JP2007316743A - Partial document retrieving program, method and device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a document retrieving device easily reusing documents. <P>SOLUTION: A partial document extracting portion 2 extracts continuous paragraphs having similar appearance frequency of a word, a paragraph having a form of explaining words, a range which is frequently reused more than a set value, continuous pages including continuous numbers at the end of a title, continuous pages of similar screen configuration of the pages, and continuous pages of similar color configuration of the pages from the document input from a document acquiring portion 1 as the partial document, and the information is saved in a storing device 7 by a partial document saving portion 3. A partial document retrieving portion 4 retrieves the partial document information of the storing device 7 according to retrieval conditions input to a retrieval condition input portion 5, and outputs the partial documents satisfying the retrieval conditions from a retrieval result output portion 6. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a partial document search apparatus that searches a part that is likely to be used for creating a new document from documents that have been created and stored, and that can be easily reused.

  Conventionally, many documents have been created and accumulated in offices and the like. When creating a new document, the existing document is partially modified or reused by extracting a part from the existing document and using it.

  As a result, the burden of creating a new document is reduced, and at the same time, the quality of the document can be improved by creating a document based on a good quality document.

  In order to make it easier to find the target document from the existing documents that have been stored, words are extracted from a subset of documents that have been extracted from the stored documents based on attribute information, search expressions, etc., and their appearance It has been proposed to obtain a frequency, rank words based on the appearance frequency, and perform a search using a part of the word as a related keyword as a search condition (see, for example, Patent Document 1).

In addition, thumbnail images of objects such as diagrams and tables in the document and position information in the document are saved, and a list of thumbnail images of objects corresponding to the type of object such as tables and figures specified by the user is displayed. It has been proposed to display and reuse an object selected from this list (see, for example, Patent Document 2).
Japanese Patent Laid-Open No. 11-25108 JP 2001-273314 A

  When reusing a document, the entire document is not necessarily required, but only a part thereof is often required. However, in the case of the one described in Patent Document 1, it is necessary to search again from the searched document to find a necessary portion, and to copy the portion and paste it on a new document. It was.

  Moreover, in the thing as described in Patent Document 2, the text is not a target, and in order to reuse the text portion, it must be as in Patent Document 1.

  Also, if you want to limit the length of something that you want to reuse, such as a sentence that represents a certain matter in several lines, or a presentation material that represents a certain matter in about three pages, The length of no one knows, so we had to check the target part of the retrieved document.

  SUMMARY OF THE INVENTION An object of the present invention is to provide a document search apparatus that can easily reuse a document.

  The invention that solves the above-described problem is to extract a related description range from a document as a partial document, store the information of the extracted partial document, and match a search condition input from the stored information The partial document to be searched is retrieved and output.

  In the present invention, a partial document that is a part of the document is extracted from the document and stored, and a document that matches the search condition is retrieved from the partial document and output. Therefore, the retrieved partial document can be used as it is.

  Here, continuous paragraphs or pages having similar word frequencies are defined as the related description range.

  In this way, continuous paragraphs or pages with similar word frequencies are used as partial documents, and the related description ranges can be extracted well.

  In addition, paragraphs or pages that are in the form of explaining words are included in the scope of the related description.

  In this way, a paragraph or page in a format that explains words is regarded as a partial document, and a related description range can be satisfactorily extracted.

  Further, a range in which the number of times of reuse is greater than the set number of times is set as the range of the related description.

  In this way, a range in which the number of times of reuse is greater than the set number of times is set as a partial document, and a related description range can be extracted satisfactorily.

  In addition, in a document in which a title is assigned to each page, consecutive pages related to the content of the title are set as the range of the related description.

  In this way, continuous pages related to the contents of the title are made partial documents, and the related description range can be extracted well.

  Further, continuous pages having similar page screen configurations are set as the related description range.

  In this way, continuous pages with similar screen configurations are used as partial documents, and the related description ranges can be extracted well.

  Further, continuous pages having similar page color configurations are set as the related description range.

