JP2010134766A - Document data processing apparatus and program thereof - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To efficiently support confirmation of the validity of numerical information, in document data. <P>SOLUTION: An important expression extracting unit 16 of a document data processing apparatus 100 reads an important expression extraction/message creation rule dictionary, from a dictionary information storage unit 13 of a storage device 12; and if an extraction pattern of the dictionary and an input sentence match, it extracts an important expression, according to the matching expression in the input sentence that matches the extraction pattern. By using the important expression extracted by the important expression extracting unit 16, a message creating unit 18 refers to a message for presenting a user with the important expression in the important expression extraction/message creation rule dictionary. The message creating unit 18 substitutes the extracted important expression into the variable portion of the message referred to and creates the text of the message for presenting the user with the important expression. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a document data processing apparatus and program for supporting calibration of numerical information in document data (including table information).

Conventionally, as a technique that can be used for supporting calibration of numerical values in document data, there has been a technique for extracting numerical information in the document data (see, for example, Patent Document 1). In this method, numerical information is extracted from sentence data, and a sentence is subjected to dependency analysis, information related to the extracted numerical information is acquired, and numerical information and an object represented by the numerical information are acquired.
JP 2005-157853 A

  However, in order to confirm the validity of the numerical information in the document data, it is not sufficient to extract the numerical information from the text data, and the user discovers the description of the numerical information from the text data. It is necessary to find a description that expresses what the numerical information means. However, this operation is not easy for the user and requires a lot of labor.

  SUMMARY OF THE INVENTION An object of the present invention is to provide a document data processing apparatus and program for efficiently supporting confirmation regarding the validity of numerical information in document data (including table information).

  That is, the present invention made to solve the above problems includes a document input means for inputting document data, a storage means for storing rule information for extracting specific information from the document data, and the document Extraction rule element acquisition means for generating an extraction pattern based on the rule element information extracted in accordance with the rule information from the data and the extraction rule template stored in the storage means, and generating specific information from the extraction pattern; Search means for searching for information corresponding to specific information extracted by the extraction rule element acquisition means from information in the document data input by the document input means; and specific information extracted by the extraction rule element acquisition means Among the above, information corresponding to the information searched by the search means, a matching means for matching the searched information, and a matching result by the matching means A document data processing apparatus and the program characterized by comprising a generating means for generating a message indicating.

Further, the extraction rule element is filtered while referring to the unnecessary word list and the selection rule stored in the storage means, and the extraction rule element acquisition means for generating an extraction pattern for the extraction rule element selected by this filtering is performed. A document data processing apparatus and a program thereof characterized by having an extraction rule element filter means for sending.

  According to the present invention, it is possible to efficiently support confirmation regarding the validity of specific information (particularly numerical information is representative) in document data (including table information). Therefore, the efficiency of document creation is improved and it is possible to support creation of a high-quality document.

Hereinafter, embodiments of the present invention will be described with reference to a plurality of drawings.
(Basic embodiment)
First, before describing the first embodiment of the present invention, the basic embodiment of the present invention will be described. This is because the operation of the basic embodiment of the present invention forms the basis of the present invention and is considered to facilitate understanding of the present invention.

FIG. 1 is a block diagram showing an example of the configuration of a document data processing apparatus 1 according to the basic embodiment of the present invention.
As shown in FIG. 1, a document data processing apparatus 1 according to a basic embodiment of the present invention includes a control unit 11 that controls processing operations of the entire apparatus, a storage device 12 that is a storage means, an input / output interface 15, and an extraction. An important expression extraction unit 16 as a means, a certainty factor calculation unit 17 as a calculation means, and a message creation unit 18 as a creation means are provided, and they are connected to each other via a bus 19.

  The storage device 12 is a storage medium configured by hardware such as a hard disk drive or a nonvolatile memory device. The storage device 12 stores a control program to be executed by the control unit 11, the important expression extraction unit 16, the certainty factor calculation unit 17, and the message creation unit 18. Further, the storage device 12 has a dictionary information storage unit 13 for storing an important expression extraction / message creation rule dictionary which is rule information, and further a color correspondence information storage unit for storing color / certainty degree correspondence information 14. The important expression extraction / message creation rule dictionary and the color / certainty correspondence information will be described later. The storage device 12 also functions as a work memory for various processes performed by the control unit 11, the important expression extraction unit 16, the certainty factor calculation unit 17, and the message creation unit 18.

  The important expression extraction unit 16 extracts an important expression in the input document data while referring to the important expression extraction / message creation rule dictionary stored in the dictionary information storage unit 13 of the storage device 12, and the extracted result is stored in the storage device 12. To remember.

  The certainty factor calculation unit 17 calculates a certainty factor indicating how reliable the content of the important expression extracted by the important expression extraction unit 16 and stored in the storage device 12 is, and stores the calculation result in the storage device 12. To do.

  The message creation unit 18 uses the important expressions extracted by the important expression extraction unit 16 and stored in the storage device 12, and refers to the important expression extraction / message creation rule dictionary stored in the dictionary information storage unit 13 of the storage device 12. Meanwhile, a message for confirmation to the user of the important expression is created.

  FIG. 2 is a block diagram showing a configuration example of a terminal device according to the embodiment as the basis of the present invention.

  The terminal device 2 is a device handled by a user who creates and reviews document data. As shown in FIG. 2, the terminal device 2 includes a control unit 21 that controls processing of the entire device, a storage device 22, an input device 23, an output device 24, and a communication interface 25, which are connected to each other via a bus 26. Is done.

  The storage device 22 is a storage medium configured by hardware such as a hard disk drive or a nonvolatile memory device. The storage device 22 stores a control program to be executed by the control unit 21 and also functions as a work memory for various processes performed by the control unit 21.

