JP3627446B2 - Search formula creation device and search formula creation method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a search formula creation apparatus and a search formula creation method , and more particularly to a search formula creation apparatus and a search formula creation method for creating a search formula for document search.
[0002]
[Prior art]
In addition to using information such as keywords and classification codes controlled based on a thesaurus, the method for searching document data as needed from a large amount of document data is also a full text search that directly references the full text of the document. There is a search (full text search) method. Compared to the former method, the latter method has fewer search leaks.
[0003]
However, in the full-text search method, a large amount of documents become hits instead of few search leaks, and it takes a lot of labor to check the contents one by one. For example, when searching for documents with content related to “regulation of import / export of hazardous substances”, if the search condition is “includes the words“ toxic substances ”,“ import / export ”, and“ regulation ””, Documents with scattered words may also become hits.
[0004]
Therefore, in order to narrow down information more accurately by this full-text search method, a method using text information in the vicinity of a search term (search keyword) designated by a user has been proposed.
[0005]
For example, Japanese Patent Laid-Open No. 4-293161 discloses a “document search method and apparatus” that outputs a search result that matches a search condition when a searcher specifies two search keywords and their positional relationship.
[0006]
In this method, when extracting the document regarding the “regulation of import / export of hazardous substances” as shown in the above example, the word “hazardous substances” and the word “import / export” are included, and the interval is two words. The search condition can be “within and includes the word“ import / export ”and the word“ regulation ”, and the interval is within two words”. Therefore, it is possible to prevent hits of documents in which search keywords are scattered.
[0007]
This method is effective when the search request is clear, such as “regulation of import / export of hazardous substances”. However, in some cases, for example, only the search keyword “regulation” is revealed in the user's search request.
[0008]
Thus, when the search request is not clear and the user designates one search keyword, an “interactive database search method and apparatus” that extracts and presents text before and after the location where the search keyword appears is presented to the user. It is shown in Japanese Patent Laid-Open No. 5-189487.
[0009]
In this method, for example, if the search keyword “regulation” is specified, texts such as “regulation of import / export of hazardous substances”, “regulation of ~”, and “regulation related to” are enumerated, and the user is required among them. Information can be obtained selectively.
[0010]
[Problems to be solved by the invention]
However, the above method has a problem that a large amount of text is presented and the burden on the user's selection work increases. One reason for this is that even if the meaning is almost the same, all texts with different wordings are presented separately. For example, “exhaust gas regulations” and “exhaust gas regulations” correspond to this.
[0011]
The present invention has been made in view of the above points, and provides a search formula creation device and a search formula that efficiently perform information presentation for narrowing down a search request for a document and reduce a user's burden for creating the search formula. The purpose is to provide a creation method .
[0012]
[Means for Solving the Problems]
In the present invention, in order to solve the above-mentioned problem, in a search expression creating apparatus for creating a search expression for document search, document data holding means for holding a plurality of document data, and a search keyword for holding an input search keyword A holding means and a neighborhood word extraction condition in which the number of forward words extracted from the front of the search keyword and the number of backward words extracted from the back among the neighborhood words located in the vicinity of the search keyword are held. Based on the neighborhood word extraction condition holding means, the search keyword and the neighborhood word extraction condition, from the plurality of document data, the neighborhood word positioned within the number of forward words counted from the search keyword, and neighborhood word extraction means from the search keyword counting backwards for extracting said neighboring word positioned in the rear word in number, neighborhood word Li from the vicinity word A neighborhood word list generating means for generating a neighborhood word, a neighborhood word list holding means for holding the neighborhood word list, and a selection of neighborhood words included in the displayed neighborhood word list and receiving the selected neighborhood words There is provided a search expression creating apparatus comprising: a keyword adding means for adding as a search keyword; and a search expression creating means for creating a search expression from the search keyword.
[0013]
When a search formula for document search is created by such a search formula creation device, the document data holding means holds a plurality of document data. Further, the search keyword holding means holds the search keyword to be input. The neighborhood word extraction condition holding means holds an inputted neighborhood word extraction condition. The neighborhood word extraction means extracts neighborhood words from a plurality of document data based on the search keyword and the neighborhood word extraction condition. The neighborhood word list generation means generates a neighborhood word list from the neighborhood words. The neighborhood word list holding unit holds the generated neighborhood word list. The keyword adding means accepts selection of a neighborhood word included in the displayed neighborhood word list, and adds the selected neighborhood word as a search keyword. Then, the search formula creating means creates a search formula from the search keyword.
