JP4489034B2 - Structured document processing apparatus, structured document processing method, and structured document processing program - Google Patents

Description

  The present invention relates to a structured document processing apparatus, a structured document processing method, and a structured document processing program for processing a structured document.

  In recent years, in order to facilitate the management and processing of document data on a computer, a document (structured document) in which a “tag” indicating a document structure is added to a local document has been widely used. Then, standardization of structured document conventions such as SGML (Standard Generalized Markup Language) and XML (extensible Markup Language) is being promoted as a convention for describing this structured document.

  In a structured document described based on such a structured document convention, for example, a specific keyword is added to text with a tag “headline” in document data, using the text divided by the tag as a search unit. Can be retrieved using the structure of the document data, for example, by detecting document data that includes the text (see, for example, “Patent Document 1”).

  For this reason, if such structured document rules are used, it is possible to search document data including a keyword desired by a user in a specified structure in a document from a database including a plurality of structured documents. Search for appropriate documents.

  In addition, a transposition file method is used in searching document data including a keyword from a large amount of document data (see, for example, “Patent Document 2”). The transposed file method is intended to speed up search processing. However, there are also disadvantages, and depending on the search keyword, the appearance frequency in all documents may be large, and the search processing cost at that time increases in accordance with the frequency. In particular, the cost of reading the index information from the external storage device appears remarkably high.

JP 2000-207409 A JP 2001-265773 A

  When performing a search process designating a structure, if keywords are frequently used, the cost of reading index information is high in the conventional method including the inverted file method. For this reason, there has been a problem that the actual number of hits at least increases the processing cost.

  It is also conceivable to further reduce the processing cost by controlling the index by further subdividing the index into the structure, thereby suppressing the number of search processes that specify the structure. However, it is not realistic because it is expected that the management cost of the index will increase due to the complexity of the structure, the storage efficiency of the index will decrease, and the consumption of resources for the effect will be significant. Further, due to segmentation by structure, performance degradation is expected when the structure is not designated or when a plurality of structures are designated.

  The present invention has been made in view of the above, and provides a structured document processing apparatus, a structured document processing method, and a structured document processing program capable of searching for a document suitable for a purpose while suppressing search processing costs. The purpose is to provide.

One aspect of the present invention relates to a structured document processing apparatus, comprising: a structured document holding unit that holds a plurality of structured documents expressed in a tree structure; and the structured document held by the structured document holding unit. A character string frequency holding unit that holds a character string included in association with a character string frequency that is an appearance frequency of the character string in the plurality of structured documents, and the character string frequency holding unit holds the character string frequency Among the character strings, for a frequent character string having the character string frequency equal to or higher than a preset threshold, the frequent character string and a structure ID indicating a position where the frequent character string appears in the tree structure of the structured document. An index information holding unit that holds index information for identifying a character string that is to be specified by the index key in association with the index key, and the character string that the character string frequency holding unit holds U , A frequent character string extracting means for extracting a frequent character string having a character string frequency equal to or higher than a predetermined threshold, and a structure ID specifying for identifying the structure ID of the frequent character string extracted by the frequent character string extracting means The structure ID specifying means assigns group identification information for identifying a group including the plurality of structure IDs to a plurality of structure IDs that meet a predetermined condition in the frequent character string. The index information holding means holds the structure ID specified by the structure ID specifying means as the index key, and the frequent character string included in the group is used as the structure ID of the index key. It holds group identification information.

According to the present invention , it is possible to search for a suitable document while reducing the search processing cost even with respect to a frequent character string.

  Embodiments of a structured document processing apparatus, a structured document processing method, and a structured document processing program according to the present invention will be described below in detail with reference to the drawings. Note that the present invention is not limited to the embodiments.

  FIG. 1 is a block diagram illustrating an overall configuration of a structured document processing apparatus 10 according to the embodiment. The structured document processing apparatus 10 includes a structured document acquisition unit 102, a character string extraction unit 104, a character string frequency calculation unit 106, a frequent character string extraction unit 108, a structure ID specification unit 110, and a structure ID frequency calculation. Unit 112, index registration unit 114, structured document database (DB) 120, character string frequency DB 122, structure ID frequency DB 124, index DB 126, search condition acquisition unit 130, keyword character string extraction unit 132, , A search unit 134, a structured document extraction unit 136, and a structured document output unit 138.