  The input device 23 is a keyboard or a mouse, for example, and accepts an operation related to creation of new document data. The output device 24 is a liquid crystal display, for example, and displays a document created and reviewed by the user.

  The communication interface 25 is communicably connected to the input / output interface 15 of the document data processing device 1 via a cable or the like, and exchanges document data created in the terminal device 2 and processing results of the document data processing device 1. .

  The communication interface 25 can be communicably connected to an external storage device (not shown) via a cable, and can also input / output document data to / from the external storage device.

  FIG. 3 is a table showing an example of an important expression extraction / message creation rule dictionary stored in the storage device 12 of the document data processing apparatus 1 according to the basic embodiment of the present invention.

  As shown in FIG. 3, in the important expression extraction / message creation rule dictionary stored in the dictionary information storage unit 13 of the storage device 12, important expression extraction patterns and messages are managed in association with each other. Here, the extraction pattern is a description of a combination of each important expression to be extracted with a regular expression. In addition, a message for confirming an important expression is created using each important expression extracted according to the extraction pattern and a message associated with the important expression extraction / message creation rule dictionary. Then, this confirmation message is presented to the user at the output device 24 of the terminal device 2.

  Note that “. *?” In the extraction pattern of the important expression extraction / message creation rule dictionary shown in FIG. 3 means zero or more characters such as characters and symbols. Further, “$ 1” and “$ 2” which are variable parts in the message of the important expression extraction / message creation rule dictionary shown in FIG. 3 are associated with the message in the important expression extraction / message creation rule dictionary. This indicates that reference to the part enclosed in parentheses in the extraction pattern is performed in the description order of the enclosed part.

  The extraction pattern in the first line of the important expression extraction / message creation rule dictionary shown in FIG. 3 is “(CO2 emission). *? ([0-9.,] + Ton)”. This pattern includes “CO2 emission amount”, then includes some character of zero or more, and includes a word indicating an amount whose unit is tons next to the character.

  Subsequently, the extraction pattern in the second row is “(CO2 emissions). *? (US | Australia | Canada). *? ([0-9.,] + Tons)”. This pattern includes “CO2 emissions”, followed by any zero or more characters, followed by any of “US”, “Australia” and “Canada”, followed by units Indicates that it contains a word indicating the amount of tons.

  The extraction pattern in the third row is “(US | Australia | Canada). *? ([0-9.,] + Ton)”. This pattern includes any of “USA”, “Australia” and “Canada”, followed by some zero or more characters, followed by a word indicating the amount in tons. It is shown that.

  FIG. 4 is a diagram showing an example of the color / confidence information stored in the document data processing apparatus 1 storage device according to the basic embodiment of the present invention in a table format.

  As shown in FIG. 4, in the color / confidence information stored in the color correspondence information storage unit 14 of the storage device 12, a number indicating the degree of certainty is associated with color information unique to the certainty. Managed.

  Next, operations of the document data processing apparatus 1 having the configuration shown in FIG. 1 and the terminal apparatus 2 shown in FIG. 2 will be described. FIG. 5 is a flowchart showing an example of the processing operation of the document data processing apparatus 1 according to the basic embodiment of the present invention. In the following description, description will be made corresponding to each step of FIG.

  FIG. 6 is a diagram showing an extraction example of important expressions to be processed by the document data processing apparatus according to the basic embodiment of the present invention.

  First, when document data is input by an operation on the input device 23 of the terminal device 2 (step S1), the control unit 21 stores the document data in the storage device 22. In this embodiment, as shown in FIG. 6, the input document data indicates that “the amount of CO2 emission per capita is the largest in the United States at 19.8 tons, and Australia and Canada each have 18 It continues at 0.0 tons and 14.2 tons. "

  When an input operation related to a request for checking an important expression in the input document data is performed on the input device 23, the control unit 21 processes the document data stored in the storage device 22 via the communication interface 25. Output to device 1. This document data is composed of text information and position information. Here, the position information is a sentence, paragraph, or page index.

  When the input / output interface 15 of the document data processing apparatus 1 inputs text information and position information from the terminal device 2, the important expression extraction unit 16 reads important information in the input document data based on the input text information and position information. In order to extract the expression, the important expression extraction / message creation rule dictionary is read from the dictionary information storage unit 13 of the storage device 12, and the first line of the extraction pattern of the dictionary is referred to (step S2). The extracted pattern is compared to determine whether or not the input sentence matches the extracted pattern (step S3).

  As shown in FIG. 6, when the input document data is “CO2 emission amount per population is...”, The important expression extraction unit 16 extracts the input document data and the important expression shown in FIG. By collating with the message creation rule dictionary, it is determined that this input document data matches the extraction pattern on the second line of the important expression extraction / message creation rule dictionary shown in FIG. That is, “CO2 emissions” is included in the input document data, followed by “is in developed countries”, followed by “US” after this expression. This is because “ga” is included following “” and “19.8 tons” is included following this expression.

  In addition, the important expression extraction unit 16 includes “CO2 emissions” in the input document data shown in FIG. 6. Following this expression, “the United States is the most developed country with 19.8 tons. Is included, followed by "Australia", followed by ", Canada, respectively", followed by "18.0 tons" Therefore, it is determined that the input document data matches the extraction pattern in the second row of the important expression extraction / message creation rule dictionary shown in FIG.

  Then, the important expression extraction unit 16 includes “CO2 emission” in the input document data shown in FIG. 6. Following this expression, “the United States is the most developed country with 19.8 tons. "Australia," is included, followed by "Canada", followed by "18.0 tons," respectively, followed by "14.2 tons" ”Is included, it is determined that the input document data matches the extraction pattern in the second row of the important expression extraction / message creation rule dictionary shown in FIG.