  In this way, continuous pages with similar page color configurations are used as partial documents, and the related description ranges can be satisfactorily extracted.

  Further, the length of the partial document is stored as the partial document information, and the length of the partial document is input as the search condition.

  In this way, a partial document is searched under a search condition including the length of the partial document, and a partial document having a required length can be searched.

  According to the present invention, a partial document that is a part of the document is extracted from the document and stored, and a document that matches the search condition is retrieved from the partial document and output. Therefore, the retrieved partial document can be used as it is, and the document can be easily reused.

  In addition, as the range of related descriptions, continuous paragraphs or pages with similar word frequencies, paragraphs or pages in a format that explains words, and ranges where the number of reuses is greater than the set number of times Or, in a document with a title on each page, a continuous page related to the content of the title, a continuous page with a similar page screen configuration, or a continuous page with a similar page color configuration. If it is a partial document, the range of related description can be extracted well, and the document can be easily reused.

  In addition, since the length of the partial document is stored as the partial document information and can be searched based on this length, the partial document having the required length can be searched, and the document can be easily reused. .

The present invention will be described below with reference to the drawings.
FIG. 1 is a diagram showing a partial document search apparatus according to an embodiment of the present invention.

  In FIG. 1, a partial document search apparatus according to this embodiment includes a document acquisition unit 1 that acquires a document to be searched for reuse, and a partial document extraction unit 2 that extracts a part from the acquired document as a partial document. The partial document information storage unit 3 that stores the extracted partial document information in the storage device 7, the partial document search unit 4 that searches the partial document according to the input search conditions, and the search that inputs the conditions for searching the partial documents A condition input unit 5 and a search result output unit 6 that outputs information of the searched partial documents are provided.

  In such a partial document search apparatus, in response to an instruction from a user, the document acquisition unit 1 acquires a document to be searched for reuse, extracts the partial document by the partial document extraction unit 2, and extracts the partial document. The information storage unit 3 stores the partial document information in the storage device 7.

  Specifically, as shown in the flowchart of FIG. 2, the document acquisition unit 1 is registered in all documents stored in the computer, documents in a certain folder, or a document management system instructed by the user. One document is acquired from the document such as the existing document, and is input to the partial document extraction unit 2 (S11).

  When the document is input, the partial document extraction unit 2 extracts the partial document based on the similarity of contents (S12).

  In this case, when a document is composed of a plurality of topics, a range of similar contents related to each topic is extracted as a partial document. A method based on the appearance frequency of words is used as a method of discriminating a range of similar contents related to a topic.

  Create a dictionary in which words related to the genre that you want to extract as topics in advance are classified for each genre, extract words in the dictionary created from each paragraph (each page in the case of a document for presentation) of each document, Or, each page) is acquired from which genre of words.

  Next, the composition of the genre in the preceding and following paragraphs is compared, and if they are similar, it is determined that the same topic is described, and a partial document is obtained.

  As a similar determination method, since the appearance frequency of words of each genre can be regarded as a vector, the appearance frequency of words of each genre in each paragraph is set as a vector, the distance between the vectors is calculated for each genre, If they are below a preset threshold, they are considered to be similar.

  The Euclidean distance can be used to calculate the distance between vectors. The Euclidean distance between two vectors u and v of dimension number n can be calculated by the following equation.

  Further, the LSI (Latent Semantic Indexing) method can be used to determine the similarity.

The LSI method is a search method using singular value decomposition. When the number of words is t and the number of documents is d, and H _td is the expression frequency of words in all documents as a matrix, H _td = U _tr D _rr V _rd by singular value decomposition.
It is decomposed as the product of three matrices. Here, D _rr is a diagonal matrix, and the dimension number r is the rank of the matrix.

Here, V _rd will rank × document number of the matrix. Each column vector becomes a feature of each document. The smaller the distance between the vectors, the more similar the documents can be considered.

  That is, for words of each genre, vectors are obtained from the appearance frequency of words in each paragraph (or each page), and are considered similar if the distance between the vectors is equal to or less than a preset threshold. The Euclidean distance described above can be used as the distance between the vectors.