[0014]
When a search expression for document search is created in this way, information for narrowing down the search request, that is, the neighborhood word extracted based on the search keyword and the neighborhood word extraction condition is efficiently presented. Can be added as a search keyword for creating a search expression, so that the burden on the user for creating the search expression can be reduced.
[0015]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a block diagram showing the principle configuration of a retrieval formula creation apparatus according to the present invention.
[0016]
The search expression creation apparatus of the present invention includes document data holding means 1, search keyword holding means 2, neighborhood word extraction condition holding means 3, neighborhood word extraction means 4, neighborhood word list generation means 5, and neighborhood word list. The holding unit 6, the keyword adding unit 7, and the search expression creating unit 8 are configured.
[0017]
Here, the document data holding means 1 holds a plurality of document data. The search keyword holding unit 2 holds the input search keyword. The neighborhood word extraction condition holding unit 3 holds an inputted neighborhood word extraction condition. The neighborhood word extraction unit 4 extracts neighborhood words from the document data based on the search keyword and the neighborhood word extraction condition. The neighborhood word list generation means 5 generates a neighborhood word list. The neighborhood word list holding unit 6 holds the generated neighborhood word list. The keyword adding means 7 adds a neighborhood word selected from the neighborhood word list as a search keyword. Then, the search formula creating means 8 creates a search formula from the search keyword.
[0018]
When a search formula for document search is created by such a search formula creation device, the search keyword input by the user is held in the search keyword holding means 2. That is, if the input search keyword is “aaaa”, the search keyword holding means 2 holds this. Similarly, the neighborhood word extraction condition input by the user is held in the neighborhood word extraction condition holding means 3.
[0019]
When the neighborhood word extraction unit 4 receives an instruction from the user, the neighborhood word extraction unit 4 extracts a neighborhood word from the document data held in the document data holding unit 1 based on the search keyword and the neighborhood word extraction condition. For example, when “aaaa, bbbb,...” And “aaaa, cccc,...” Are stored as document data including the search keyword “aaaa”, it is confirmed that the neighborhood word extraction condition is met. , “Bbbb” and “cccc” are extracted as neighborhood words.
[0020]
At this time, the ID or the like of the document data as the extraction source is associated with the extracted neighboring word, but the same neighboring word may be extracted from a plurality of document data, and the relationship between the neighboring word and the extracting source document data Is not necessarily 1: 1. Note that the neighborhood word in this specification refers to an independent word that exists in the vicinity of the search keyword (the range is determined by the neighborhood word extraction condition) and has meaning alone.
[0021]
The neighborhood word list generation means 5 generates a neighborhood word list with the extracted neighborhood words. That is, if the extracted neighborhood words are “bbbb” and “cccc”, they are listed. The neighborhood word list holding unit 6 holds the generated neighborhood word list. When the user who sees the displayed neighborhood word list selects an arbitrary neighborhood word in the list, the keyword adding means 7 accepts this selection and adds the selected neighborhood word as a search keyword to the search keyword holding means 2. To do. If the user selects the neighborhood word “bbbb”, this word is added as a new search keyword.
[0022]
Then, upon receiving a user instruction, the search formula creation unit 8 creates a search formula from the search keyword held in the search keyword holding unit 2. That is, if “aaa” and “bbbb” are held as search keywords at this time, a search expression is created from these words.
[0023]
When a search expression is created in this way, information for narrowing down the search request, that is, neighboring words are efficiently presented, and this neighboring word can be added as a search keyword for creating a search expression. It is possible to reduce the burden on the user.
[0024]
Here, the configuration and operation in the case where the retrieval formula creation apparatus of the present invention is implemented in combination with a document retrieval apparatus will be described.
FIG. 2 is a hardware configuration diagram of a document search apparatus to which the search expression creation apparatus of the present invention is applied.
[0025]
The document search apparatus 10 to which the search expression creation apparatus of the present invention is applied has a document information storage unit 20, receives input from the user via the input device 30, and outputs information to the user via the output device 40. Here, as the input device 30, a pointing device such as a mouse, a keyboard, or the like is applicable. As the output device 40, a monitor, a printer or the like can be applied.