  If the input sentence and the extraction pattern to be matched do not match (NO in step S3), the important expression extraction unit 16 refers to the extraction pattern of the next line in the important expression extraction / message creation rule dictionary, The processing after S2 is performed again. In this embodiment, it is assumed that the input sentence matches the extraction pattern of any line in the important expression extraction / message creation rule dictionary.

  When the input sentence and the extraction pattern match (YES in step S3), the important expression extraction unit 16 extracts the important expression based on the matching expression with the extraction pattern in the input sentence (step S4).

  For example, the important expression extraction unit 16 extracts among the expressions in the input sentence shown in FIG. 6 and the matching expressions with each element of the extraction pattern in the important expression extraction / message creation rule dictionary shown in FIG. Based on the expression enclosed in parentheses in the pattern, as shown in FIG. 6, “CO2 emissions US 19.8 tons” is extracted as the first pattern of the important expression extraction process result, “CO2 emission amount Australia 18.0 tons” is extracted as the second pattern, and “CO2 emission amount Canada 14.2 tons” is extracted as the third pattern of the important expression extraction processing result.

  The important expression extraction unit 16 performs the process of step S2 when the extraction of the important expression accompanying the collation between the end of the input sentence and the important expression extraction / message creation rule dictionary is not completed (NO in step S5). Return.

  When the important expression extraction unit 16 completes the extraction of the important expression accompanying the collation with the important expression extraction / message creation rule dictionary until the end of the input sentence (YES in step S5), the certainty factor calculation unit 17 For each pattern of the important expression extraction processing result extracted by the expression extraction unit 16, a certainty factor indicating how reliable the contents of the important expression belonging to this pattern as the important expression in the input document data is calculated (step S6). . Here, the greater the certainty value, the higher the reliability.

  Specifically, the certainty factor calculation unit 17 calculates the certainty factor based on the number of important expressions included in the extracted pattern. Here, when the number of important expressions included in the extracted pattern is N, the certainty factor is N-1.

  That is, as shown in FIG. 6, the extracted pattern is “CO2 emissions US 19.8 tons”, “CO2 emissions Australia 18.0 tons”, “CO2 emissions Canada 14.2”. In the case of “Ton”, since there are three important expressions included in this pattern, the certainty factor is “2”. The certainty factor calculation unit 17 stores the certainty factor information thus calculated in the storage device 12 together with the corresponding important expression pattern.

  FIG. 7 is a diagram showing an example of creating a message to be processed by the document data processing apparatus 1 according to the basic embodiment of the present invention.

  Next, the message creating unit 18 uses the important expression extracted by the important expression extracting unit 16 and the certainty factor calculated by the certainty factor calculating unit 17, a message for presenting an important expression to the user, that is, an important expression. Create a confirmation message to the user of the expression.

  Specifically, the message creation unit 18 selects one of the patterns for the extracted processing result of the important expression extraction, and extracts the extracted pattern in the important expression extraction / message creation rule dictionary stored in the storage device 12. Among them, when extracting a pattern related to a selected important expression, a message associated with the extracted expression matched in the important expression extraction / message creation rule dictionary is referred to in the important expression extraction / message creation rule dictionary (step S7).

  For example, if the extracted pattern is “CO2 emission US 19.8 tons” as described above, the extraction that matches in the important expression extraction / message creation rule dictionary shown in FIG. Since the pattern is “(CO2 emissions). *? (US | Australia | Canada). *? ([0-9.,] + Tons)” on the second line, the message referenced is the same on the two lines. The second question is "Is $ 1 for $ 2 correct for $ 3?"

  Then, the message creating unit 18 substitutes the extracted important expression into $ 1 and $ 2 which are the variable parts of the referenced message (step S8), and creates a message word for presenting the important expression to the user.

  As described above, when the extracted pattern is “CO2 emission US 19.8 tons” and the referenced message is “is $ 1 of $ 2 correct with $ 3?” The important expression assigned to “1” is the first expression “CO2 emissions” in the extracted pattern, and the important expression assigned to “$ 2” is the second expression “US” in the extracted pattern. The important expression assigned to “$ 3” is the third expression “19.8 tons” in the extracted pattern.

  Therefore, the character string of the created message is “Are the US CO2 emissions 19.8 tons correct?”.

  Further, as described above, when the extracted second pattern is “CO2 emission amount Australia 18.0 tons”, the character string of the created message is “Australian CO2 emission amount is 18.0 tons”. If the extracted third pattern is “CO2 emissions Canada 14.2 tons” as described above, the string of the message created is “Canada CO2 emissions”. Is the amount correct at 14.2 tons? "

  The message creation unit 18 then refers to the color / confidence level correspondence information stored in the color correspondence information storage unit 14 of the storage device 12 based on the confidence level calculated by the confidence level calculation unit 17 for each message. Then, the highlight color of the message is determined (step S9).

  If the extracted patterns are “CO2 emissions US 19.8 tons”, “CO2 emissions Australia 18.0 tons” and “CO2 emissions Canada 14.2 tons”, the certainty corresponding to these patterns is Both are “2”. The color code associated with the certainty factor “2” in the color / certainty factor correspondence information illustrated in FIG. 4 is “#DCDCDC”. Therefore, the created message is highlighted in the color corresponding to the code.

  In this way, the message creation unit 18 outputs the created colored message to the terminal device 2 via the input / output interface 15.

  In addition, instead of distinguishing the emphasized color of the message based on the certainty factor, the message creating unit 18 may distinguish the font size and the message wording itself using the color / confidence correspondence information as another correspondence information. Good.

  When the communication interface 25 of the terminal device 2 inputs a message output from the document data processing device 1, the control unit 21 stores this message in the storage device 22. And the control part 21 acquires the message memorize | stored in the memory | storage device 22, and outputs it to the output device 24 with the input document data memorize | stored in the memory | storage device 22 (step S10).

  FIG. 8 is a diagram showing an example of message output by the terminal device 2 according to the basic embodiment of the present invention.

