JPH07219962A - Keyword creation device - Google Patents

Abstract

(57) [Summary] [Purpose] To efficiently improve the efficiency of keyword creation work by efficiently reviewing and correcting the search word candidates extracted from text data and their readings. [Structure] Referring to a keyword dictionary in which nouns and their readings that may become a search word for a CD-ROM are registered, search word candidates extracted from text data and their readings read from the keyword dictionary A keyword creating device that displays keyword data including and on a display and creates a keyword suitable as a search word on the display, creates display data from the keyword data read from the floppy disk 16, and displays the keyword data. A display data creation unit 18 to be displayed on the display 12, an instruction unit (pen) 14 that instructs calibration on the display 12, a recognition unit 12B that recognizes the instructed calibration content, and a keyword based on the recognized calibration content. The correction unit 20 directly corrects the data.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention relates to a keyword creating device,
In particular, the present invention relates to a keyword creating apparatus suitable for applying to creating keywords used for high-speed retrieval of text stored in a CD-ROM by direct access.

[0002]

2. Description of the Related Art Various publications such as dictionaries are CDs.
-Recorded and stored in ROM and published. This CD-
In ROM publications, direct access is generally performed by using a search word (keyword) in order to perform a high-speed search for the content (explanatory text, explanatory text) of the stored text. As shown conceptually in FIG. 14, this keyword is created by extracting a large number of nouns from the text data to be stored in the CD-ROM as candidates for search words, and creating keyword data. A keyword suitable as a search word is determined from the above, and it is recorded in a CD-ROM together with text such as a sentence to be a publication.

A conventional keyword extraction system using a natural language processing system is used for keyword creation, and all nouns are extracted from the text data with kana (reading) and can be used as keywords from the noun data. There is a work in which keyword data having a certain property is created, and the keyword data is reviewed and modified so that only nouns suitable for a search word are used as a keyword.

[0004]

However, when a keyword is created by reviewing and correcting the keyword data extracted from the text data by the keyword extraction system, the reviewer conventionally considers the keyword data and the text body containing the keyword. While looking at the printed list in contrast with the above, if there is inappropriate keyword data, proofread the correction content by writing in red in the list, and then the proofreader receives the list and writes it in red. According to the contents, since the proofreading (editing) work and the correction work are performed separately, such as the correction of the same keyword data as the list displayed on the display, the keyword creation work is too burdensome. There was a problem.

The present invention has been made to solve the above-mentioned conventional problems, and it is possible to efficiently perform the editing / correcting work of the keyword data extracted as the search word candidates from the text data, and the efficiency of the keyword creating work. An object of the present invention is to provide a keyword creating device that can improve the keyword.

[0006]

The present invention is a CD-ROM.
Word candidate in which a noun that may become a search word for words and its pronunciation are registered, and a search word candidate extracted from the CD-ROM text data and the search word candidate read from the keyword dictionary A keyword creating device for displaying keyword data including the reading of, on a display, and editing / correcting the keyword data on the display to create a keyword suitable as a search word,
Display data creation means for creating display data from the keyword data and displaying the keyword data on a display, instructing means for instructing calibration of the keyword data displayed on the display, and the instructing means for instructing The problem is solved by providing a recognizing unit that recognizes the proofreading contents and a correcting unit that corrects the keyword data based on the proofreading contents recognized by the recognizing unit.

In the present invention, and in the above-mentioned keyword creating device, when the display data creating means includes a plurality of identical search word candidates in the keyword data, a new search word candidate first appearing on the display and 2 The present invention has a function for distinguishing and displaying already-exposed search word candidates appearing after the th, and the correction means has a function of reflecting the result of correcting the new search word candidate in the already-existing search word candidate and the already-existing search word candidate selected. It is configured to have a function of correcting only the above.

