JP2005011301A - Document processor and document processing program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a document processor by which what context employs a noticeable phrase in each part in a document is efficiently checked with a simple operation. <P>SOLUTION: When a user specifies a keyword, a document analysis section extracts a given range containing the keyword for each part emerging the keyword in the document as a noticeable portion. The noticeable portion is divided into a foregoing part, the keyword, and a later part for display on an image plane. The user selects the second keyword on the image plane and performs processing including the concentration of a result, keyword connection, and keyword change, and the like. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a document processing apparatus and program for efficiently retrieving and extracting important description parts from a document described in a natural language.
[0002]
[Prior art]
An example of the document processing apparatus as described above is an information extraction apparatus described in Patent Document 1. The outline procedure of document processing by this apparatus is as follows.
The subject information extraction unit extracts a subject presentation phrase from each word obtained by morphological analysis of the document.
A word-to-word similarity analysis unit calculates the similarity of each word.
The important phrase extracting unit extracts words having high similarity to the words constituting the subject presentation phrase as important phrases.
-An output part displays the said important phrase linked with the word which comprises the said subject presentation phrase.
[0003]
Patent Document 1 also discloses an apparatus configured to allow a user to specify a subject word / phrase in the process. In this apparatus, first, subject presentation words and related important phrases are extracted from the processing target document by the above procedure, and the contents of the document are displayed with these words underlined. The user can select one of the underlined phrases as a new subject presentation phrase. When the user selects a phrase, the device extracts words having high similarity to the words constituting the phrase as new important phrases, and redisplays the contents of the document with these phrases underlined.
[Patent Document 1] JP 2000-298673 A
[Problems to be solved by the invention]
When examining the contents of a document, there are cases where it is desired to examine in what context the attention word is actually used. For this purpose, it is necessary to examine the descriptions before and after each word where the word of interest appears in the document. When the apparatus described in Patent Document 1 is evaluated from such a viewpoint, the following can be said. First, the method of linking and displaying the subject words and key words only shows the relationship between multiple words or phrases extracted from the entire document, so what context each phrase actually means It is not possible to know what is being used in. On the other hand, in the method of displaying the contents of the document with the underlined important words / phrases, not only the descriptions before and after the attention words / phrases but also the descriptions of other parts are displayed. Therefore, for example, when the document is long, it takes a considerable amount of time to examine the descriptions before and after the word / phrase at all locations where the word / phrase appears in the document.
[0005]
The present invention has been made to solve the above-mentioned problems, and the object of the present invention is to determine in what context the attention word is used in each part of the document by a simple operation. An object of the present invention is to provide a document processing apparatus capable of efficiently examining.
[0006]
[Means for Solving the Problems]
The first document processing apparatus according to the present invention, which has been made to solve the above problems,
A keyword specification section for allowing the user to specify the first keyword,
An attention part extraction unit that extracts a description of a predetermined range including the first keyword as an attention part for each occurrence of the first keyword in the document to be processed;
An attention part display unit that separates each attention part into a preceding part preceding the first keyword, a first keyword, and a subsequent part following the first keyword, and displays it on the screen;
A keyword selection unit for allowing the user to select any of the phrases from the front part or the rear part on the screen as a second keyword; and
When the second keyword is selected from the preceding part, the subsequent part of each target part including the second keyword is selectively displayed, and when the second keyword is selected from the succeeding part, the second part is selectively displayed. A display update unit for selectively displaying the preceding part of each target part including the second keyword,
It is characterized by having.
[0007]
The second document processing apparatus according to the present invention is
A keyword specification section for allowing the user to specify the first keyword,
An attention part extraction unit that extracts a description of a predetermined range including the first keyword as an attention part for each occurrence of the first keyword in the document to be processed;
An attention part display unit that separates each attention part into a preceding part preceding the first keyword, a first keyword, and a subsequent part following the first keyword, and displays it on the screen;
A keyword selection unit for allowing the user to select any of the phrases from the front part or the rear part on the screen as a second keyword; and
A keyword changing unit for instructing the target part extracting unit to extract a target part using the second keyword selected by the user through the keyword selecting unit as a new first keyword;
It is characterized by having.
[0008]
A document processing program according to the present invention causes a computer to function as the first or second document processing apparatus.