  The structured document acquisition unit 102 acquires a structured document from the outside. Here, the structured document will be described. FIG. 2 is a diagram for explaining a structured document. Here, an example of a structured document in which XML is described as a structured document rule is shown. Note that the structured document is not limited to XML having a structured document convention, and may be another structured document such as SGML.

  As shown in FIG. 2, in XML, a “tag” is added to a sentence in the ground. In the example illustrated in FIG. 2, “<doc>”, “</ doc>”, “<title>”, “<p>”, and the like are tags. The document data is structured by the “tag”. In XML, a tag is expressed by enclosing a name representing the contents of the tag with “<” and “>”.

  A tag sandwiched between “<” and “>” is a start tag, and a tag sandwiched between “</” and “>” is an end tag. By sandwiching the text between a start tag and an end tag, the text is structured. Text between the start tag and the end tag is referred to as a text element.

In the example shown in FIG. 2, the text “XML database” is surrounded by “<title>” and “</ title>”. Thereby, a structure of “title” is given to this text.

  FIG. 3 is a diagram conceptually showing the structure of the structured document shown in FIG. In this way, the structured document shown in FIG. 2 is represented by a tree structure. Each text element divided by the tag is represented as a tree-structured node. Furthermore, each tag is also represented as a node. Each node is given an element ID for identifying each node.

  For example, an element ID “1” is assigned to the tag “doc”, and an element ID “2” is assigned to “title”. The element ID “3” is assigned to the text element “XML database”. Thus, the element ID is an ID for identifying each element in the structured document.

  FIG. 4 is a diagram for explaining the structure ID in the structured document. Each tag element and each text element are arranged at any position in the tree structure. The structure ID identifies this arrangement position. That is, the structure ID is information for identifying a route from “root”.

  For example, the structure ID “1” is assigned to the tag “doc”, and the structure ID “2” is assigned to “title”. In addition, the three “p (paragraph)” included in “chapter” have the same route from “root”. Therefore, the same structure ID “5” is assigned to these three “p”.

  Also, in the structure document other than the structure document shown in FIG. 4, it is assumed that “doc” exists as a branch of “root” and “title” exists as a branch of “doc” similarly to the structure document shown in FIG. . In this case, the same structure ID as the “title” in the structure document shown in FIG. 4, that is, the structure ID “2” is given to this “title”. As described above, the structure ID is the same value if the structural position in the structured document is the same regardless of the structured document.

  FIG. 5 is a diagram schematically showing the data structure of the structured document DB 120. As described above, the structured document DB 120 holds the structured document in association with the document ID for identifying the structured document.

  The character string extraction unit 104 in FIG. 1 extracts a character string from a structured document registered in the structured document DB 120 according to a preset condition. Then, the number of extracted character strings is registered in the character string frequency DB 122.

  FIG. 6 is a diagram for explaining a character string extracted by the character string extraction unit 104. FIG. 6 shows one text element. The character string is a partial character string included in the text element. For example, “data”, “base”, and the like are extracted as character strings. The preset condition is, for example, a condition that alphanumeric characters and katakana are extracted in units of three characters and hiragana is extracted in units of two characters.

  FIG. 7 is a diagram for explaining the character string extraction process. As described above, when extracting a character string in a predetermined character unit according to the condition, it may be extracted by shifting one character at a time. Thereby, omission of extraction of a character string can be prevented. Note that the method of extracting a character string is not limited to the present embodiment, and for example, a character string may be extracted in units of words using Japanese morphological analysis.

  The character string frequency calculation unit 106 illustrated in FIG. 1 calculates the appearance frequency of each character string specified by the character string extraction unit 104, that is, the character string frequency. Here, the character string frequency is the number of times a target character string appears in a structured document registered in the structured document DB 120. For example, when a plurality of structured documents are registered in the structured document DB 120, the number of times a target character string appears in all the plurality of structured documents.

  For example, the structured document shown in FIG. 2 includes four character strings “data”. In this case, the character string frequency is 4. The character string frequency calculation unit 106 further adds the character string frequencies of the character string “data” included in the other structured documents, and adds the character “data” in the structured document registered in the structured document DB 120. Calculate the string frequency of the column.

  The character string frequency calculation unit 106 registers the character string frequency of each character string in the character string frequency DB 122 in association with the character string.

  FIG. 8 is a diagram schematically showing the data configuration of the character string frequency DB 122. The character string extracted by the character string extraction unit 104 and the character string frequency calculated by the character string extraction unit 104 are stored in association with each character string.