According to the present invention, in the above-mentioned keyword creating device, the display data creating means has a function of displaying the search word candidates on the display in comparison with the text containing the search word candidates. It is what

[0009]

In the present invention, as shown in the flow chart of FIG. 1, keyword data (calibration monitor) including search word candidates and their readings is displayed on the display, and calibration is performed by the indicating means on the display. The proofreading contents are recognized, and the corresponding keyword data is corrected based on the recognition, so that the reviewer can see the display while looking at the keyword data displayed using the pointing device such as a pen. On the other hand, just by instructing calibration (correction) with the same feeling as conventional calibration work,
Since the correction according to the proofreading contents can be performed at the same time, the efficiency of the keyword data creation work can be significantly improved.

Further, in the present invention, when the keyword data includes a plurality of identical search word candidates, the display data creating means appears first on the display and the second and subsequent new search word candidates. It has a function of distinguishing it from existing search word candidates, and the correcting means reflects the result of modifying the new search word candidates in the existing search word candidates, and the function of correcting only the selected previous search word candidates. In the case of having the following, it becomes possible to clearly distinguish the new search word candidate and the already-existing same search word candidate on the display, and for the same search word candidate, if the new search word candidate is reviewed and corrected only once. As a result, it is possible to significantly improve the efficiency of the review / correction work. In addition, for a specific already-existing search word candidate, it is selected to affect other same search word candidates. Since it is possible to make unique corrections without doing so, it is possible to make appropriate corrections in response to individual requests for the same search word candidate, and it is possible to further improve the efficiency of keyword creation work. .

Further, in the present invention, when the display data creating means has a function of displaying the keyword data in comparison with the corresponding text, the search word candidates in the text of the text data are used. Since the person can be taken into consideration, it is possible to accurately edit and correct the search word candidate.

[0012]

Embodiments of the present invention will now be described in detail with reference to the drawings.

FIG. 2 is an explanatory diagram showing a schematic structure of a keyword creating system according to an embodiment of the present invention and a processing flow, and FIGS. 3 and 4 are flowcharts showing a processing procedure by the system. .

The keyword creating system of the present embodiment includes a keyword data creating device 6 for creating a text data stored in a floppy disk FD1 by inputting the text and referring to the keyword dictionary to extract the keyword data. The created keyword data is input by the floppy disk FD2 in which the created keyword data is stored, the keyword data is reviewed and modified to create a keyword, and the keyword is stored in the floppy disk FD3. And a device 8. However, the storage medium used is not limited to a floppy disk.

In the present embodiment, the keyword data is created according to the procedure shown in FIG.

That is, first, the text data (body text) corresponding to the text to be stored in the CD-ROM is stored in the floppy disk.
It is input to the data creating device 6 by the disk FD1 (step 110), the noun of the search word candidate is extracted from this text data by referring to the keyword dictionary created in advance (step 112), and the extracted search word candidate is extracted. Reading is read from the keyword dictionary (step 114), and keyword data including the search word candidate and its reading is created (step 116). FIG. 5 conceptually shows the above processing. The created keyword data is stored in the floppy disk FD2 and edited / detailed later.
Used for correction work.

As the keyword data creating device 6, for example, a workstation can be used,
As the input text data, data processed for CD-ROM, for example, created in MS-DOS format can be used.

Only the words (nouns) that can be keywords are registered in the keyword dictionary used for extracting the nouns of the search word candidates from the text data.
The content of the matter is a matching key and keyword information as conceptually shown in FIG. 6, and the keyword information includes (A) search word (keyword) candidate, (B) its reading,
(C) Compound word division position, (D) related word, and (E) related word reading are included. Note that the compound word is a noun in which a plurality of unit nouns are connected, and the related word is a noun related to a search word (including a compound unit noun (expanded form)).
Means

In order to extract search word candidates from this keyword dictionary, if there is a noun corresponding to the matching key in the text, the keyword (search word) candidate corresponding to the matching key and its keyword information are read from the keyword dictionary. Done by. This keyword dictionary is created in advance by selecting a noun that can be a keyword from data such as index words and headwords of dictionaries.

An example of the contents of the keyword dictionary is shown below. (Example) Matching key: Mediterranean trade Keyword information ・ Keyword: Mediterranean trade ・ Reading ： Cichu Kaiboueki ・ Division position ： Mediterranean ● Trade ・ Related words ： Mediterranean, trade, Europe, trade ・ Related word reading ： Chichukai, Boueki

The keyword data created in step 116 described above is input to the review / correction device 8 of this embodiment and processed according to the following procedure to create a keyword.