  In addition, it is possible to determine in advance which genre each paragraph (or each page) belongs to, and if paragraphs (or pages) of the same genre continue, they may be combined into a partial document.

  In this case, the appearance frequency of a word in the dictionary classified by genre is acquired from each paragraph (or each page), and if the appearance frequency of the word for each genre is equal to or higher than a preset threshold, the paragraph (or Page) belongs to the genre. If the appearance frequency of words of any genre does not exceed the threshold, it is not regarded as a partial document.

  As an example, a case will be described in which partial documents are extracted from the following documents with content similarity. The basic operation is possible not only for text-only documents but also for text-image mixed documents.

“Our laboratory is developing multimedia search technology as a method for extracting useful information from a large amount of multimedia information such as images, video, and sound. You can search for the information you want by placing information, overviewing the space and walking through it.
We are also developing robot technology that supports people at home and in the office. Real space can be recognized based on information obtained from various sensors, and stable bipedal walking can be performed with advanced posture control algorithms. "

  This document contains two different topics: multimedia search and robots.

  In the first paragraph, words registered in the genre of “multimedia information” such as “image”, “video”, “sound”, “multimedia information” appear, and this paragraph is determined to be a topic related to multimedia information Is done.

  In the next paragraph, words registered in the genre of “robot” such as “robot”, “posture control”, “biped walking” appear, and this paragraph is determined as a topic related to robots. , Because the topic of each paragraph is different, it is a separate partial document.

  Next, the partial document extraction unit 2 extracts a partial document based on the explanatory text (S13).

  When a paragraph (the page in the case of a presentation document) has a format that explains a specific keyword (thing) such as “to”, the paragraph is set as one partial document.

  Also, there may be a case where the description is not completed in one paragraph, such as when describing the keyword in more detail. For this reason, it is determined whether or not the explanation is continued in the following paragraphs by the following method, and if it is determined that the explanation is continued, they are set as one partial document.

  If the words appearing in the subsequent paragraph (or page) and the previous paragraph (or page) are similar, it is determined that the explanation continues.

  The similarity is determined by using the word appearance frequency of each paragraph as a vector and similar when the Euclidean distance of each vector is equal to or less than a threshold, or by determining the number of dimensions of the word appearance frequency vector by the LSI method described above. It is assumed that it is similar to the case where the Euclidean distance is equal to or smaller than the threshold value.

  If the first keyword is included in the subsequent paragraph (or page), the threshold value may be increased.

  As an example, a case will be described in which a partial document is extracted from the following document using an explanatory text. The basic operation is possible not only for text-only documents but also for text-image mixed documents.

“Multimedia search uses a technique called self-organizing map for image layout.
The self-organizing map is an unsupervised learning method based on a neural network, and can be arranged in a two-dimensional space while keeping the distribution of the original multidimensional vector space as much as possible. By repeating learning, a portion having a high density in the original vector space is arranged in a wide range of the two-dimensional space, and a portion having a low density in the vector space is arranged in a narrow range of the two-dimensional space.
By using the self-organizing map, it is possible to grasp the rough density distribution of the original multidimensional vector space on the two-dimensional space. It is also possible to make it easier to find similar objects by arranging them as close as possible.
Next, in multimedia search, the user walks through a virtual space where an image is displayed to find desired information. In addition to simply walking through, it is also possible to display an image matching the keyword prominently in the foreground using a keyword or the like. "

  In this document, the keyword “self-organizing map” is explained in the second paragraph. In this paragraph, words such as “placement”, “vector space”, “two-dimensional space”, “distribution”, “density” appear. In the following paragraph, not only the keyword “self-organizing map”, which is the source of the explanation, but also “placement”, “vector space”, “two-dimensional space”, “distribution”, “density” Words such as appear as well. Thus, it is determined that the two paragraphs are continuous and explain the “self-organizing map”.