  The frequent character string extraction unit 108 illustrated in FIG. 1 extracts a frequent character string based on the character string frequency registered in the character string frequency DB 122. Here, the frequent character string is a character string whose character string frequency is larger than a preset threshold value. For example, the threshold is set to “10000”. Thus, the threshold value is set as an absolute value.

  As another example, a relative value based on the number of character strings included in all structured documents registered in the structured document DB 120 may be set as a threshold value. For example, a ratio “0.8” of a predetermined number of character strings when the number of character strings included in all structured documents is 1 may be set as a relative threshold value.

  The structure ID specifying unit 110 specifies the arrangement position in the structured document, that is, the structure ID, of each frequent character string included in the structured document held in the structured document DB 120. Then, the structure ID is registered in the structure ID frequency DB 124 in association with the frequent character string.

  The structure ID frequency calculating unit 112 calculates the appearance frequency of the structure ID specified by the structure ID specifying unit 110, that is, the structure ID frequency. Here, the structure ID frequency is the number of predetermined frequent character strings arranged in a predetermined structure ID in the structure document registered in the structured document DB 120.

  For example, if “data” is a frequent character in the tree structure shown in FIG. 4, the structure ID frequency of the frequent character string “data” in the structure ID “5” is “3”. The structure ID frequency in the structure ID “3” is “1”. Furthermore, the structure ID frequency is calculated by adding the structure ID frequencies for the same structure ID in other structured documents.

  FIG. 9 is a diagram schematically illustrating the data configuration of the structure ID frequency DB 124. The frequent character string extracted by the frequent character string extraction unit 108 is associated with the structure ID in which the frequent character string appears. Further, the structure ID frequency is associated with the structure ID.

  The index registration unit 114 illustrated in FIG. 1 registers an index key and index information in the index DB 126 based on information registered in the character string frequency DB 122 and the structure ID frequency DB 124. Specifically, the character string extracted by the character string extraction unit 104 is registered as an index key. The index information includes the document ID of the structured document including the character string that is the index key, the element ID of the text element including the character string, and the offset. Here, the offset is information indicating the position of the character string in the text element. The offset corresponds to the in-element position information. For frequent character strings, a character string and a structure ID are registered as index keys, and index information is registered in association with each index key.

  In this way, the index information is information for specifying a character string. By using the index information, the text element including each character string, the position of the character string in the text element, and the structured document including the character string are specified. can do. Furthermore, for a frequent character string, the search processing cost in the frequent character string can be suppressed by using the character string and the structure ID as a search key.

  FIG. 10 is a diagram for explaining the offset. As shown in FIG. 10, in the text element “database...”, The offset position is set in order with the first character “de” as the offset position “1”. Therefore, for example, since the character string “data” is three characters starting from the first character in the text element, the offset is “1”. Since the first character “-” of the character string “Tabe” is the second character from the beginning, the offset of the character string “Tabe” is “2”.

  The index DB 126 holds an index table. More specifically, a document unit index table including a table for each document ID and a character string unit index table including all character strings regardless of the structured document are held.

  FIG. 11 is a diagram schematically showing the data configuration of the document unit index table 127 for each document ID held in the index DB 126. The document unit index table holds a character string as an index key and index information in association with each other. Further, the index information includes a document ID, an element ID, and an offset. In the document unit index table, even if the character strings are the same index key, if the document IDs are different, each character string is registered as an index key.

  FIG. 12 is a diagram schematically showing the data configuration of the character string unit index table 128 for each character string held in the index DB 126. Similarly to the document unit index table 127 described with reference to FIG. 11, the character string unit index table 128 holds character strings as index keys and index information in association with each other. The document unit index table 127 is a table in units of documents. In the character string unit index table 128, when the same character string exists in a plurality of structured documents, the search condition for any of these character strings is also set. Are associated with a predetermined character string as an index key. Therefore, a plurality of structured documents can be searched, and structured documents and text elements including specific character strings can be easily detected.

  In the character string unit index table 128, the frequent character strings are further divided in units of structure IDs, and the structure IDs are registered as index keys in addition to the character strings. Each index information is registered in association with the structure ID.

  Further, the index registering unit 114 sets a plurality of structure IDs as one structure ID group for the structure IDs of frequent character strings whose structure ID frequency is lower than a threshold value set in advance in the structure ID frequency DB 124. Then, the group ID for identifying the structure ID group is registered in the character string unit index table 128 as the structure ID. Thereby, even if there are many types of structure IDs, an increase in management cost due to subdivision can be suppressed.