  The output device 24 displays the message on a monitor by associating the message with the corresponding character string in the input document data as shown in FIG.

  As described above, the document data processing apparatus 1 according to the basic embodiment of the present invention automatically extracts important expressions in user-created document data and uses the contents and the extracted location as confirmation messages to the user. It comes to present. By doing so, the efficiency of the confirmation / correction work for the notation content, particularly the expression that the document creator has determined to be important, is improved, and the quality of the document is improved.

  Further, the document data processing apparatus 1 according to the basic embodiment of the present invention changes the message presentation form using the certainty of the extracted important expression. By doing so, the noticeable degree of each message can be conveyed to the user. Therefore, the user can visually determine a message to be noted.

(First embodiment)
Next, a first embodiment of the present invention will be described. Note that, in the configuration of the document data processing apparatus 100 according to the following description, detailed description of the same parts as those shown in FIG. 1 is omitted. Although not shown, it is assumed that the terminal device 2 is also provided in the following description.

  FIG. 9 is a block diagram showing a configuration example of the document data processing apparatus 100 according to the first embodiment of the present invention.

  As shown in FIG. 9, the document data processing apparatus 100 according to the basic embodiment of the present invention has a table information search unit 31 and a collation processing unit 32 as compared with the basic embodiment of the present invention described above. The extraction rule element acquisition unit 34 is further connected to each other via the bus 19. The storage device 12 further includes an extraction rule template storage unit 33.

  The table information search unit 31 acquires table information in the document data, searches for item information and limited information in the table information while referring to the extraction result by the important expression extraction unit 16, and stores the search result in the storage device 12. Remember. The item information is an expression corresponding to the subject of the confirmation message among the expressions in one pattern of the important expressions extracted by the important expression extracting unit 16. The limitation information is an expression related to limitation of numerical values in the pattern among the expressions in one pattern of the extracted important expression.

  The collation processing unit 32 compares the extraction result by the important expression extraction unit 16 and the extraction result by the table information search unit 31 stored in the storage device 12. The extraction rule template storage unit 33 and the extraction rule element acquisition unit 34 will be described later.

[Table information extraction process]
First, the “table information extraction process” using the document data processing apparatus 100 having the configuration shown in FIG. 9 will be described. FIG. 10 is a flowchart showing an example of a processing operation related to the “table information extraction process” of the document data processing apparatus 100 according to the first embodiment of the present invention. In the following description, description will be given corresponding to each step of FIG. FIG. 11 is a diagram showing an example of numerical information extracted by the document data processing apparatus 100 according to the first embodiment of the present invention.

  First, the document data processing apparatus 100 performs the same processing from step S1 to step S5 described in the basic embodiment of the present invention (steps S21 to S25).

  Then, the table information search unit 31 of the document data processing apparatus 1 acquires the table information included in the input document data, refers to the item information and the limited information of the extraction result obtained by the important expression extraction unit 16, and determines the items from the table information. Information and limited information are acquired (step S26).

  Here, as an example, the configuration of the table information in the input document data is the configuration shown in FIG. In this table information, “CO2 emission (tons)” is described in the second column and the first row, “United States” is described in the first column, the second row, “Australia” is described in the first column, the fourth row, and 1 "Canada" is described in the sixth column. In this case, the table information search unit 31 acquires “CO2 emissions” as item information in the table information, and acquires “USA”, “Australia”, and “Canada” as limited information in the table information.

  Further, the table information search unit 31 acquires numerical information from the table information based on the item information and limited information acquired from the table information in the input document data (step S27).

  When the table information in the input document data has the configuration shown in FIG. 11, the table information search unit 31 has two columns and one row that is a description location of the item information “CO2 emission (tons)” in the table information. “19.8” described in the second column and the second row, which is the same row as the first column and the second row, which is the description location of the limited information “US”, and the item information “CO2 emissions” and Obtained as numerical information corresponding to the limited information “USA”.

  Further, the table information search unit 31 is the same column as the second column and first row, which is the description location of the item information “CO2 emission (tons)” in the table information, and is the description location of the limited information “Australia” 1 “18.0” described in the second column and the fourth row, which is the same row as the fourth column, is acquired as numerical information corresponding to the item information “CO2 emission” and the limited information “Australia”.

  Further, the table information search unit 31 is the same column as the second column and the first row, which is the description location of the item information “CO2 emission (tons)” in the table information, and is the description location of the limited information “Canada” 1 The numerical information “15.2” described in the second column and the sixth row, which is the same row as the sixth column, is acquired as the numerical information corresponding to the item information “CO2 emissions” and the limited information “Canada”.

  FIG. 12 is a diagram illustrating an example of information to be collated by the document data processing apparatus 100 according to the “table information extraction process”.

  For this example, as shown in FIG. 12, the first pattern of the important expression extraction processing result is “CO2 emission US 19.8 tons”, and the second pattern of the important expression extraction processing result is “ “CO2 emissions Australia 18.0 tons” and the third pattern of the important expression extraction processing result is “CO2 emissions Canada 14.2 tons”.

  As shown in FIG. 12, the table information search unit 31 removes the unit from the item information acquired from the table information as shown in FIG. 11, and attaches the unit in the acquired item information to the numerical information. Information obtained by adding the acquired limited information is used as a table information search processing result. Therefore, the pattern of the table information search processing result based on the table information shown in FIG. 11 is “CO2 emissions US 19.8 tons”, “CO2 emissions Australia 18.0 tons” and “CO2 emissions Canada”. 15.2 tons ".

  The collation processing unit 32 collates the extraction result by the important expression extraction unit 16 stored in the storage device 12 and the table information retrieval processing result by the table information retrieval unit 31 (step S28).

