are determined as search results from the sorted node list (step S610). .

  Then, under the control of the output control unit 208, the nodes determined as the search results in step S610 are displayed on the display unit 209 in descending order of the degree of match of the search conditions (step S611), and the series of processing ends.

(Example of score calculated by calculation unit 204)
Next, an example of the score calculated by the calculation unit 204 will be described. FIG. 7 is an explanatory diagram illustrating an example of a score calculated by the calculation unit 204.

  FIG. 7 shows the association between the node list shown in FIG. 5 and the score of each element node calculated by the calculation unit 204. In FIG. 7, the score calculated by the calculation unit 204 is set in a column “score1”. The search character string used in the calculation process at this time is “XML, tag, scheme”.

  For example, from FIG. 7, it is assumed that the score “38” calculated by the calculation unit 204 is associated with the element node indicated by the path “/ article / body / sec / p1” to which the index “4” is assigned. Judgment can be made. Here, this score “38” is calculated by the following TF-IDF calculation formula (2).

  38 (TFIDF: score) = 1 (TF: number of occurrences of the search character string “XML” in the text node) × 20 (IDF: log (total number of text nodes / number of text nodes including the search character string “XML”)) +1 (TF: appearance number of search character string “scheme” in text node) × 18 (IDF: log (total number of text nodes / number of text nodes including search character string “scheme”)) (2)

  Further, it may be determined that the score “80” calculated by the calculation unit 204 is associated with the element node indicated by the path “/ article / body / sec / p2” to which the index “5” is assigned. it can. The score “80” is calculated by the following TF-IDF calculation formula (3).

  80 (TFIDF: score) = 2 (TF: number of occurrences of the search character string “tag” in the text node) × 40 (IDF: log (total number of text nodes / number of text nodes including the search character string “tag”)) ... (3)

  Note that the score calculated by the calculation unit 204 may be included in the node list, or may be included in a table different from the node list.

(Example of score added by adding unit 206)
Next, an example of the score added by the adding unit 206 will be described. FIG. 8 is an explanatory diagram illustrating an example of scores added by the adding unit 206.

  FIG. 8 shows the association between the node list shown in FIG. 5, the score of each element node calculated by the calculation unit 204, and the score added by the addition unit 206. In FIG. 8, the score added by the adding unit 206 is set in a column “score2”.

  For example, from FIG. 8, it is determined that the score “118” added by the adding unit 204 is associated with the element node indicated by the path “/ article / body / sec” to which the index “3” is assigned. be able to. Here, the score “118” includes the score “38” of the element node to which the index “4” is assigned and the score “80” of the element node to which the index “5” is assigned. It is an addition.

  Prior to the addition process, the determination unit 205 determines whether or not the indexes “4” and “5” are added to the index “3”. The determination conditions used in the determination process at this time are “add when the total score is 50 or more” and “add when the number of connected nodes is 100 or less”. Since the indexes “4” and “5” satisfy the above condition, it is determined that the scores are added to the score of the index “3”.

  Further, as shown in FIG. 8, the scores of the indexes “4” and “5” are set to “0” by the adding unit 204. This is to prevent the index “3” and the indexes “4” and “5” from being determined as search results in duplicate.

  Similarly, from FIG. 8, the score “66” added by the adding unit 204 is associated with the element node indicated by the path “/ article / body / sec / title / name” to which the index “9” is assigned. Can be determined. Here, the score “66” includes the score “22” of the element node to which the index “10” is assigned and the score “44” of the element node to which the index “11” is assigned. It is an addition. The scores of indexes “10” and “11” are set to “0” by the adding unit 204.

(Example of node list sorted by determination unit 207)
Next, an example of the node list sorted by the determination unit 207 will be described. FIG. 9 is an explanatory diagram illustrating an example of a node list sorted by the determination unit 207.