  The search condition acquisition unit 130 illustrated in FIG. 1 acquires search conditions via an input / output device or the like. Here, the search condition includes a search keyword and a search range.

  The keyword character string extraction unit 132 extracts a keyword character string from the search condition acquired by the search condition acquisition unit 130. Here, the keyword character string is a character string included in the keyword. The process of extracting the character string is the same as the process of extracting the character string from the text element by the character string extraction unit 104.

  The search unit 134 searches the index DB 126 for an index key that matches the keyword character string extracted by the keyword character string extraction unit 132. The structured document extraction unit 136 extracts a structured document specified by the index information associated with the index key specified by the search unit 134 in the index DB 126 from the structured document DB 120. The structured document output unit 138 outputs the structured document extracted by the structured document extraction unit 136 as a search result.

  FIG. 13 is a flowchart showing index registration processing in the structured document processing apparatus 10. First, a user inputs a structured document using an input / output device such as a mouse, a keyboard, or a display monitor. The structured document acquisition unit 102 of the structured document processing apparatus 10 acquires the structured document input from the user in this way (step S100).

  Next, the structured document acquired by the structured document acquisition unit 102 is registered in the structured document DB 120 in association with the structured document ID (step S102). Next, the character string extraction unit 104 extracts a character string from each structured document registered in the structured document DB 120 (step S104). The extracted character string is registered in the document unit index table 127 together with index information as an index key. Next, the character string frequency calculation unit 106 calculates the character string frequency of each character string extracted by the character string extraction unit 104 (step S106). Each character string is associated with the character string frequency and registered in the character string frequency DB 122 (step S108).

  Next, the frequent character string extraction unit 108 extracts a frequent character string based on the character string frequency of the character string registered in the character string frequency DB 122 (step S110). Next, the structure ID specifying unit 110 specifies the structure ID of the frequent character string extracted by the frequent character string extracting unit 108 (step S112). Next, the structure ID frequency calculating unit 112 calculates the appearance frequency of the frequent character string for each structure ID specified by the structure ID specifying unit 110, that is, the structure ID frequency (step S114). The structure ID of the structure to which each frequent character string belongs and the structure ID are associated with each other and registered in the structure ID frequency DB 124 (step S116). Next, the index registration unit 114 registers the index key and the index information in the index DB 126 in association with each other based on the information registered in the character string frequency DB 122 and the structure ID frequency DB 124 (step S118). Specifically, the frequent character string and the structure ID are registered in the character string unit index table 128 together with the index information as an index key. This completes the index registration process.

  Here, the index registration process will be described more specifically. For example, assume that a character string and a character string frequency as shown in FIG. 8 are registered. Then, it is assumed that the threshold used for frequent character string determination is set to “10000”. In this case, “data” is extracted as a frequent character string. Further, the structure ID frequency for each structure ID of the frequent character string “data” included in all structured documents registered in the structured document DB 120 is registered in the structure ID frequency DB 124.

  Assume that the threshold value of the structure ID frequency that should include the structure ID as an index key is set to “500” in the character string unit index table 128. In this case, the structure IDs corresponding to the structure ID “6” and the structure ID “19” shown in FIG. 9 are integrated as the same structure ID group, and a group ID “G1” for identifying the structure ID group is given. As shown in FIG. 12, in the character string unit index table 128, the index information for the frequent character string “data” associated with the structure ID “6” and the structure ID “19” identifies the structure ID group. To be registered in association with an index key including “G1” and the character string “data”.

  For frequent character strings, frequent character strings can be distributed by assigning an index key to each structure ID. As a result, the search processing cost for designating the structure can be reduced.

  Furthermore, when the same character string (frequent character string) is divided for each structure ID, there is a possibility that problems such as a significant increase in the number of index keys may cause data storage efficiency degradation and system resource pressure. is there. Therefore, as described above, those with a small number of corresponding index information are collected into one group, so that it is possible to suppress an increase in the number of index keys while reducing the search processing cost for the frequent character strings as described above. .

  As another example, instead of grouping those whose structure ID frequency is less than or equal to the threshold value, those having similar paths in the tree structure of the structured document may be grouped. As another example, structure IDs having the same tag name may be integrated regardless of the route from “root”, and the total appearance frequency of each integrated structure ID may be stored as the structure ID frequency. Thus, the conditions for grouping are not limited to embodiment.