Then, the message creating unit 18 creates a message to be presented to the user using the important expressions extracted by the important expression extracting unit 16, the results by the table information searching unit 31, and the matching processing unit 32 (step S29). ). Further, the message creation unit 18 outputs the created message to the terminal device 2 via the input / output interface 15. When the communication interface 25 of the terminal device 2 inputs a message from the message creation unit 18, the control unit 21 associates this message with the corresponding part in the input document data stored in the storage device 22 and monitors it in the output device 24. Display (step S30).

[Display process taking certainty into account]
Next, the “display process with certainty factor” using the document data processing apparatus 100 having the configuration shown in FIG. 9 will be described. FIG. 13 is a flowchart showing an example of a processing operation related to the “display process taking certainty factor into account” of the document data processing apparatus 100 according to the first embodiment of the present invention. In the following description, description will be given corresponding to each step of FIG.

FIG. 14 is a diagram showing an example of various data to be processed by the document data processing apparatus 100 according to the first embodiment of the present invention. “????” in one pattern of the important expression extraction process result shown in FIG. 14 means an expression that has not been extracted in comparison with the table information search process result to be collated by the collation processing unit 32.

  First, the document data processing apparatus 100 performs the same processing as the processing from step S21 to step S27 described in the “table information extraction process” (steps S41 to S47).

  In this embodiment, as shown in FIG. 14, the first pattern of the important expression extraction processing result is “CO2 emission amount 19.8 tons”, and the second pattern of the important expression extraction processing result is “CO2 emission amount”. “Australia 18.0 tons” and the third pattern of the important expression extraction processing result is “CO2 emissions Canada 14.2 tons”.

  The collation processing unit 32 collates the extraction result by the important expression extraction unit 16 and the extraction result by the table information search unit 31 stored in the storage device 12 (step S48). In this embodiment, the collation processing unit 32 individually collates each element in one pattern of the extracted important expression extraction process result and each element in one pattern of the collation target in the table information search process result, The collation result for each element is output in collation order.

  The collation processing unit 32 outputs “true” as a collation result when there is an element included in one pattern of the important expression extraction process result and one pattern to be collated in the table information search process result, and the important expression extraction process If there is an element missing in one of the pattern of the result and the pattern of the table information search processing result, or if there is an element that does not match in both of the objects to be collated, collation with “false” Output the result.

  Then, the certainty factor calculation unit 17 determines the content of the important expression belonging to this pattern in the input document data for each pattern of the important expression extraction processing result extracted by the important expression extraction unit 16 based on the collation result. A certainty factor indicating how reliable the expression is can be calculated (step S49).

  Specifically, the certainty factor calculation unit 17 calculates the certainty factor based on the number of elements included in one pattern of the important expression extraction processing result. Here, when the number of elements included in the one pattern is N, the certainty factor is N-1. The certainty factor calculation unit 17 stores the information on the certainty factor calculated in this way in the storage device 12 together with the pattern of the important expression extraction processing result to be collated.

  Further, the message creation unit 18 stores each message in the color correspondence information storage unit 14 of the storage device 12 based on the certainty factor calculated by the certainty factor calculation unit 17, and the color / confidence shown in FIG. 4. With reference to the degree correspondence information, the highlight color of the message is determined (step S50).

  The message creating unit 18 creates a message to be presented to the user using the important expressions extracted by the important expression extracting unit 16, the results obtained by the table information searching unit 31, and the matching processing unit 32 (step S51). Here, the message creating unit 18 sets “true” for all the matching results for each element in one pattern of the important expression extraction processing result and the one pattern element of the table information search processing result that is the matching target of the pattern. In the case of, include the character string “* The information in the sentence and the table match.” Is included in the message, and both “true” and “false” are included in the matching result of each element of one pattern. If the message contains the character string “* The information in the sentence and the table partially match.” And the matching results for each element of one pattern are all “false” , "* The text and the information in the table do not match."

  FIG. 14 shows a collation result between the important expression extraction process result and the table information search process result. In this example, the first pattern of the extracted important expression extraction processing result is “CO2 emission 19.8 tons”, and the table information search processing result is compared with the first pattern of the important expression extraction processing result. The target pattern is “CO2 emissions in the United States 19.8 tons”, and “CO2 emissions” is included in both. “US” is included in the table information search processing results, but is missing in the important expression extraction processing results. Since “19.8 tons” is included in both, the collation result is “true false true” as shown in FIG.

  In this case, the character string of the first message created by the message creation unit 18 is “* The information in the sentence and the information in the table partially match. Information in the sentence: CO2 emissions 19.8 tons Information: CO2 emissions in the United States of America 19.8 tons.

  Here, since the number of elements in the first pattern of the important expression extraction processing result is two, the certainty factor calculated by the certainty factor calculation unit 17 is “1”. In this case, the color code associated with the certainty factor “1” in the color / confidence information shown in FIG. 4 is “# 999999”.

  Further, the second pattern of the extracted important expression extraction result is “CO2 emission Australia 18.0 tons”, and the table information search process result is compared with the second pattern of the important expression extraction process result. The target pattern is “CO2 emissions Australia 18.0 tons”, and all the elements in this pattern match in both of the verification targets, so the verification result becomes “true true true” as shown in FIG.

  In this case, the message created by the message creation unit 18 is “* The information in the text and the information in the table are the same. Information in the text: CO2 emissions Australia 18.0 tons Information in the table: CO2 emissions Australia 18. 0 tons ".

  Here, since the number of elements in the first pattern of the important expression extraction processing result is 3, the certainty factor calculated by the certainty factor calculation unit 17 is “2”. In this case, the color code associated with the certainty factor “2” in the color / certainty factor correspondence information shown in FIG. 4 is “#DCDCDC”.