  9 is based on the score added by the adding unit 206 as shown in FIG. 8 by the determining unit 207 and the score calculated by the calculating unit 204 as shown in FIG. The node list shown is sorted in descending order of matching degree of search conditions. From the node list thus sorted, the determination unit 207 determines a predetermined number of nodes as search results from the top.

  For example, when the number of searches is specified as three, the determination unit 207 receives the element node to which the index “3” is assigned and the index “8” from the node list illustrated in FIG. Three element nodes, that is, the element node and the element node to which the index “9” is assigned, are determined as search results.

(Example of search result displayed on display unit 209)
Next, an example of the search result displayed on the display unit 209 will be described. FIG. 10 is an explanatory diagram illustrating an example of a search result displayed on the display unit 209.

  FIG. 10 shows a search result displayed on the display unit 209 as a result of performing the document search process according to the procedure described above with reference to FIG. 6 on the XML document shown in FIG. As shown in FIG. 10, in performing the document search process, the search target document “c: ¥ documents ¥ 0123.xml”, the search condition “XML, tag, scheme”, and the number of searches “3 (cases)” are designated by the user. Has been.

  When the “search” button is pressed, a document search process is performed on the search target document “c: \ documents \ 0123.xml”. As a result, the search target document “c: \ documents \ 0123.xml” is executed. The top three nodes determined from "are displayed as search results.

  As described above, according to the document search apparatus 100 according to the present embodiment, an XML document is acquired, a node list is generated from the acquired XML document, an input of a search condition is received, and the generated node list For each node indicated in, a score indicating the degree of match of the search condition is calculated based on the input search condition, and a predetermined matching condition is satisfied for each node indicated in the generated node list Whether or not, and the score of the node determined to satisfy the predetermined matching condition is added to the score of the parent node to which the node belongs, and based on the added score and the calculated score, The generated node list is sorted in descending order of matching degree of search conditions, and a predetermined number of nodes are determined as search results from the sorted node list. The over-de, has a structure in which the matching degree of the search condition displayed in descending order.

  Thereby, in a document search process, when a node having a text node satisfies a predetermined matching condition, a parent node to which this node belongs can be handled as a node having a high matching degree of the search condition. In addition, a required number of nodes having a higher matching degree of the search condition can be searched from the XML document. Furthermore, the retrieved nodes can be displayed in an appropriate order. Therefore, it is possible to search for an appropriate unit and number of partial documents, and to display the searched partial documents in an appropriate order. As a result, the search accuracy and usability in the document search process can be improved. Can be planned.

  Note that the document search apparatus, document search method, and document search program according to the present invention can be applied to document search for documents other than XML documents as long as the document set has a hierarchical structure. Further, the present invention is not limited to a filed document, and can be applied to a document search for a databased document, for example. Furthermore, not only documents filed in a single file or documents databased in a single database, but also document searches for documents filed in multiple files and documents databased in multiple databases. Can be applied.

  The document search method described in this embodiment can be realized by executing a program prepared in advance on a computer such as a personal computer or a workstation. This program is recorded on a computer-readable recording medium such as a hard disk, a flexible disk, a CD-ROM, an MO, and a DVD, and is executed by being read from the recording medium by the computer. The program may be a transmission medium that can be distributed via a network such as the Internet.

  As described above, the document search apparatus, the document search method, and the document search program according to the present invention are a personal computer that searches a hierarchically-structured document set for a node that matches a search condition input by a natural sentence, Suitable for use in document servers, document search software, etc.

It is a block diagram which shows an example of the hardware constitutions of the document search device concerning this embodiment. It is a block diagram which shows the functional structure of the document search device concerning this embodiment. It is explanatory drawing which shows an example of the XML document used for the document search apparatus concerning embodiment of this invention. It is a flowchart which shows an example of the production | generation procedure of the node list by a production | generation part. It is explanatory drawing which shows an example of the node list produced | generated by the production | generation part. It is a flowchart which shows an example of the procedure of the document search process by the document search apparatus concerning embodiment of this invention. It is explanatory drawing which shows an example of the score calculated by the calculation part. It is explanatory drawing which shows an example of the score added by the addition part. It is explanatory drawing which shows an example of the node list sorted by the determination part. It is explanatory drawing which shows an example of the search result displayed on the display part.