[0026]
The document search apparatus 10 includes a text search processing unit 11, a search result storage unit 12, a neighborhood word extraction processing unit 13, a neighborhood word storage unit 14, a neighborhood word detection processing unit 15 in the same sentence, and a part of speech classification processing unit. 16.
[0027]
Here, the text search processing unit 11 is configured by a CPU or the like, and searches for document information from the document information storage unit 20 based on a search keyword or a search expression input via the input device 30. In the search, a known technique such as a full text search is applied.
[0028]
The search result storage unit 12 includes a memory or the like, and stores the search result of the text search processing unit 11. That is, the sentence number of the sentence that matches the search keyword and the search expression, the ID of the document including the sentence, and the like are stored in association with each other. The search result stored here is output via the output device 40.
[0029]
The neighborhood word extraction processing unit 13 is composed of a CPU or the like, and extracts neighborhood words from the search result document data stored in the search result storage unit 12 based on the neighborhood word extraction conditions input via the input device 30. To do. In order to extract a neighborhood word, a morphological analysis is performed on a sentence that matches a search keyword or a search expression, and a word that meets the neighborhood word extraction condition may be detected.
[0030]
The neighborhood word extraction condition is input in a format such as “[m] search keyword [n]”. This indicates that m independent words that have a meaning independently and n independent words that have a meaning alone and that are located behind the search keyword are extracted as neighborhood words. . When “search keyword 1 * [m] search keyword 2 [n]” (* is a logical product) is input, a document including both search keyword 1 and search keyword 2 is targeted as “[ m] Neighboring words satisfying the search keyword 2 [n] ”may be extracted.
[0031]
The morphological analysis performed at this time is a technique widely known as a basic technique of natural language processing, and is detailed in “Basic techniques of natural language processing” (written by Hirogo Nomura, The Institute of Electronics, Information and Communication Engineers, 1988) and the like.
[0032]
The neighborhood word storage unit 14 includes a memory or the like, and stores the neighborhood words extracted by the neighborhood word extraction processing unit 13. The neighborhood word is stored in association with the sentence number of the original sentence from which the neighborhood word is extracted, the document ID, and the like. Further, the neighborhood words stored here are input to the part of speech classification processing unit 16.
[0033]
The same sentence neighboring word detection processing unit 15 includes the added search keyword when the user selects an arbitrary neighboring word from the list of neighboring words output via the output device 40 and adds the search keyword. A sentence is searched in the neighborhood word storage unit 14, and a neighborhood word included in the sentence is detected. The detected neighborhood word is input to the part-of-speech classification processing unit 16.
[0034]
The part-of-speech classification processing unit 16 classifies the input neighborhood words into parts of speech, that is, nouns, verbs, etc., and generates a neighborhood word list. The generated neighborhood word list is output via the output device 40.
[0035]
Next, a procedure for performing a document search with the document search apparatus 10 will be described.
FIG. 3 is a flowchart showing a procedure for searching for a document by a document search apparatus to which the search expression creation apparatus of the present invention is applied. Hereinafter, it demonstrates along a step number.
[S1] The text search processing unit 11 receives a search keyword input via the input device 30.
[S2] The text search processing unit 11 searches the document information storage unit 20 with the received search keyword, and uses the sentence number of the sentence including the search keyword and the ID of the document data including the sentence as a search result. The data is stored in the storage unit 12.
[S3] The neighborhood word extraction processing unit 13 receives a neighborhood word extraction condition input via the input device 30.
[S4] The neighborhood word extraction processing unit 13 selects one sentence including the input search keyword or search expression, and morphologically analyzes the sentence. In this morphological analysis, if nouns appear continuously, that part is treated as a single compound word. Thus, for example, when the word “morpheme” and the word “analysis” appear in succession, the word “morpheme analysis” appears.
[S5] The neighborhood word extraction processing unit 13 extracts neighborhood words from the morphological-analyzed sentence based on the search keyword and the neighborhood word extraction condition, and determines the sentence number of the sentence and the ID of the document including the sentence It is stored in the neighborhood word storage unit 14 in association with the word.
[S6] The neighborhood word extraction processing unit 13 determines whether there is document data in the search result storage unit 12 that has not undergone the processing of steps S4 to S5. If there is unprocessed document data, the process proceeds to step S4 again to select an unprocessed sentence. If there is no unprocessed document data, the process proceeds to step S7.