[0009]
DETAILED DESCRIPTION OF THE INVENTION
The procedure of document processing by the first document processing apparatus according to the present invention is as follows.
[0010]
(Specify the first keyword)
The user designates the first keyword through the keyword designation unit. For example, the keyword specifying unit displays a keyword input text box on the screen, and the user inputs an arbitrary phrase as the first keyword in the text box.
[0011]
(Extract attention part)
The attention part extraction unit extracts a description of a predetermined range including the first keyword from the document as the attention part. The attention portion is extracted for each location where the first keyword appears in the document. The extraction range is determined according to the following rules, for example.
Example 1: A range starting from the next character after the nearest phrase that appears before the first keyword and ending at the nearest phrase that appears after the first keyword.
Example 2: A range including a predetermined number (for example, 10) of morphemes before and after the first keyword as a reference.
Example 3: Range defined as the logical product of Example 1 and Example 2.
[0012]
(Display the attention part on the screen)
The attention part display unit displays each attention part extracted as described above on the screen by separating it into a first part preceding the first keyword, a first keyword, and a subsequent part following the first keyword. . Examples of display formats include the following.
Example 1: A format that is simply divided into three parts, the first part, the first keyword, and the latter part (hereinafter referred to as “simple three-part display”).
Example 2: A format in which the former part and the latter part are disassembled into morphemes and displayed (hereinafter referred to as “morpheme decomposition display”). Examples of the display format include a tree format (each node corresponds to a morpheme), a table format (each cell corresponds to a morpheme), a split writing format (split using a separating character such as a slash or a blank), and the like.
[0013]
(Selection of second keyword)
The user selects the second keyword from the preceding part or the succeeding part on the screen through the keyword selection unit. For example, in the case of simple three-division display, the user appropriately operates an input device such as a mouse to select (highlight) an arbitrary phrase in the front part or the rear part. In the case of morpheme decomposition display, the user selects an arbitrary morpheme from a large number of morphemes displayed on the screen.
[0014]
(Selection display of front part or rear part)
The display update unit selectively displays the front part or the rear part of each part of interest according to the selection of the second keyword. When the second keyword is selected in the preceding part, the subsequent part of the attention part including the keyword is selectively displayed. When the second keyword is selected in the latter part, the preceding part of the attention part including the keyword is displayed. Selectively display. Here, “selectively display” means that the preceding part or the following part of the attention part including the second keyword (hereinafter referred to as “corresponding part”) is the preceding part or the following part of the attention part not including the second keyword. This is to distinguish and display from a part (hereinafter referred to as “non-applicable part”). Specific examples are shown below.
Example 1: Display only relevant parts on the screen and hide non-relevant parts.
Example 2: Display the relevant part and the non-relevant part in different colors.
[0015]
The procedure of document processing by the second document processing apparatus according to the present invention is as follows.
[0016]
First, similarly to the first document processing apparatus, when the user designates the first keyword, the attention part extraction unit extracts the attention part corresponding to the first keyword, and the attention part display unit displays each attention part. Separately display the first part, the first keyword, and the second part. Next, when the user selects a second keyword from the front part or the rear part of any target part displayed on the screen and executes a predetermined keyword changing operation, the keyword changing unit newly sets the second keyword. It passes to the attention part extraction part as the first keyword. Receiving this, the attention part extraction part extracts the attention part corresponding to the new first keyword. The attention part display unit separates each extracted attention part into a front part preceding the first keyword, a first keyword, and a rear part following the first keyword, and displays them on the screen. Thus, the user can change the keyword as many times as necessary.
[0017]
In the document processing apparatus according to the present invention, when the user selects the second keyword through the keyword selection unit and executes a predetermined keyword linking operation, the words included in the range from the second keyword to the first keyword are linked. There may be further provided a keyword linking unit for passing the character string formed as a new first keyword to the target portion display unit.
[0018]
In the document processing apparatus according to the present invention, when a user selects a second keyword through the keyword selection unit and executes a predetermined original text reference operation, an attention part including the first keyword and the second keyword appears in the original text. You may further provide the original text display part which displays a description part.
[0019]
In the document processing apparatus according to the present invention, when the processing target document is a structured document composed of a plurality of parts, a processing target limiting unit that targets only a specific part of the document may be further provided. Examples of structured documents include documents created according to a prescribed format, such as patent application documents and academic papers, and documents written in a markup language such as HTML or XML.