The external appearance of the editing / correcting device 8 is shown in FIG.
In addition, as shown in FIG. 8, the functional configuration is provided with a display 12 on the apparatus main body 10, a pen (instruction unit) 14 for inputting calibration data on the display 12, and a floppy capable of inputting and storing data. The device main body 10 has a display data creation unit 18 and a keyword data correction unit 20 built therein.

The display data creation unit 18 recognizes whether it is a first-appearing new search word candidate or a second or later already appearing search word candidate.
And a display positioning unit 18B that determines the position of the new and existing search word candidates so that they can be displayed separately on the display 12, and the floppy disk (see FIG.
(Corresponding to the FD2 of the above) 16 has a function of displaying the keyword data read from the display unit 12A of the display 12 in a predetermined arrangement pattern.

The display 12 has a display unit 12
When the proofreading is instructed by the pen 14 with respect to the keyword data displayed on A, the proofreading section 12B has a recognizing section 12B which can recognize the proofreading content. Is directly transmitted to the keyword data correction unit 20, and the correction unit 20 corrects the corresponding search word candidate in the keyword data read from the floppy disk 16, and when the correction result is confirmed, the data is copied to the floppy disk. It is adapted to be written on the disc 16.

The reviewing / correcting apparatus 8 of this embodiment is programmed with various processing procedures including the processing shown in the flowchart of FIG. 4 described in detail below.
It is designed to exert editing / correcting functions.

First, in step 210, the above step 1
Search word candidates created in 16 and their readings (kana),
Keyword data including a body text (text data) is input from the floppy disk 16 to the display data creating section 18.

Next, the display data creating section 18
With respect to the input keyword data, the new / existing recognition unit 18A determines whether the search word candidate is new or has already appeared, and the display positioning unit 18B determines the display position of each of the text and the search word candidate. , By outputting the information to the display 12, the search word candidates are compared with the body text, and when the job includes a plurality of the same search word candidates, the first appearing new search word candidate and the second and subsequent words. It is displayed on the display unit 12A of the display 12 so as to be clearly distinguished from the already-existing search word candidates appearing in (step 212).

FIG. 9 schematically shows a correction (calibration) screen displayed in the above step 212 when the operation of reviewing / correcting the keyword data is performed. The text, keywords, correction results, and menus displayed at the top are divided into display areas for display.

In the body text display area, the body text portion containing the search word candidates to be reviewed / corrected is displayed.
In addition, the corresponding search word candidates are displayed with the lightness and darkness reversed so that they can be clearly distinguished from other character parts. Further, since the body text screen and the keyword screen do not correspond to each other on a line-by-line basis, by selecting a search word candidate, the relevant part of the text including the search word candidate is displayed. Therefore, since the text can be referred to when reviewing / correcting a search word candidate described later, the operation can be performed accurately.

In the keyword display area of the display screen shown in FIG. 9, search word candidates consisting of characters represented by symbols such as ○ △ are sequentially displayed one by one from the top, and these search word candidates are displayed in the keyword dictionary. Are extracted with reference to, and all the text data such as sentences are displayed in order from the beginning.

Further, when there are a plurality of the same search word candidates in the keyword data, a new search word candidate of the first appearance (in the figure, as conceptually shown in FIG. 10 is the same correction screen as in FIG. 9). Is displayed only as "new"), whereas the second and subsequent candidate words that appear (only displayed as "already" in the figure)
Are shifted by one character and are displayed separately.

The compound word is 3 in the keyword display area.
At the same time as the division position is added and displayed as shown in the line, it is expanded and displayed from the beginning into derivative words (unit nouns and their connected words) at the same time as shown in lines 4 and 5. Information about the division position and derivative words is fetched from the keyword dictionary when the keyword data is created.

Also, numbers, affixes, hiragana, transposition nouns, 1
In order to make it easier to understand words that require careful review, such as text keywords, they are highlighted.