[S7] The neighborhood word storage unit 14 inputs all the stored neighborhood words to the part of speech classification processing unit 16. The part-of-speech classification processing unit 16 classifies the neighborhood words and generates a neighborhood word list.
[S8] The part of speech classification processing unit 16 outputs the generated neighborhood word list via the output device 40.
[S9] The neighborhood word detection processing unit 15 in the same sentence determines whether or not a search keyword to be added has been selected from the user via the input device 30. Here, one of the neighborhood words included in the neighborhood word list displayed in step S8 is selected. If the search keyword to be added is selected, the process proceeds to step S10. If no additional search keyword is selected, the process proceeds to step S11.
[S10] Since there are a plurality of search keywords, a search expression is created. Details of this will be described later. After the search expression is created, the process proceeds again to step S4.
[S11] If there is a search expression created by the processes so far, the search expression is input to the text search processing unit 11. The text search processing unit 11 searches the document information storage unit 20 based on the search formula. The search result is stored in the search result storage unit 12. Note that this processing may be omitted when a search expression is not created.
[S12] The search result storage unit 12 outputs the stored search results via the output device 40.
[0036]
Here, the creation of the search expression performed in step S10 will be described. When a display device such as a monitor is applied as the output device 40, a list of neighboring words is displayed on this display device at this time.
[0037]
FIG. 4 is a diagram illustrating a display screen example of the neighborhood word list.
On the display screen 50, a neighborhood word list 51, a start button 56, and an end button 57 are displayed.
[0038]
The user selects the start button 56 with a pointing device such as a mouse when adding a search keyword and detecting a nearby word in the same sentence. Further, when there is no need to add a search keyword, the end button 57 is selected.
[0039]
The neighborhood word list 51 includes a neighborhood word extraction condition display field 52, a neighborhood word <noun> display field 53, a neighborhood word <verb> display field 54, and a neighborhood word <other> display field 55. . The neighborhood word <noun> display field 53 and the neighborhood word <verb> display field 54 are each accompanied by a scroll bar.
[0040]
In the neighborhood word extraction condition display field 52, the inputted neighborhood word extraction condition is displayed. In the neighborhood word <noun> display field 53, nouns are displayed among the extracted neighborhood words. Of the extracted neighborhood words, verbs are displayed in the neighborhood word <verb> display field 54. In the neighborhood word <other> display field 54, the extracted neighborhood words that are neither noun nor verb are displayed.
[0041]
Here, an example is shown in which a document search is performed with the search keyword as “restriction” and the neighborhood word extraction condition as “[2] restriction”. This is equivalent to the neighborhood word extraction condition “[2] restriction [0]”.
[0042]
When the search keyword “restriction” is input via the input device 30 shown in FIG. 2, the text search processing unit 11 sets the search keyword “restriction” for the document data stored in the document information storage unit 20. Based on the document search. This search result is stored in the search result storage unit 12.
[0043]
The neighborhood word extraction processing unit 13 that has received the neighborhood word extraction condition “[2] restriction” via the input device 30 performs morphological analysis on the document data of the search result stored in the search result storage unit 12. Each time the search keyword “regulation” appears, two words that appear in front of the search keyword “regulation” and have meaning alone are extracted and set as neighborhood words. The extracted neighborhood words are stored in the neighborhood word storage unit 14 and are classified by the part of speech classification processing unit 16 to generate a neighborhood word list 51.
[0044]
Thus, by classifying the extracted neighborhood words by part of speech, the user can easily find an appropriate word when adding a search keyword.
Although the case where only the word that appears in front of the search keyword is used as the neighborhood word has been described here, the word that appears behind the search keyword may be extracted as the neighborhood word depending on the neighborhood word extraction condition. In that case, the neighborhood word that appears in the front and the neighborhood word that appears in the back may be displayed separately.
[0045]
Although the neighborhood words are displayed separately for each part of speech here, they can be displayed without being divided.
Next, a case where the user adds a search keyword will be described.
[0046]
If the user is interested in “regulations on transfer, transaction, import / export”, the user designates the neighborhood words “transfer”, “transaction”, “import / export” with a pointing device such as a mouse, and then selects the start button 56. To do.
[0047]
FIG. 5 is a diagram illustrating a state in which a search keyword to be added is designated and a nearby word in the same sentence is detected.
A neighborhood word list 51a is displayed on the display screen 50a.