[0020]
【The invention's effect】
As described above, according to the document processing apparatus of the present invention, each description part in the original sentence including the attention word (first keyword) is displayed together with the preceding part and the subsequent part. As a result, the user can easily confirm the meaning of the word used in the sentence. In addition, by selecting an arbitrary word or phrase before or after the attention word or phrase as the second keyword, it is possible to narrow down the search result or perform a search with a new keyword with a simple operation.
[0021]
【Example】
A patent publication analysis apparatus which is an embodiment of a document processing apparatus according to the present invention will be described with reference to the drawings. FIG. 1 is a schematic configuration diagram of a patent publication analysis apparatus (hereinafter referred to as the present apparatus) of the present embodiment. The apparatus 1 includes a document analysis unit 18 configured by operating a predetermined program on a computer including an input device 10, a storage device 12, a display device 14, a main unit 16, and the like. The storage device 12 stores a file of a published patent publication (hereinafter simply referred to as a document). The main unit 16 includes a CPU, a memory (none of which are shown), and the like.
[0022]
The operation of this apparatus will be described with reference to the flowchart of FIG.
[0023]
First, the user designates a target document file stored in the storage device 12 by a predetermined operation of the input device 10 (step S10). In response to this operation, the document analysis unit 18 reads the sentence saved in the designated document file (step S12), and decomposes the sentence into a large number of morphemes by morphological analysis (step S14). These morphemes are held in memory in the form of a bidirectional list.
[0024]
Next, the document analysis unit 18 displays a keyword analysis screen on the display device 14 (step S16). FIG. 3 shows a schematic configuration of the keyword analysis screen. The keyword analysis screen 20 includes a document information display area 202, an analysis target item designation area 204, an analysis condition setting area 206, and an analysis result display area 208. In the document information display area 202, basic attribute information of the document is displayed. The analysis target item designation area 204 is provided with check boxes corresponding to the respective components of the document. With this check box, the user can specify only desired items in the document as keyword analysis targets. In the analysis condition setting area 206, an “analysis key” text box 210, a “depth” text box 212, and an “execute analysis” button 214 are provided. In the “analysis key” text box 210, a phrase that is the first keyword of the present invention is entered. In the “depth” text box 212, a numerical value for designating the size of the target portion is input. That is, the “depth” indicates the maximum value of the number of morphemes that can be included in the front part and the rear part of each target part.
[0025]
When the user inputs a keyword (first keyword) in the “analysis key” text box 210 and presses the “execute analysis” button 214, the document analysis unit 18 scans the morpheme list to search for the keyword (step S1). S18) For each occurrence of the keyword, the range of the portion of interest corresponding to the keyword (that is, the front part and the rear part) is determined (step S20). The range of interest is determined according to the following rules.
・ Rule for determining the first part: When a punctuation point is included in the range of morphemes for the number of “depths” preceding the keyword, the part from the morpheme following the punctuation point to the morpheme immediately before the keyword is defined as the previous part. . If no punctuation is included in the range, the range is defined as the previous part as it is.
Rules for determining the latter part: When a punctuation point is included in the range of morphemes for the number of “depths” following the keyword, the part from the morpheme immediately after the keyword to the punctuation point is determined as the subsequent part. When the range does not include a punctuation mark, the range is determined as the latter part as it is.
[0026]
Next, the document analysis unit 18 separates the attention part (the front part and the rear part) extracted for each occurrence of the keyword into morphemes and displays them in the analysis result display area 208 in a tree format (step S24). A procedure for generating the morpheme tree will be described with reference to FIGS. Now, consider processing of a sample document having the contents as shown in FIG. The keyword is “morpheme”, the depth is “8”, and the analysis target items are “name of invention”, “summary”, and “claims”. The underline in the sample document is given for explanation and does not constitute the content of the document.