  In addition, the third pattern of the extracted important expression extraction processing result is “CO2 emission Canada 14.2 tons”, and the table information search processing result is compared with the third pattern of the important expression extraction processing result. The target pattern is “CO2 emission amount Canada 14.2 tons”, and all the elements in this pattern match in both of the verification targets, so the verification result is “true true true” as shown in FIG.

  In this case, the message created by the message creation unit 18 is as follows: “* The information in the text is the same as the information in the table. Information in the text: CO2 emissions Canada 14.2 tons Information in the table: CO2 emissions Canada 14. 2 tons ".

  Here, since the number of elements in the first pattern of the important expression extraction processing result is 3, the certainty factor calculated by the certainty factor calculation unit 17 is “2”. In this case, the color code associated with the certainty factor “2” in the color / certainty factor correspondence information shown in FIG. 4 is “#DCDCDC”.

  The message creation unit 18 outputs the created colored message to the terminal device 2 via the input / output interface 15. When the communication interface 25 of the terminal device 2 inputs a message from the message creation unit 18, the control unit 21 associates this message with the corresponding part in the input document data and displays it on the output device 24 (step S52).

[Usage process of extraction rule template]
Furthermore, an “extraction rule template use process” using the document data processing apparatus 100 having the configuration shown in FIG. 9 will be described. The extraction rule element acquisition unit 34 constituting the document data processing apparatus 100 acquires the table information in the document data through the “table information extraction process” as described above, and defines the table definition information and the matrix constituting the table. While referring to the position information, the axis item information and unit information in the table information can be extracted and classified as rule element information.

Here, the extraction rule element acquisition unit 34 acquires the extraction rule template stored in the extraction rule template storage unit 33 of the storage device 12, substitutes the classified rule element information into the extraction rule template, and extracts the extraction pattern. Is generated. The generated extraction pattern is stored in the dictionary information storage unit 13 of the storage device 12.

The axis item information is a row and column heading that defines the meaning of the data described in the table, and the unit information is information (for example, millions of yen, tons) indicating the unit of the data in the table. . That is, these pieces of information are expressions related to the limitation of numerical values in the pattern among the expressions in one pattern of the important expressions extracted in this way.

The extraction rule template is an abstraction of the extraction pattern as shown in FIG. This extraction rule template does not describe specific elements (for example, CO2 emissions, United States, tons), such as Subject (country), Country (country), Branch (department), Unit (unit) This is a model of a pattern in which only the extracted classification names are combined according to a predetermined order and description rules. The rule element information is specific element information to be assigned to each classification name in order to embody the extraction rule template.

  As described above, the storage device 12 uses the extraction rule template storage unit 33 for storing the extraction rule template for generating the extraction rule information and the extraction pattern and message generated from the extraction rule element acquisition unit 34. It has a dictionary information storage unit 13 for storing the creation rule dictionary and a color correspondence information storage unit 14 for storing color / certainty degree correspondence information.

  FIG. 15 is a flowchart showing an example of the processing operation related to the “extraction rule template use process” of the document data processing apparatus 100 according to the first embodiment of the present invention. In the following description, description will be given corresponding to each step of FIG.

FIG. 16 is based on the processing from the extraction rule element information acquisition (step S52) by the document data processing apparatus 100 according to the “extraction rule template use process” to the end (necessity) determination of important expression extraction (step S57). It is a figure which shows an example of the various data of a process target.

First, the document data processing apparatus 100 performs the document data input process (step S61) in the same manner as the process of step S41 (that is, step S21 = step S1) described in the “display process with certainty factor”. In this embodiment, the input document data is the input table and the input sentence “Sales of department A is 5.2 million yen” shown in FIG.

  The extraction rule element acquisition unit 34 acquires rule element information for the input table (step S62). In order to acquire this rule element information, first, axis item information and unit information of the table are extracted while referring to the definition information of the table. In the example shown in FIG. 16, “sales” and “profit” from the column headings, “A department”, “B department”, “C department” from the row headings, and “million yen” from the numerical area of the row. Useful information.

Note that the table definition information used as reference information is either external reference information that clearly indicates the position and type (meaning) of the heading or unit described in the table, or information based on the position (matrix number and order) of each element. This is reference information representing an extraction rule.

  Then, by referring to the definition information of the table and the position information (matrix number and order) of each element, information classification and arrangement are performed on the extracted information. In the example illustrated in FIG. 16, three classes “Subject”, “Branch”, and “Unit” are generated for the information extracted from the input table.

  Next, the extraction rule element acquisition unit 34 acquires the extraction rule template stored in the extraction rule template storage unit 33 of the storage device 12, and substitutes the rule element information generated in step S62 into the extraction rule template. Thus, an extraction pattern is generated and stored in the dictionary information storage unit 13 of the storage device 12 (step S63). In the example shown in FIG. 16, for the extraction rule template “Subject. *? ([0-9.,] + Unit)”, the specific extraction rule elements acquired from the table are classified into classification names (Subject, Unit). Substituting every time generates the following two extraction patterns.

1-1 Sales. *? ([0-9,] + 10,000 yen)
1-2 Profit. *? ([0-9,] + 10,000 yen)
When the extraction rule element is substituted for the extraction rule template “Subject. *? Branch. *? ([0-9.,] + Unit)”, the following six extraction patterns are generated.

2-1 “Department A. *? Sales. *? ([0-9.,] + Million yen)”
2-2 “Department B. *? Net sales. *? ([0-9.,] + Million yen)”
2-3 “C department. *? Sales. *? ([0-9.,] + Million yen)”
2-4 “Department A. *? Profit. *? ([0-9.,] + Million yen)”
2-5 “Department B. *? Profit. *? ([0-9.,] + Million yen)”
2-6 “C section. *? Profit. *? ([0-9.,] + Million yen)”
The extracted pattern generated in this way is referred to, and the same processing as the processing in steps S42 to S45 (that is, steps S21 to S25 = steps S1 to S5) described in “Display process with certainty factor” is performed (step S64). To S67), a specific expression (for example, an important expression such as a numerical value) is extracted from the text. In the example shown in FIG. 16, for the input sentence “Sales of department A is 5.2 million yen”, reference is made to Extraction Pattern 2-1, and numerical values such as “Sales A department of 5.2 million yen” Extract important expressions.