[0048]
When the neighborhood word storage unit 14 previously extracts neighborhood words based on the neighborhood word extraction condition “[2] restriction”, each neighborhood word is associated with the ID of the original document from which the neighborhood word is extracted. Since it is stored, when a search keyword is added, the neighborhood word detection processing unit 15 in the same sentence detects the neighborhood word existing in the same sentence from the neighborhood word storage unit 14.
[0049]
Therefore, if "transfer", "transaction", and "import / export" are additionally specified as search keywords, documents that contain the word "regulation" and also contain the words "transfer" or "transaction" or "import / export". , Two words that appear in front of the word “regulation” and have meaning alone are detected as neighboring words and displayed in the neighboring word list 51a. The number of neighboring words presented in the neighboring word list 51 a is smaller than the number of neighboring words presented in the neighboring word list 51.
[0050]
If the user is further interested in the “Rules for Transfer, Trading, Import / Export of Wildlife and Rare Wild Animals”, the user specifies the neighborhood words “wildlife” and “rare wildlife” and then starts. Button 56 is selected.
[0051]
FIG. 6 is a diagram illustrating a state in which a search keyword to be added is designated and a nearby word in the same sentence is detected.
A neighborhood word list 51b is displayed on the display screen 50b.
[0052]
If "wildlife" or "rare wildlife" is additionally specified as a search keyword, it will contain the word "regulation", the word "transfer" or "transaction" or "import / export", and "wildlife" Alternatively, from a document including the word “rare wildlife”, two words that appear before the word “regulation” and have meaning alone are detected as neighboring words and displayed in the neighboring word list 51b. The number of neighboring words presented in the neighboring word list 51b is further smaller than the number of neighboring words presented in the neighboring word list 51a.
[0053]
Therefore, at this time, the document information storage unit 20 stores the search expression “regulation * (transfer + transaction + import / export) * (wildlife + rare wildlife)” (* indicates a logical product and + indicates a logical sum). It can be seen that documents satisfying the above are stored. When the user determines that the number of neighboring words has sufficiently decreased in this way, the user may create a search expression from the displayed neighboring word extraction conditions.
[0054]
This search formula deletes [m] and [n] of the character string displayed in the neighborhood word extraction condition display fields 52, 52a, 52b..., “:” Is changed to “*”, “,” Can be created by converting the symbol to “+”, the CPU or the like may process it, or the user may input it.
[0055]
As described above, in the document search apparatus 10 to which the search expression creation apparatus of the present invention is applied, the neighborhood words are extracted and displayed for the search keyword, and the neighborhood words that the user is interested in can be narrowed down and specified one after another. It is possible to create an appropriate search expression without requiring complicated operations.
[0056]
In addition, as described above, neighboring words can be selected step by step, so that the search keywords to be input can be narrowed down from those having a broad concept to those having a narrow concept sequentially. Therefore, even when the user himself does not grasp the search intention, the required document can be searched.
[0057]
In the example shown in FIG. 4 to FIG. 6, the neighborhood word “import / export” disappears when moving from FIG. 5 to FIG. 6. This indicates that there is no document in which the word “import / export” is compatible with the words “wildlife” or “rare wildlife” as a neighborhood word. Therefore, the search formula in this case may be “regulation * (transfer + transaction) * (wildlife + rare wildlife)”.
[0058]
【The invention's effect】
As described above, the search expression creating apparatus of the present invention efficiently presents information for narrowing a search request, that is, a neighborhood word extracted based on a search keyword and a neighborhood word extraction condition, and searches for this neighborhood word. Since it can be added as a search keyword for creating a formula, it is possible to reduce the burden on the user for creating the search formula.
[Brief description of the drawings]
FIG. 1 is a block diagram showing the principle configuration of a retrieval formula creation apparatus of the present invention.
FIG. 2 is a hardware configuration diagram of a document search apparatus to which the search expression creation apparatus of the present invention is applied.
FIG. 3 is a flowchart showing a procedure for searching for a document by a document search apparatus to which the search expression creation apparatus of the present invention is applied.
FIG. 4 is a diagram illustrating a display screen example of a neighborhood word list.
FIG. 5 is a diagram (No. 1) showing a state in which a search keyword to be added is designated and a neighborhood word in the same sentence is detected.