[0027]
Referring to FIG. 4, the keyword “morpheme” appears in six places in the document (see the underlined portion in the sample document). Therefore, this document includes six attention portions in the analysis target range. Table A in FIG. 5 shows the contents of each target portion. Reference numerals P1 to P6 indicate the former part corresponding to the appearance locations of the six keywords, and reference signs S1 to S6 indicate the latter part. The preceding part P6 corresponding to the sixth appearance location is composed of four morphemes from the morpheme next to the punctuation point to the morpheme immediately before the keyword. Further, the rear portion S5 corresponding to the fifth appearance location is composed of five morphemes from the morpheme immediately after the keyword to the punctuation point. The other front and rear parts are each composed of eight morphemes.
[0028]
Next, when the front part P1 to P6 is rearranged using “front 1”, “front 2”,..., “Front 8” as sort keys, the result is as shown in Table B of FIG. When the rear portions S1 to S6 are rearranged using “rear 8” as a sort key, the result is as shown in Table C of FIG. When a plurality of identical phrases appearing continuously in the same column of Table C are combined into one, Table D in FIG. 6 is obtained. If the keywords and each phrase are regarded as nodes of the tree, a tree structure as shown in FIG. 3 is obtained.
[0029]
When any word (morpheme) displayed in the analysis result display area 208 is selected as a second keyword and a predetermined operation is executed, a pop-up menu 209 is displayed on the screen. From this menu 215, the user can select one of the processes described below and cause the document analysis unit 18 to execute it.
[0030]
(Narrow down)
The narrowing-down process is a process for selectively displaying only a portion of interest including both the first keyword and the second keyword. In the present apparatus 1, when the second keyword is selected in the preceding part, the subsequent part is selectively displayed, and when the second keyword is selected in the succeeding part, the preceding part is selectively displayed. As an example, consider the case where the word “document” indicated by reference numeral 216 in FIG. 3 is selected and “narrow down” in the menu 215 is selected. Referring to FIG. 3, the word denoted by reference numeral 216 appears four morphemes before the keyword (“morpheme”). Referring to Table A in FIG. 5, two parts P1 and P4 include the word “document” in the “front 4” column. The contents of the rear stage portions S1 and S4 corresponding to the front stage portions P1 and P4 are exactly the same. Therefore, as shown in FIG. 7, the document analysis unit 18 displays only the branches corresponding to the contents of S1 and S4 in the latter part of the morpheme tree.
[0031]
(Keyword link)
The keyword linking process is a process of executing document analysis by setting a character string formed by linking morphemes included in the range from the first keyword to the second keyword as a new first keyword. The second keyword may be selected in either the front part or the rear part. As an example, consider a case where the word “emphasis” indicated by reference numeral 218 in FIG. 3 is selected and “keyword concatenation” in the menu 215 is selected. In this case, the document analysis unit 18 generates a phrase “emphasize morpheme” by concatenating morphemes included from the first keyword “morpheme” to the second keyword “enhancement”, and generates the new first keyword. Is set in the “analysis key” text box 210. Thereafter, when the user presses the “analysis execution” button 212, the document analysis unit 18 analyzes the document by the procedure described above, and displays a tree as shown in FIG. 8 in the analysis result display area 208.
[0032]
(Keyword change)
The keyword changing process is a process of executing document analysis by setting the word / phrase selected in the analysis result display area 208 as a new first keyword. For example, consider a case where the word “emphasis” indicated by reference numeral 218 in FIG. 3 is selected and “change keyword” in the menu 215 is selected. In this case, the document analysis unit 18 sets “emphasis” as a new keyword in the “analysis key” text box 210 instead of the current keyword “morpheme”. After that, when the user presses the “execute analysis” button 212, the document analysis unit 18 analyzes the document according to the procedure described above, and displays a new tree (not shown) in the analysis result display area 208.
[0033]
(See original)
The original text reference process is a process of displaying the description range of the original text including words from the first keyword to the second keyword in a separate window. For example, consider a case where the word “document” indicated by reference numeral 216 in FIG. 3 is selected and “reference original text” is selected from the menu 215. In response to this operation, the document analysis unit 18 generates a source text reference window 220, and displays the source text in a form in which the phrase “document content is a morpheme” formed by concatenating morphemes from the first keyword to the second keyword is highlighted. Displayed (in the example of FIG. 3, the corresponding word is underlined). If the corresponding phrase appears in multiple places in the original text, all the corresponding phrases are highlighted. By using the “Previous” button 222 and the “Next” button 224, the user can sequentially display a plurality of corresponding words / phrases in the window 220.