Further, the correction result display area of FIG. 9 shows the result of correction of the search word candidates, and the menu display area shows the type of processing mode that can be selected and executed on the current display screen and the current processing mode. The mode name to be selected when moving from the screen to another mode screen is displayed.

FIG. 11 shows a concrete example of the review / correction screen schematically shown in FIG. However, the division position of the compound word is ●, and the menu display area is shown at the lower end of the screen.

After the keyword data (the text and a group of search word candidates) is displayed in step 212, whether or not each search word candidate is suitable as a search word on the correction screen of FIG. If the division position attached to the compound word is not appropriate, it is corrected using the pen 14 on the display 12. This correction is performed by instructing the position of the correction (calibration) menu and the keyword data displayed on the display 12 with the pen 14 so that the recognition content of the menu content, the target search word candidate, the correction position, etc. is recognized by the recognition unit 12B. Therefore, the keyword data correction unit 20 directly executes the correction based on the correction contents.

The review / correction device 8 of this embodiment has the following characteristic correction functions.

In the case of the same search word candidate, when a new search word candidate that appears for the first time is corrected, all of the above-mentioned already-exposed search word candidates are automatically displayed as the same corrected nouns. In addition, since the new search word candidates have been modified, the existing search word candidates that are in the same modified state are selected and individually selected, that is, other identical search word candidates are selected. It is possible to make original corrections without affecting it, and cancel the corrections made at the same time when a new search word candidate is corrected, and only the specific existing search word candidates are in the state before correction. It is also possible to return to.

In this embodiment, the review / correction work is performed as follows. First, if the search word candidates displayed in the keyword display area are new (step 214), all the search word candidates are reviewed regardless of whether or not there is another same search word candidate, and it is necessary to review them. Make any corrections. On the contrary, in the above step 214, when it is not new from the state where it is displayed distinctly on the screen and it is an already-exposed search word candidate, it is usually already corrected at the stage of new search word candidate, so it is reviewed.・ If you want to modify the already-existing search word candidates independently without modifying it,
In step 216, the revision / correction is performed as in the case of the new one (step 218). In addition, when modifying the already-existing search word candidates, if the new reference mode is selected in the menu display area, it is possible to move to the screen in which the same search word candidates that appear for the first time are displayed, and the contents can also be confirmed. .

When the search word candidates are corrected in steps 214 to 218 described above, the result is displayed in the correction result display area. That is, FIG. 9 shows the result of deleting the search word candidate on the second line and adding one character to the search word candidate on the fifth line to make it four characters. However, the correction result of the same search word candidate is not displayed except for the first appearance word, and is displayed in the already-existing search word candidate only when itself is corrected.

An example of the correction executed here and its method will be described below. The following corrections (1) to (4) and (7) can be executed in both the keyword display area and the text display area.

(1) All delete Select the all delete mode in the menu and touch the search word candidate to be deleted once with the pen. Although illustration is omitted, FIG.
The delete modes in the menu display area of include all delete and partial delete.

(2) Partial deletion Select the partial deletion mode in the menu, touch the character twice with the pen to delete one character, and the first character and the end character with the pen to delete one continuous character. Touch each time.

(3) Addition After selecting the addition mode in the menu and touching the first character or the last character of the search word candidate corresponding to the position to be added once,
Enter the character (string) to add with the pen.

(4) New addition Select the new addition mode of the menu and touch the beginning character and the ending character of the character string to be used as the keyword existing in the text in the text display area, or touch the keyword once. Add keywords by typing with the pen in the display area.

(5) Correction of compound word division position When the compound word division mode of the menu is selected and the character in the search word is touched once with the pen, the division mark (/) appears immediately before that.
Is inserted and the already displayed division mark is deleted by touching it once.

(6) Determination of second division position of compound word When a compound word has a plurality of ways of division, for example, the compound word is "physicist" and the first division position is "physical / scholar", When the second division position is "physics / person", after determining the first division position by the method of (5) above, the compound word second division mode in the menu is selected. Then, the compound word with the first division position determined in (5) above can be copied by touching it once with a pen, and the same operation as in (5) above can be performed on the copied compound word. Then, the second division position is determined.