Explanation of symbols

100 Document Retrieval Device 101 CPU
102 ROM
103 RAM
104 HDD
105 HD
106 FDD
107 FD
108 CD-RW drive 109 CD-RW
110 Display 111 Keyboard 112 Mouse 113 Network I / F
114 communication cable 120 bus 201 acquisition unit 202 generation unit 203 input unit 204 calculation unit 205 determination unit 206 addition unit 207 determination unit 208 output control unit 209 display unit

Claims (6)

A document search device that searches for a node that matches a search condition input by a natural sentence from a hierarchically structured document set,
Obtaining means for obtaining the document set;
Generating means for generating a node list from the document set acquired by the acquiring means;
Input means for receiving input of the search condition;
Calculation means for calculating a score indicating the degree of match of the search condition based on the search condition input by the input means for each node indicated in the node list generated by the generation means;
Determination means for determining whether or not a predetermined conformity condition is satisfied based on the score calculated by the calculation means for each node indicated in the node list generated by the generation means;
Adding means for adding a score of a node determined to satisfy a predetermined matching condition by the determining means to a score of a parent node to which the node belongs;
Based on the score added by the adding means and the score calculated by the calculating means, a node having a high degree of match of the search condition is selected as a search result from the node list generated by the generating means. A decision means to decide;
A document retrieval apparatus comprising:   2. The document search apparatus according to claim 1, further comprising output control means for controlling output so that nodes determined by the determination means are displayed in descending order of matching degree of the search conditions.   The determination unit sorts the node list generated by the generation unit in descending order of matching degree of the search condition based on the score added by the addition unit and the score calculated by the calculation unit. 3. The document search apparatus according to claim 1, wherein a predetermined number of nodes are determined as search results from the top of the sorted node list.   The calculation means uses the TF-IDF method to calculate the degree of match of the search conditions based on the search conditions input by the input means for each node indicated in the node list generated by the generation means. The document search apparatus according to claim 1, wherein a score to be calculated is calculated. A document search method for searching a node that matches a search condition input by a natural sentence from a hierarchically structured document set,
An acquisition step of acquiring the document set;
A generation step of generating a node list from the document set acquired by the acquisition step;
An input step for receiving an input of the search condition;
A calculation step of calculating a score indicating a degree of match of the search condition based on the search condition input by the input step for each node indicated in the node list generated by the generation step;
A determination step for determining whether or not a predetermined matching condition is satisfied based on the score calculated by the calculation step for each node indicated in the node list generated by the generation step;
An adding step of adding the score of the node determined to satisfy the predetermined matching condition by the determining step to the score of the parent node to which the node belongs;
Based on the score added by the adding step and the score calculated by the calculating step, a node having a high degree of match of the search condition is selected as a search result from the node list generated by the generating step. A decision process to decide;
A document retrieval method characterized by causing a computer to execute. A document search program for searching a node that matches a search condition input by a natural sentence from a hierarchically structured document set,
An acquisition step of acquiring the document set;
A generation step of generating a node list from the document set acquired by the acquisition step;
An input step for receiving an input of the search condition;
A calculation step of calculating a score indicating a degree of match of the search condition based on the search condition input by the input step for each node indicated in the node list generated by the generation step;
A determination step for determining whether or not a predetermined matching condition is satisfied based on the score calculated by the calculation step for each node indicated in the node list generated by the generation step;
An adding step of adding the score of the node determined to satisfy the predetermined matching condition by the determining step to the score of the parent node to which the node belongs;
Based on the score added by the adding step and the score calculated by the calculating step, a node having a high degree of match of the search condition is selected as a search result from the node list generated by the generating step. A decision process to decide;
Document search program characterized by causing a computer to execute.

Images