  Thereafter, the same processing as the processing of steps S46 to S52 described in "Display process taking certainty factor into account" is performed (steps S68 to 74).

  As described above, the document data processing apparatus 100 according to the first embodiment of the present invention extracts a specific expression in document data created by the user in the processing operation related to the “extraction rule template use process”. The extraction pattern can be automatically generated by using other related information in the document (for example, table information) even if it is not created in advance or cannot be created in advance. . Therefore, the efficiency of checking and correcting the consistency between the sentence portion of the document data and the table information is improved.

(Second Embodiment)
Next, a second embodiment of the present invention will be described. Note that, in the configuration of the document data processing apparatus 200 according to the present embodiment, the following detailed description of the same parts as those shown in the first embodiment will be omitted. Although not shown, it is assumed that the terminal device 2 is also provided in the following description.

FIG. 17 is a block diagram showing a configuration example of the document data processing apparatus 200 according to the second embodiment of the present invention. As shown in FIG. 17, the document data processing apparatus 200 according to the second embodiment of the present invention further includes an extraction rule element filter unit 36 as compared with the first embodiment. The unit 36 is connected to the bus 19. The storage device 12 further includes a filter information storage unit 35. Therefore, the basic functions are the same as those of the first embodiment of the present invention, but the functions of the filter information storage unit 35 and the extraction rule element filter unit 36 are added.

The extraction rule element filter unit 36 refers to an extraction rule element acquired from document data such as a table, while referring to an unnecessary word list and a selection rule stored in the filter information storage unit 35 of the storage device 12. The selected extraction rule element is sent to the extraction rule element acquisition unit 34 to generate an extraction pattern.

Further, the extraction rule element filter unit 36 can total the appearance frequencies of the body of each extraction pattern with respect to the extraction patterns generated by the extraction rule element acquisition unit 34, and can select the extraction pattern by comparison with the threshold value of the appearance frequency. The extraction pattern having the appearance frequency equal to or higher than the threshold value is transmitted to the extraction rule element acquisition unit 34 and the selected result is stored in the dictionary information storage unit 13 of the storage device 12.

Next, the operation of the document data processing apparatus 200 having the configuration shown in FIG. 17 will be described. FIG. 18 is a flowchart showing an example of the processing operation of the document data processing apparatus 200 according to the second embodiment of the present invention. In the following explanation, explanation will be given corresponding to each step of FIG.

FIG. 19 shows processing by extraction rule element filtering (unnecessary word processing... Step S83, selection processing... Step S84 and appearance frequency statistical processing... Step S86) of the document data processing apparatus 200 according to the second embodiment of the present invention. It is a figure which shows an example of the various data of object. FIG. 20 is a diagram showing an example of various data to be processed by the appearance frequency statistical process (step S86) of the document data processing apparatus according to the second embodiment of the present invention.

First, the document data processing apparatus 200 performs the same processing as the processing in steps S61 to S62 described in the “extraction rule template usage process” in the first embodiment (steps S81 to S82). In the second embodiment, “A department”, “B department”, “Total”, “FY 2007”, and “Consolidation” are extracted as branch extraction rule elements as in the example shown in FIG. .

  The extraction rule element filter unit 36 refers to the unnecessary word list stored in the filter information storage unit 35 of the storage device 12 and selects an extraction rule element (step S83). In the unnecessary word list, general expressions (for example, totals, averages, etc.) that are often described in the table (as table information), which are not appropriate as extraction rule elements, are registered as unnecessary words in advance. In the example shown in FIG. 19, since “total” is in the unnecessary word list, it is determined that it is not appropriate as an extraction rule element, and is excluded from the extraction rule element.

  Next, the extraction rule element filter unit 36 refers to the selection rule stored in the filter information storage unit 35 of the storage device 12 and selects an extraction rule element (step S84). Here, the selection rule is a rule at the time of selection defined in advance in order to select an extraction rule element. For example, in the example shown in FIG. 19, since there is a selection rule “Date must not be included in Branch”, if “07 year” is determined to be a date expression, it is excluded from the extraction rule element.

  Then, the extraction rule elements selected by the unnecessary word processing (step S83) and the selection processing (step S84) are transmitted to the extraction rule element acquisition unit 34, and the “extraction rule template use process” of the first embodiment is performed. The same process as the process of step S63 described is performed (step S85), an extraction pattern is generated, and transmitted to the extraction rule element filter unit 36.

  Next, the extraction rule element filter unit 36 adds up the frequency of the appearance of each extraction pattern in the document data with respect to the extraction patterns transmitted from the extraction rule element acquisition unit 34. Then, referring to a predetermined threshold value, only an extraction pattern whose appearance frequency is higher than the threshold value is transmitted to the extraction rule element acquisition unit 34 and registered in the dictionary information storage unit 13 (step S86). For example, in the example shown in FIG. 20, the appearance frequency of pattern 1, pattern 2, and pattern 3 is 5 times, 3 times, and 0 times, respectively. Here, when the threshold value is 0, the pattern 3 is excluded from the extraction pattern generated by the extraction rule element acquisition unit 34 as an inappropriate pattern.

  Thereafter, the same processing as Steps S64 to S74 described in the “Extraction Rule Template Usage Process” of the first embodiment is performed (Steps S87 to 97).

  As described above, the document data processing apparatus 200 according to the second embodiment of the present invention checks the extraction rule elements extracted from the table and removes inappropriate elements and patterns, thereby removing the document data. It is possible to improve the accuracy of checking the consistency between the text portion in the table and the table information.