  In the next paragraph, the above-mentioned word does not appear in turn, and words such as “display”, “walk-through”, and “keyword” appear. Therefore, it is determined that this paragraph is not continuous, An organized map is extracted from “from” to “can be easily searched by arranging” as one partial document.

  Next, the partial document extraction unit 2 extracts partial documents based on the reuse frequency (S14).

  When a user has created a document for a certain period of time, or when a plurality of users are creating a document with the same theme, the user may reuse the same range many times. A range in which the number of times of reuse is large is set as a partial document.

  In this case, the keyboard input and mouse operation performed by the user are monitored, and which part of each document is inserted into the new document by copy and paste or file insertion during document editing is recorded. As a result, a range inserted in a new document that is equal to or greater than a preset threshold is set as a partial document.

  Next, the partial document extraction unit 2 extracts partial documents based on the relevance between pages (S15).

  In the case of a multi-page document where the title of each page is a serial number or the like, the related range is set as a partial document.

  For example, if the title is “search method (1)”, “search method (2)”, “search method (3)”, etc., these three pages are regarded as related, and one partial document And

  The method of determining the relevance is to divide the text used in the title of each page into words, calculate the percentage of the same word used in the previous and next pages, and if it is above a preset threshold, the title Are determined to be similar and related.

As a simple calculation method, when the number of words in two title sentences is t ₁ and t ₂ and the number of commonly used words is c, the similarity s is expressed as s = ((c / t ₁ ) + (c / t ₂ )) / 2
And s has a value between 0 and 1, and a larger value can be considered to be more similar.

  Alternatively, only the numbers in parentheses in the title sentence may be extracted, and the range in which the numbers increase one by one may be regarded as related.

  In addition, using the LSI method described above, the appearance frequency of words included in the sentence of the title of the page is converted into a feature vector, the Euclidean distance between the vectors is calculated, and when the title sentence is equal to or less than a preset threshold, They may be considered similar and related.

  Further, the relevance may be determined based on the appearance between pages. When explaining one figure over many pages, a similar figure may appear many times. In such a case, features such as color and shape are extracted from the page, and it is determined whether they are related by their similarity.

  A technique called a color histogram can be used for color similarity. In the color histogram, the color of each pixel in the image is classified into any of several color categories, and the similarity is determined based on the ratio of the pixels classified into each color segmented.

  For layout similarity, a color layout technique can be used. In the color layout, an image is divided into several partial areas, the average color of each area is calculated, and the similarity is determined based on the similarity of the average color in each area.

  In addition, by analyzing the structure of the page, it is possible to use the feature that the figure is arranged at which position on each page and the text is arranged at which position. Wavelet transform can be used to determine this form of similarity. In the wavelet transform, an image is analyzed, and a position on the screen and a frequency in the vertical and horizontal directions at the position (information on whether there is a small change or a large change) are output.

  Since the output of the method for extracting the appearance features can be expressed as a vector, the above-mentioned Euclidean distance can be used for calculating the similarity.

  These partial document extraction processes may execute all of the above-described methods, may be executed in combination with a plurality of methods, or may extract partial documents by any one method. Good. In addition, a plurality of partial documents may be extracted from one document, or there may be a portion overlapping the range of each partial document.

  When partial documents in the same range are extracted by a plurality of partial document extraction methods, they may be combined into one partial document.

  The partial document extraction unit 2 outputs the partial document information extracted in this way to the partial document information storage unit 3 and stores it in the storage device 7 (S16).

  For each partial document, a new document independent of the original document may be created, or only information such as the position and length in the original document may be stored.

  Information to be stored includes each document information (file name (URL: Uniform Resource Locater), creation date, title, creator, etc.) and information of each partial document (document ID, position in document, length, partial document) Extraction method, text, image, etc.).

  In order to speed up the search process, a search index may be created in advance. As a method, a general text search method (N-gram method, full-text search method by morphological analysis, etc.), an image similarity search method, or the like can be used.

  Further, when the document is reused, there is a possibility that many reused parts appear in the partial document unit. For this reason, when partial documents having similar contents among a plurality of documents are extracted, it is conceivable that only one partial document is presented as a search result.