  FIG. 14 is a flowchart showing search processing in the structured document processing apparatus 10. First, a user inputs search conditions using an input / output device. The search condition acquisition unit 130 of the structured document processing apparatus 10 acquires the search condition input by the user in this way (step S200). Next, the keyword character string extraction unit 132 extracts a keyword character string from keywords included in the search condition acquired by the search condition acquisition unit 130 (step S202).

  Next, the search unit 134 searches the index DB 126 for a character string that matches the keyword character string extracted by the keyword character string extraction unit 132 (step S204). At this time, if a search range is specified in the search condition acquired by the search condition acquisition unit 130, the search is performed within the specified search range. If the search condition includes the designation of the structure ID, an index key matching the structure ID and the keyword character string is searched.

  The structured document extraction unit 136 extracts a corresponding structured document from the structured document DB 120 based on the index information associated with the index key detected by the search unit 134 in the index DB 126 (step S206). As another example, only the corresponding part of the structured document may be extracted. Next, the structured document output unit 138 outputs the structured document extracted by the structured document extraction unit 136 as a search result (step S208). This completes the search process.

  FIG. 15 is a diagram illustrating a hardware configuration of the structured document processing apparatus 10 according to the embodiment. The structured document processing apparatus 10 is structured in accordance with a ROM 52 storing a structured document processing program for executing an index registration process and a search process in the structured document processing apparatus 10 and a program in the ROM 52 as a hardware configuration. A CPU 51 that controls each part of the document processing apparatus 10, a RAM 53 that stores various data necessary for controlling the structured document processing apparatus 10, a communication I / F 57 that communicates by connecting to a network, and each part are connected. And a bus 62.

  The structured document processing program in the structured document processing apparatus 10 described above is a file in an installable or executable format and can be read by a computer such as a CD-ROM, floppy (R) disk (FD), or DVD. The program may be recorded on a recording medium.

  In this case, the structured document processing program is loaded on the main storage device by being read from the recording medium and executed by the structured document processing device 10, and each unit described in the software configuration is loaded on the main storage device. It is to be generated.

  Further, the structured document processing program of the present embodiment may be provided by being stored on a computer connected to a network such as the Internet and downloaded via the network.

  As described above, the present invention has been described using the embodiment, but various changes or improvements can be added to the above embodiment.

  As such a first modification, it is not necessary to perform processing for integrating a plurality of structure IDs into one group. In this case, the structured document processing apparatus 10 does not have to include the structure ID frequency calculation unit 112. And the process which calculates structure ID frequency and integrates into a group based on structure ID frequency is abbreviate | omitted.

  As a second modification, the structured document processing apparatus 10 may not hold the document unit index table 127 shown in FIG.

It is a block diagram which shows the whole structure of the structured document processing apparatus 10 concerning embodiment. It is a figure for demonstrating a structured document. It is a figure which shows notionally the structure of the structured document shown in FIG. It is a figure for demonstrating structure ID in a structured document. It is a figure which shows typically the data structure of structured document DB120. It is a figure for demonstrating the character string which the character string extraction part 104 extracts. It is a figure for demonstrating the process of character string extraction. It is a figure which shows typically the data structure of character string frequency DB122. It is a figure which shows typically the data structure of structure ID frequency DB124. It is a figure for demonstrating offset. It is a figure which shows typically the data structure of the document unit index table 127 for every document ID which index DB126 hold | maintains. It is a figure which shows typically the data structure of the character string unit index table 128 for every character string which index DB126 hold | maintains. 4 is a flowchart showing index registration processing in the structured document processing apparatus 10. 4 is a flowchart showing search processing in the structured document processing apparatus 10. It is a figure which shows the hardware constitutions of the structured document processing apparatus 10 concerning embodiment.