(7) Division into different keywords Search word candidates that should be originally divided into two (excluding compound words)
, Are displayed in a linked manner, when the division mode of the menu is selected and the character in the search word candidate is touched once, it is divided into two keywords immediately before that.

At the time of the above-mentioned revision / correction, it is possible to ensure the correctness of the revision by referring to the display screen of the revision judgment support information from the revision screen. As this support information, for example, there is a compound word division position matching check list (division pattern) as shown in FIG. 12, and with reference to this list, it is possible to match division positions with other same compound words. Has become. In this division position matching check list, only compound words having a plurality of division methods are displayed as a division pattern with different division positions. From the screen of FIG. 12, it is possible to return to the review / correction screen of FIG. 9 by selecting “Return” in the menu.

As the proofreading decision support information, a proofreading history in which search word candidates and readings corrected in the work up to that point are compiled, a keyword list in which already-fixed keywords and readings are compiled, and cautions are required in reviewing An example is a caution word list that summarizes search word candidates that require.

When the correction processing of steps 214 to 218 described above is completed for all the cases of this job (step 220), it is confirmed as a keyword (step 222).

In step 222, the
Correct the search word candidates displayed on the correction screen,
After confirming it as a keyword, if the kana correction mode in the menu is selected, the kana correction screen shown in FIG. 13 can be displayed. The "expansion" shown below the compound word (corresponding to keyword 3) displayed in the keyword display area of FIG. 11 is an expansion and display of this compound word for each of its constituent words (unit nouns). It means that

On the kana correction screen shown in FIG. 13, kana is reviewed and corrected as described below (step 224).

The display screen of FIG. 13 is provided with display areas for the text, keywords, kana, and menus. The keyword display area includes steps 214 to 214.
All the keywords determined by the processing of 222 are displayed in a list in comparison with the text, and the kana display area is used to read the determined keywords (including compound words and their expanded forms) and related words of keywords other than compound words. Is displayed.
These kana are taken from the keyword dictionary when the keyword data is created, and when the keyword is fixed, the kana is automatically shaken and displayed for the keyword.

An example of kana correction executed here and its method will be described. Some of the keywords themselves are written in kana, so this kana correction can be performed in either the keyword display area or the kana display area.

Kana correction includes all deletion, partial deletion, and addition as in the case of correction of search word candidates, and this correction can be performed by substantially the same operation as in the case of search word candidates. In addition to these correction modes, this kana correction includes a new addition mode and a copy mode.

The new addition is a correction to be made when the same keyword has a plurality of readings. Select the new addition mode in the menu and display the keyword 2 and the expanded form on the 5th line on the display screen of FIG. Then, input another reading with the pen in the kana display area to add a new reading.

Copying is a correction to be made when the keyword itself is written in kana. By selecting the copy mode in the menu and touching the keyword once with the pen, the keyword is directly copied to the kana display area. And input as reading.

When the correction of the kana is completed by the above-mentioned editing / correction processing of step 224, the kana is fixed (step 226), and the fixed keyword and kana are stored in the floppy disk 16 for example. Creation is completed (step 228).

It should be noted that from the kana review / correction screen of FIG. 13, the search word candidate review / correction screen of FIG. 9 can be displayed.

According to this embodiment described in detail above, the CD-R
When creating a keyword to be stored in the OM, the reviewer instructs the keyword data displayed on the display 12 with the pen 14 so that the correction based on the correction content can be performed at the same time as the operation of the pen 14 is performed. Since it can be performed, it is possible to significantly improve the efficiency of the task of creating a keyword.

Further, even when a plurality of identical search word candidates exist in the text data, it is possible to clearly distinguish between the new search word candidate that first appears and the existing search word candidate that appears after the second search word candidate. In addition, by reviewing and correcting the new search word candidates, the same corrections are made to the already-existing search word candidates at the same time, and the already-existing search word candidates are unique without affecting other same search word candidates. Since the correction can be made, it is possible to appropriately review and correct the search word candidates, and the keyword creation work can be efficiently performed.