  In this case, only the information of one representative partial document is registered as the partial document information described above, and the partial document information similar to the partial document is stored in a separate table to speed up the search process. Can do. In order to determine the similarity between partial documents, the above-described similarity determination based on the appearance frequency of words and similarity determination based on the appearance of a page can be used.

  Next, the partial document extraction unit 2 determines whether all the documents instructed by the user have been processed (S17). If all the documents have not been processed, the process returns to S11 to acquire the next document. When all processing is complete, the process ends.

  In this way, the partial document is extracted from the document designated by the user, and the information is stored in the storage device 7.

  FIG. 3 is a diagram illustrating an example of partial documents extracted from each document by each partial document extraction method.

  In the example of FIG. 3, a part of the partial document extracted with the reuse frequency is extracted as a partial document by the similarity of contents or by the explanatory text, and is the same as the partial document extracted by the relevance between pages. Is a partial document extracted by similarity of contents.

  FIG. 4 is a diagram illustrating a data configuration example of partial document information based on the partial document extracted in the example of FIG. In this example, the information of the same partial documents extracted by the relevance between pages and the similarity of contents are combined into one (5th line).

  FIG. 5 is a flowchart for explaining a process of searching for a partial document from the partial document information thus stored.

  Upon receiving a search request from the user, the search condition input unit 5 displays a search condition input screen as shown in FIG. 6 and searches for keywords, similar images, partial document lengths, partial document extraction methods, and the like. An input of conditions is requested (S21).

  The length of the partial document is the length of the partial document extracted by the partial document extraction process. By inputting the length of necessary information such as a short one (concise) or a long one (details), a partial document having an appropriate length can be searched.

  In the case of a document with a clear number of pages, such as a presentation material, the number of pages is preferable for specifying the length.

  When a part of a page is included as a partial document, the size of the area (or the ratio of the area size to one page) may be designated.

  In addition, when a text-centered document is included, the number of lines may be designated as the length.

  By specifying the length of the partial document when searching, specify the length of the necessary part, such as the part explained in detail, the part explained briefly, the part explained about 3 pages Search.

  When the search condition is input, the search condition input unit 5 inputs the input search condition to the partial document search unit 4.

  The partial document search unit 4 searches the partial document information stored in the storage device 7 according to the input search condition, acquires partial document information that matches the search condition (S22), and acquires it as a search result. The partial document information is input to the search result output unit 6.

  The search result output unit 6 outputs a search result display screen as shown in FIG. 7, for example, based on the input partial document information (S23).

  In the example of FIG. 7, the file name of the original document, the length of the partial document, the length of the original document, the summary text, the thumbnail image, etc. are displayed as the partial document information.

  From the ratio between the length of the original document and the length of the partial document, it is possible to determine the relevance between the partial document and the original document and to determine the ease of reuse.

  For example, if the length of the original document is much longer than the length of the partial document, the relationship between the keyword and the entire document is low because the original document is long but there are few parts related to the specified keyword. .

  On the contrary, when the length of the partial document and the length of the original document are not so different, it is understood that the whole document is related.

  In addition, when there are many short partial documents related to a keyword in one document, it can be understood that the entire document is related to the keyword but is difficult to quote.

  During such screen display, the search result output unit 6 determines whether the user has selected the copy of the partial document by clicking the button labeled “Cite page” on the screen (S24), If selected, the contents of the corresponding partial document are copied to the clipboard (S25).

  If the copy of the partial document is not selected, the search result output unit 6 determines whether the user has selected the display of the partial document by clicking the thumbnail image of the partial document on the screen (S26). If so, the corresponding application is activated to display the partial document (S27).

  If the display of the partial document is not selected, the search result output unit 6 determines whether the user has selected the re-search by clicking the button labeled “Search screen” on the screen (S28), If it is selected, the process returns to S21 and the search condition input unit 5 displays the search screen. If not selected, the process returns to S24 to determine whether a copy of the partial document is selected.