[0034]
As mentioned above, although one Example of the document processing apparatus concerning this invention was described, the said Example is only an example to the last. Some modifications are given below.
(Example 1) The format of displaying the analysis result in the analysis result display area 208 may be a table format as shown in FIGS.
(Example 2) The type of document to be processed is not limited to a patent publication, but may be another type of document.
(Example 3) Reading of a document may be performed via a network.
(Example 4) The input device, the display device, and the main unit may communicate with each other via a network.
[0035]
As a specific form of Example 4, an application of the present invention to an Internet information search system (hereinafter referred to as a search server) can be considered. The system operation in this case is as follows. First, the user transmits a search keyword to a search server (corresponding to the document analysis unit 18 of the main unit 16) using his / her terminal (corresponding to the input device 10 and the display device 14 in FIG. 1). Upon receiving this, the search server analyzes the contents of each WEB site stored in the storage device in the natural language (such as an HTML document) according to the procedure described above, and transmits the result to the user terminal. To do. On the screen of the terminal that receives the result, the extracted portion from each document including the keyword is displayed in a predetermined format such as a tree format or a table format. Further, on this screen, a link to the WEB site from which each document is published is also displayed. When the user selects the second keyword on the screen and presses a button such as “narrow down” or “keyword link”, a processing request corresponding to the pressed button is transmitted from the terminal to the server. Upon receiving this request, the server executes the requested process and returns the result to the terminal. In this way, the user can narrow down candidate documents suitable for the purpose while confirming in which context the keyword is used in each document. When a document suitable for the purpose is found, the WEB site that is the source of the document can be easily browsed through the link.
[0036]
In addition to the above, various forms of embodiments are conceivable within the spirit and scope of the present invention.
[Brief description of the drawings]
FIG. 1 is a schematic configuration diagram of a patent publication analysis apparatus which is an embodiment of a document processing apparatus according to the present invention.
FIG. 2 is a flowchart showing the operation of the apparatus shown in FIG.
FIG. 3 is a schematic configuration diagram of a keyword analysis screen.
FIG. 4 is a diagram showing the contents of a sample document.
FIG. 5 is a diagram for explaining a procedure for generating a morpheme tree;
FIG. 6 is a diagram for explaining a procedure for generating a morpheme tree.
FIG. 7 shows an example of a morpheme tree obtained by a narrowing process.
FIG. 8 shows an example of a morpheme tree obtained by keyword concatenation processing.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 10 ... Input device 12 ... Memory | storage device 14 ... Display apparatus 16 ... Main body unit 18 ... Document analysis part

Claims (4)

A keyword specification section for allowing the user to specify the first keyword,
An attention part extraction unit that extracts a description of a predetermined range including the first keyword as an attention part for each occurrence of the first keyword in the document to be processed;
An attention part display unit that separates each attention part into a preceding part preceding the first keyword, a first keyword, and a subsequent part following the first keyword, and displays it on the screen;
A keyword selection unit for allowing the user to select any of the phrases from the front part or the rear part on the screen as a second keyword; and
When the second keyword is selected from the preceding part, the subsequent part of each target part including the second keyword is selectively displayed, and when the second keyword is selected from the succeeding part, the second part is selectively displayed. A display update unit for selectively displaying the preceding part of each target part including the second keyword,
A document processing apparatus comprising: A keyword specification section for allowing the user to specify the first keyword,
An attention part extraction unit that extracts a description of a predetermined range including the first keyword as an attention part for each occurrence of the first keyword in the document to be processed;
An attention part display unit that separates each attention part into a preceding part preceding the first keyword, a first keyword, and a subsequent part following the first keyword, and displays it on the screen;
A keyword selection unit for allowing the user to select any of the phrases from the front part or the rear part on the screen as a second keyword; and
A keyword changing unit for instructing the target part extracting unit to extract a target part using the second keyword selected by the user through the keyword selecting unit as a new first keyword;
A document processing apparatus comprising: The attention part display unit disassembles and displays each of the front part and the rear part into morphemes,
The document processing apparatus according to claim 1, wherein the keyword selection unit causes the user to select any morpheme included in the front part and the rear part as the second keyword. A document processing program for causing a computer to function as the document processing apparatus according to claim 1.

Images