Explanation of symbols

10 Structured Document Processing Device 51 CPU
52 ROM
53 RAM
57 Communication I / F
62 Bus 102 Structured Document Acquisition Unit 104 Character String Extraction Unit 106 Character String Frequency Calculation Unit 108 Frequent Character String Extraction Unit 110 Structure ID Identification Unit 112 Structure ID Frequency Calculation Unit 114 Index Registration Unit 120 Structured Document DB
122 Character string frequency DB
124 Structure ID frequency DB
126 Index DB
127 Document Unit Index Table 128 Character String Unit Index Table 130 Search Condition Acquisition Unit 132 Keyword Character String Extraction Unit 134 Search Unit 136 Structured Document Extraction Unit 138 Structured Document Output Unit

Claims (4)

Structured document holding means for holding a plurality of structured documents expressed in a tree structure ;
A character string frequency holding unit that holds a character string included in the structured document held by the structured document holding unit and a character string frequency that is an appearance frequency of the character string in the plurality of structured documents in association with each other. When,
Among the character strings held in the character string frequency holding means, for the frequent character strings having the character string frequency equal to or higher than a preset threshold, the frequent character string and the tree structure of the structured document Index information holding means for holding index information for identifying a character string to be specified by the index key in association with the index key using a structure ID indicating a position where a frequent character string appears as an index key ;
Of the character strings held by the character string frequency holding means, a frequent character string extracting means for extracting a frequent character string having the character string frequency equal to or higher than a predetermined threshold;
And Structure ID specifying means for specifying the structure ID of the frequently appearing character string the frequently appearing character string extraction unit and extracted
With
The structural ID identifying means, among the frequently appearing character string, the plurality of structures ID matches the predetermined condition, to grant group identification information identifying a group including the plurality of structures ID,
The index information holding unit holds the structure ID specified by the structural ID specifying unit as said index key, for the frequently appearing character string included in the group, the group identified as the structural ID of the index key A structured document processing apparatus characterized by holding information. Of the character strings held by the character string frequency holding means, the structure ID of a frequent character string whose character string frequency is equal to or higher than a predetermined threshold, and the appearance frequency of the character string at the position indicated by the structure ID It further comprises a structure ID holding means for holding a certain structure ID frequency in association with each other,
The structure ID specifying means includes the group identification information for a plurality of structure IDs whose associated structure ID frequencies are equal to or less than a predetermined threshold among the frequent character strings held by the structure ID holding means. The structured document processing apparatus according to claim 1 , wherein: Computer
Structured document holding means, the structured document holding step of holding a plurality of structured documents represented in a tree structure,
String frequency holding unit, associates the character string which the structured document holding means is included in the structured document that holds, the string frequency is the appearance frequency in the plurality of structured documents of the string and string frequency holding step of holding,
Index information holding means, among the character strings stored in the character string frequency holding means, said character string frequency is preset threshold or more frequently appearing character string to a corresponding frequently appearing character string, the structured Index information for identifying a character string to be specified by the index key is registered in association with the index key using a structure ID indicating a position where the frequent character string appears in the tree structure of the document. An index creation step ;
A frequent character string extracting step, wherein the frequent character string extracting unit extracts a frequent character string having the character string frequency equal to or higher than a predetermined threshold among the character strings held by the character string frequency holding unit;
A structure ID specifying means for specifying the structure ID of the frequent character string extracted by the frequent character string extracting means;
Run
The structure ID specifying means assigns group identification information for identifying a group including the plurality of structure IDs to a plurality of structure IDs that meet a predetermined condition in the frequent character strings,
The index information holding unit holds the structure ID specified by the structure ID specifying unit as the index key, and the frequently identified character string included in the group has the group identification as the structure ID of the index key. A structured document processing method characterized by holding information . On the computer ,
Structured document holding means, the structured document holding step of holding a plurality of structured documents represented in a tree structure,
String frequency holding unit, associates the character string which the structured document holding means is included in the structured document that holds, the string frequency is the appearance frequency in the plurality of structured documents of the string and string frequency holding step of holding,
Index information holding means, among the character strings stored in the character string frequency holding means, said character string frequency is preset threshold or more frequently appearing character string to a corresponding frequently appearing character string, the structured Index information for identifying a character string to be specified by the index key is registered in association with the index key using a structure ID indicating a position where the frequent character string appears in the tree structure of the document. An index creation step;
A frequent character string extracting step, wherein the frequent character string extracting unit extracts a frequent character string having the character string frequency equal to or higher than a predetermined threshold among the character strings held by the character string frequency holding unit;
A structure ID specifying means for specifying the structure ID of the frequent character string extracted by the frequent character string extracting means;
And execute
The structure ID specifying unit assigns group identification information for identifying a group including the plurality of structure IDs to a plurality of structure IDs that meet a predetermined condition in the frequent character string,
The search information holding unit holds the structure ID specified by the structure ID specifying unit as the index key, and the frequently identified character string included in the group has the group identification as the structure ID of the index key. A structured document processing program for holding information .

Images