  In this way, a part of the document is extracted and saved as a partial document, searched from the saved partial document according to the user's conditions, and the content of the partial document is copied to the clipboard with one click from the search result display screen. Therefore, it is possible to reduce the trouble of reusing a document such as searching for a target portion from the retrieved documents.

  When a button labeled “Display similar documents” is clicked on the search result screen of FIG. 7, partial documents having similar contents between the plurality of documents described above are extracted. When only the information of one partial document is registered and the information of the partial document similar to that partial document is saved in another table, the information of the partial document saved in another table is displayed and underlined. When the file name of the original document with is clicked, the corresponding application is started and the original document is displayed.

  FIG. 8 is a diagram illustrating another example of the search result display screen. In the example of FIG. 8, the thumbnail images of the partial documents as search results are arranged in a line in the horizontal direction. Documents with different vertical positions are different partial documents, arranged in the vertical direction in the order of search results. Although only images are displayed in the figure, partial document information may be displayed above or to the left of each image.

  In this display, you can move the viewpoint up, down, left and right by clicking the arrow buttons in the upper right window, zoom in by clicking the “Z +” button, and click the “Z−” button. By doing so, it is possible to zoom back and display a portion of the partial document in an enlarged manner, or display the whole in an overhead view.

  Next, processing for creating a template document from a partial document of the partial document search apparatus according to the present embodiment will be described.

  For example, when creating a document composed of a plurality of topics such as a company technology introduction, it may be desired to create a single document by connecting partial documents that explain each topic. In such a case, a document that introduces the same topic is not necessarily found even if the existing introduction material is reused. Therefore, it is necessary to search and connect the partial documents.

  Even if an existing document introducing the same topic is found, the length of the appropriate description and the content of the description may differ depending on the person introduced and the explanation time. For example, when explaining detailed technical details).

  Therefore, by enumerating necessary topic items, partial documents related to the topic can be searched, and a template document can be created from the searched partial documents.

  FIG. 9 is a diagram showing a flow of creating a template document. First, the user creates a table of contents document listing the topic items necessary for the document to be created. Then, a template document creation instruction is given to the partial document search apparatus of the present embodiment.

  When the search condition input unit 5 receives an instruction to create a template document, as shown in the flowchart of FIG. A screen as shown in FIG. 6 is displayed, and input of the table of contents document and the length of each topic is requested (S31).

  When the table of contents document and the length of each topic are input, the search condition input unit 5 extracts keywords used for the search from the items specified in the table of contents document (S32). For keyword extraction, a morphological analysis method or the like can be used.

  In the example of FIG. 9, for example, “cross media” and “search” from topic 1 are “similar”, “image”, and “search” from topic 2, and “office document” and “search” are from topic 3 From the topic 4, “video” and “search” are extracted as keywords.

  When keywords are extracted for each topic item, the search condition input unit 5 inputs the keyword for each extracted topic item and the length of each topic to the partial document search unit 4.

  The partial document search unit 4 searches for a partial document using the keyword and length input for each topic (S33).

  Next, the partial document search unit 4 selects an appropriate combination of partial documents from the partial documents for each topic obtained as a search result (S34).

Since the partial documents searched for each topic may have the same length but may have different levels of explanation, an appropriate combination is selected by the following method.
(1) When the original documents of partial documents used in the same document are the same document, it is assumed that the combination of these partial documents is an appropriate combination. When similar partial documents are extracted at the time of extracting the partial documents described above, if there are similar partial documents whose original documents are the same document, it is assumed that the combination is appropriate.
(2) Similar usage, words are used Even if the topics are different, if they are explained in similar words, it is assumed that the combination is appropriate. If a word is extracted using the morphological analysis described above and a common word is used, it is assumed that the combination is appropriate. Similarly, sentence endings may be extracted using morphological analysis, and the same endings (in other words, different keys) may be combined appropriately.
(3) Similar screen configuration and color usage are used. The screen configuration is similar or the screen color usage is similar is selected as an appropriate combination. For the determination of the similarity of the screen configuration and the similarity of the color usage, the method used for the above-described determination of the appearance similarity of the page at the time of partial document extraction can be used.