Further, according to the present embodiment, the search word candidates can be corrected while comparing the texts, so that the editing / correction work can be performed accurately.

Further, according to the present embodiment, after the keyword is fixed, the reading (kana) is fixed only for the fixed keyword. Therefore, it is possible to efficiently perform the keyword creation work in this respect as well. Become.

The present invention has been specifically described above, but the present invention is not limited to the above-mentioned embodiments, and various modifications can be made without departing from the scope of the invention.

For example, the reviewing / correcting device is not limited to the one shown in the above embodiment, and is the same as the instructing means for instructing the calibration of the keyword data displayed on the display. It can be arbitrarily changed as long as it has a means for executing correction based on the proofreading contents.

[0067]

As described above, according to the present invention,
When a keyword for a CD-ROM is created, the calibration based on the calibration content is executed at the same time by simply instructing the calibration on the keyword data displayed on the display by the instructing means. Since it is possible to create efficiently, it is possible to reduce the load of keyword creation work.

[Brief description of drawings]

FIG. 1 is a flowchart showing a flow of basic processing executed in the present invention.

FIG. 2 is an explanatory diagram showing a schematic configuration of a keyword creating system according to an embodiment of the present invention.

FIG. 3 is a flowchart showing a procedure for creating keyword data in the embodiment.

FIG. 4 is a flowchart showing a procedure for creating a keyword from keyword data.

FIG. 5 is an explanatory view conceptually showing a process of extracting a search word candidate with reference to a keyword dictionary.

FIG. 6 is an explanatory diagram showing an example of the contents of a keyword dictionary.

FIG. 7 is a schematic perspective view showing a keyword review / correction device of an embodiment.

FIG. 8 is a block diagram showing a schematic configuration of functional units included in the keyword review / correction device.

FIG. 9 is a conceptual diagram showing a review / correction screen for search word candidates according to the embodiment.

FIG. 10 is an explanatory diagram showing a display example of new and already-existing same search word candidates on the review / correction screen of FIG. 9.

FIG. 11 is an explanatory diagram showing a specific example of the review / correction screen for search word candidates shown in FIG. 9.

FIG. 12 is a conceptual diagram showing a compound word division position matching check list screen.

FIG. 13 is a conceptual diagram showing a Kana review / correction screen according to the embodiment.

FIG. 14 is a conceptual diagram showing a CD-ROM keyword creating method and its usage.

[Explanation of symbols]

 6 ... Keyword data creation device 8 ... Keyword review / correction device 10 ... Device body 12 ... Display 12A ... Display unit 12B ... Recognition unit 14 ... Pen 16 ... Floppy disk 18 ... Display data creation unit 18A ... New / previous recognition unit 18B ... Display positioning unit 20 ... Keyword data correction unit

Claims (3)

[Claims] 1. A search word candidate and a keyword extracted from CD-ROM text data with reference to a keyword dictionary in which nouns and their readings that may become a CD-ROM search word are registered. A keyword creating device for displaying keyword data including reading of the search word candidates read from a dictionary on a display, and editing / correcting the keyword data on the display to create a keyword suitable as a search word. Display data is created from the display data and the keyword data is displayed on the display, display data creating means, instruction means for instructing calibration of the keyword data displayed on the display, and calibration performed by the instructing means. The recognition means for recognizing the content and the keyword data based on the proofreading content recognized by the recognition means. A keyword creating device comprising: a correction unit that corrects the data. 2. The display data creating means according to claim 1, wherein when the keyword data includes a plurality of same search word candidates, a new search word candidate that appears first on the display and a second search word candidate or later appearing on the display. It has a function to distinguish and display the already-existing search word candidates, and the correction means can reflect only the result of the modification of the new search word candidates to the already-existing search word candidates and can correct only the selected previously-search word candidates. A keyword creating device having a function. 3. The display data creating means according to claim 1, wherein the display data creating means has a function of displaying a search word candidate on a display in comparison with a text containing the search word candidate. Keyword creating device.