  Note that the present invention is not limited to the above-described embodiments as they are, and can be embodied by modifying the components without departing from the scope of the invention in the implementation stage. In addition, various inventions can be formed by appropriately combining a plurality of constituent elements disclosed in the respective embodiments. For example, some components may be omitted from all the components shown in the embodiment. Furthermore, you may combine the component covering different embodiment suitably.

The block diagram which shows the structural example of the document data processing apparatus according to embodiment which becomes the foundation of this invention. The block diagram which shows the structural example of the terminal device according to embodiment which becomes the foundation of this invention. The figure which shows an example of the important expression extraction and message creation rule dictionary memorize | stored in the memory | storage device of the document data processing apparatus according to embodiment which becomes the fundamental of this invention in a table format. The figure which shows an example of the color corresponding | compatible information memorize | stored in the memory | storage device of the document data processing apparatus according to embodiment based on this invention in a table format. 6 is a flowchart showing an example of a processing operation of the document data processing apparatus according to the basic embodiment of the present invention. The figure which shows the example of extraction of the important expression of the process target by the document data processing apparatus according to embodiment based on this invention. The figure which shows the example of preparation of the message of the process target by the document data processing apparatus according to embodiment based on this invention. The figure which shows the example of a message output by the terminal device according to embodiment based on this invention. 1 is a block diagram showing a configuration example of a document data processing apparatus according to a first embodiment of the present invention. 6 is a flowchart showing an example of a processing operation related to a “table information extraction process” of the document data processing apparatus according to the first embodiment of the present invention. The figure which shows an example of the numerical information extracted by the "table information extraction process" of the document data processing apparatus according to the first embodiment of the present invention. The figure which shows an example of the information of the collation object by the "table information extraction process" of the document data processing apparatus according to the first embodiment of the present invention. 6 is a flowchart illustrating an example of a processing operation related to a “display process taking certainty factor into account” of the document data processing apparatus according to the first embodiment of the present invention. The figure which shows an example of the various data of the process target by the "display process which considered the certainty factor" of the document data processing apparatus according to the 1st Embodiment of this invention. 6 is a flowchart showing an example of a processing operation related to an “extraction rule template use process” of the document data processing apparatus according to the first embodiment of the present invention. The figure which shows an example of the extraction process of the pattern element from a table | surface by the "use process of an extraction rule model" of the document data processing apparatus according to the 1st Embodiment of this invention. The block diagram which shows the structural example of the document data processing apparatus according to the 2nd Embodiment of this invention. 12 is a flowchart showing an example of processing operation of the document data processing apparatus according to the second embodiment of the present invention. The figure which shows an example of a pattern element filter process by the document data processing apparatus according to the 2nd Embodiment of this invention. The figure which shows an example of an appearance frequency statistical process by the document data processing apparatus according to the 2nd Embodiment of this invention.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Document data processing apparatus, 2 ... Terminal device, 11, 21 ... Control part, 12, 22 ... Storage device, 13 ... Dictionary information storage part, 14 ... Color correspondence information storage part, 15 ... Input / output interface, 16 ... Important Expression extraction unit, 17 ... Certainty factor calculation unit, 18 ... Message creation unit, 19, 26 ... Bus, 23 ... Input device, 24 ... Output device, 25 ... Communication interface, 31 ... Table information search unit, 32 ... Collation processing unit 33 ... Extraction rule template storage unit, 34 ... Extraction rule element acquisition unit, 35 ... Filter information storage unit, 36 ... Extraction rule element filter unit.

Claims (5)

A document input means for inputting document data;
Storage means for storing rule information for extracting specific information from the document data;
Extraction rule element acquisition that generates an extraction pattern based on the rule element information extracted from the document data in accordance with the rule information and the extraction rule template stored in the storage means, and generates specific information from the extraction pattern Means,
Search means for searching for information corresponding to specific information extracted by the extraction rule element acquisition means from information in the document data input by the document input means;
Of the specific information extracted by the extraction rule element acquisition means, information corresponding to the information searched by the search means and a matching means for matching the searched information;
A document data processing apparatus comprising: a creation unit that creates a message indicating a collation result by the collation unit.   Extraction rule elements are filtered while referring to unnecessary word lists and selection rules stored in the storage means, and extraction rule elements selected by this filtering are extracted and sent to an extraction rule element acquisition means for generating an extraction pattern 2. The document data processing apparatus according to claim 1, further comprising rule element filtering means.   The extraction rule element filter means totals the frequency of occurrence of the extraction pattern generated by the extraction rule element acquisition means in the document data, and the extraction rule element acquisition means, when the frequency is equal to or less than a threshold, 3. The document data processing apparatus according to claim 2, wherein expressions occurring at the frequency are excluded from the document. In order to provide a function to extract specific information from document data using a computer,
A document input function for inputting document data;
A storage function for storing rule information for extracting specific information from the document data;
An extraction rule for generating an extraction pattern based on the rule element information extracted from the document data in accordance with the rule information and the extraction rule template stored in the storage unit of the computer, and generating specific information from the extraction pattern An element acquisition function and a search function for searching for information corresponding to specific information extracted by the extraction rule element acquisition unit of the computer from information in the document data input by the document input means;
Among the specific information extracted by the extraction rule element acquisition means, information corresponding to the information searched by the search unit of the computer and a matching function for matching the searched information;
A document data processing program for causing the computer to execute a creation function for creating a message indicating a collation result by the collation means.   To extraction rule element acquisition means for filtering extraction rule elements while referring to unnecessary word lists and selection rules stored in the storage unit of the computer and generating an extraction pattern for the extraction rule elements selected by this filtering 5. The document data processing program according to claim 4, which causes the computer to execute an extraction rule element filter function to be sent.

Images