  The partial document search unit 4 inputs the partial documents of each topic selected as an appropriate combination in this way to the search result output unit 6.

  The search result output unit 6 outputs the input partial document for each topic, for example, 4. The thumbnail images are displayed side by side in the horizontal direction as shown in FIG.

  It should be noted that a plurality of appropriate partial documents may be selected. As shown in the lower figure, a plurality of candidates may be displayed side by side in the vertical direction so that the user can select them.

  Then, when there are a plurality of candidates, the search result output unit 6 selects one partial document from the plurality of candidates, and inputs a template document creation instruction. The selected partial documents are connected to create a template document (S36), and the corresponding application is activated to display the template document.

  In this way, the partial documents related to the table of contents items created by the user are respectively searched, and an appropriate combination of partial documents is selected from the searched partial documents, and the template document is generated by the selected partial documents. A template document suitable for the user's purpose can be created simply by creating a table of contents item.

  In this embodiment, an appropriate combination of partial documents is selected and displayed. However, only a prescribed number is presented in a certain order such as the similarity order of keywords or the order of reuse frequency to the user. You may make it select.

  As described above, in this embodiment, a part of a document is extracted and saved as a partial document, searched from the saved partial document according to the user's condition, and the content of the partial document is displayed with one click from the search result display screen. Is copied to the clipboard, so that it is possible to reduce the trouble of reusing a document such as searching for a target portion from the retrieved documents.

It is a figure which shows the partial document search device of one Embodiment of this invention, and is the block diagram. It is a flowchart for demonstrating the partial document extraction process of this embodiment. It is a figure which shows the example of the partial document extraction of this embodiment. It is a figure which shows the example of a data structure of the partial document information of this embodiment. It is a flowchart for demonstrating the partial document search process of this embodiment. It is a figure which shows the example of the search condition input screen of this embodiment. It is a figure which shows the example of the search result display screen of this embodiment. It is a figure which shows the other example of the search result display screen of this embodiment. It is a figure which shows the flow of template document preparation of this embodiment. It is a flowchart for demonstrating the template document preparation process of this embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Document acquisition part 2 Partial document extraction part 3 Partial document information preservation | save part 4 Partial document search part 5 Search condition input part 6 Search result output part 7 Storage device

Claims (10)

Extracting a related description range from a document to a computer to form a partial document;
Storing the extracted partial document information;
Searching and outputting the partial document that matches the input search condition from the stored information;
A program for running Making the computer a series of paragraphs or pages with similar word frequencies, the scope of the related description;
The program according to claim 1, further comprising: Making the computer a paragraph or page in a format that explains words, the scope of the related description;
The program according to claim 1 or 2, wherein the program is further executed. A step in which a range in which the number of times of reuse is set in the computer is larger than a set number of times is the range of the related description;
The program according to any one of claims 1 to 3, wherein the program is further executed. In the computer, in a document in which a title is assigned to each page, a continuous page related to the content of the title is included in the related description range;
The program according to any one of claims 1 to 4, wherein the program is further executed. The step of setting the related description to a range of pages that are similar to the computer in the screen configuration of pages,
The program according to any one of claims 1 to 5, wherein the program is further executed. Making the computer the range of the said description that the continuous page where the color composition of a page is similar,
The program according to any one of claims 1 to 6, wherein the program is further executed. Storing the length of the partial document as information of the partial document in the computer;
Inputting a length of a partial document as the search condition;
The program according to any one of claims 1 to 7, wherein the program is further executed. A partial document extraction unit that extracts a related description range from a document to form a partial document;
A partial document information storage unit for storing extracted partial document information;
A partial document search unit that searches for the partial document that matches the search condition input from the stored information;
A partial document search device comprising: Extracting a range of related descriptions from a document to form a partial document;
Storing the extracted partial document information;
Searching and outputting the partial document that matches the input search condition from the stored information; and
The partial document search method characterized by performing this.