FIG. 6 is a diagram (part 2) illustrating a state in which a search keyword to be added is specified and a nearby word in the same sentence is detected.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 Document data holding means 2 Search keyword holding means 3 Neighborhood word extraction condition holding means 4 Neighborhood word extraction means 5 Neighborhood word list generation means 6 Neighborhood word list holding means 7 Keyword addition means 8 Search expression creation means

Claims (4)

In a search expression creation device for creating a search expression for document search,
Document data holding means for holding a plurality of document data;
Search keyword holding means for holding the input search keyword,
Among the neighborhood words located in the vicinity of the search keyword, the neighborhood word extraction condition that holds the neighborhood word extraction condition in which the number of forward words extracted from the front of the search keyword and the number of backward words extracted from the back are individually specified Holding means;
Based on the search keyword and the neighborhood word extraction condition, from the plurality of document data, the neighborhood word positioned within the number of forward words counted forward from the search keyword, and counted backward from the search keyword Neighboring word extraction means for extracting the neighboring words located within the number of backward words ;
Neighborhood word list generating means for generating a neighborhood word list from the neighborhood words;
Neighborhood word list holding means for holding the neighborhood word list;
A keyword adding means for accepting selection of a neighboring word included in the displayed neighboring word list and adding the selected neighboring word as the search keyword;
Search expression creating means for creating a search expression from the search keyword;
A search expression creation device characterized by comprising: In a document search device that searches a document from a plurality of document data,
  Document data holding means for holding the plurality of document data;
  Search keyword holding means for holding the input search keyword,
  Among neighboring words located in the vicinity of the search keyword, a neighboring word extraction condition that holds a neighboring word extraction condition in which the number of forward words extracted from the front of the search keyword and the number of backward words extracted from the back are individually specified. Holding means;
  Based on the search keyword and the neighborhood word extraction condition, from the plurality of document data, the neighborhood word positioned within the number of forward words counted forward from the search keyword, and counted backward from the search keyword Neighboring word extraction means for extracting the neighboring words located within the number of backward words;
  Neighborhood word list generating means for generating a neighborhood word list from the neighborhood words;
  Neighborhood word list holding means for holding the neighborhood word list;
  A keyword adding means for accepting selection of a neighborhood word included in the displayed neighborhood word list and adding the selected neighborhood word as the search keyword;
  Search expression creating means for creating a search expression from the search keyword;
  Search means for searching for a document from the plurality of document data using the search formula;
  A document search apparatus characterized by comprising: In a search expression creation method for creating a search expression for document search,
  A document data holding unit holding a plurality of document data;
  A search keyword holding means for holding the input search keyword;
  The neighborhood word extraction condition holding means is a neighborhood word extraction in which the number of forward words extracted from the front of the search keyword and the number of backward words extracted from the back are individually specified from the neighborhood words located in the vicinity of the search keyword Maintaining a condition;
  Based on the search keyword and the neighborhood word extraction condition, a neighborhood word extraction unit counts the neighborhood words that are located within the number of forward words from the plurality of document data, counting forward from the search keyword, and the search keyword Extracting the neighborhood words located in the backward word count counting backward from
  A neighborhood word list generating means generating a neighborhood word list from the neighborhood words;
  A neighborhood word list holding means holding the neighborhood word list;
  A keyword adding unit that accepts selection of a nearby word included in the displayed neighboring word list and adds the selected neighboring word as the search keyword;
  A search formula creating means creating a search formula from the search keyword;
  A search expression creating method characterized by comprising: In a document retrieval method for retrieving a document from a plurality of document data,
  Document data holding means holding the plurality of document data;
  A search keyword holding means for holding the input search keyword;
  The neighborhood word extraction condition holding means is a neighborhood word extraction in which the number of forward words extracted from the front of the search keyword and the number of backward words extracted from the back are individually specified from the neighborhood words located in the vicinity of the search keyword Maintaining a condition;
  Based on the search keyword and the neighborhood word extraction condition, a neighborhood word extraction unit counts the neighborhood words that are located within the number of forward words from the plurality of document data, counting forward from the search keyword, and the search keyword Extracting the neighborhood words located in the backward word count counting backward from
  A neighborhood word list generating means generating a neighborhood word list from the neighborhood words;
  A neighborhood word list holding unit holding the neighborhood word list;
  A keyword adding unit that accepts selection of a nearby word included in the displayed neighboring word list and adds the selected neighboring word as the search keyword;
  A search expression creating means creating a search expression from the search keyword;
  A search means for searching for a document from the document data using the search formula;
  A document search method characterized by comprising:

Images