JP5047209B2 - Error conversion pointing device and method for indicating error conversion based on conversion break position - Google Patents

Description

  The present invention relates to an erroneous conversion indication device and method for indicating an erroneous conversion of a homonym included in a sentence composed of converted character strings.

  Conventionally, there is a Japanese homonym error detection / correction method using probabilistic LSA (Non-Patent Document 1) as a method for detecting and correcting homonym error due to kana-kanji conversion.

  In this method, the presence / absence of erroneous conversion is determined using the likelihood defined based on the local appearance probability modeled by ngram of the homonym and the global appearance probability modeled by PLSA.

Takuya Sanna, September Sadamitsu, Mikio Yamamoto “Detection and correction of Japanese homonyms using probabilistic LSA”, Journal of Information Processing Society, September 2004, Vol. 45, no. 9, p. 1-9

However, since the difference in the conversion break position at the time of input in the sentence after conversion to Kanji is not considered, for example, homonyms with different conversion positions (test target words such as “practice result”, “practice trainee”), etc. ) Could not be detected.

Therefore, an object of the present invention is to point out a conversion error even when the conversion delimiter positions differ for homonyms (inspection words) in a sentence.

  The present invention provides the following solutions.

(1) contained in the sentence which is constituted by the character string after conversion, erroneous conversion a false conversion pointed device to point out, conversion delimiting position storage for storing the converted delimiter position after kanji conversion kana string Means for extracting a word to be inspected to extract the character string after the kanji conversion corresponding to the longest common part of the kana character string when the whole sentence after the kanji conversion is converted into a kana character string And a conversion delimiter position extracting unit for extracting a conversion delimiter position of each of the test target words extracted by the test target word extracting unit by referring to the conversion delimiter position storage unit , and the conversion delimiter position extracting unit If the extracted transform delimiter position differs between the inspection target word by erroneous conversion determination to determine whether there is a conversion false before Symbol inspection target word extracted by said object word extraction means Comprising a stage, and a feature that when it is determined that there is a mis-converted by the erroneous conversion determination means, points out that there is a conversion erroneous the inspection target word extracted by said object word extraction means An erroneous conversion indication device.

According to the configuration of (1), the conversion delimiter position storage unit stores the conversion delimiter position after kanji conversion of the kana character string, and the inspection target word extraction unit stores the inspection target included in the sentence after the kanji conversion. A conversion delimiter position extracting unit extracts a conversion delimiter position of each test target word, and an erroneous conversion determination unit extracts the test target word extracted when the conversion delimiter position differs between the test target words. It is determined whether or not there is an erroneous conversion. If it is determined that there is an erroneous conversion, it is pointed out that there is an erroneous conversion in the word to be examined .

Accordingly, the erroneous conversion indication device can indicate a conversion error even when the conversion delimiter positions are different for the homonyms (test target words) in the sentence.

(2) The inspection target word extracting unit includes a morpheme dividing unit that divides the sentence after the kanji conversion into morphemes, a kana conversion unit that converts the morphemes divided by the morpheme dividing unit into a kana character string, and a kana conversion unit. comprising a pseudonym morpheme connecting means for connecting the morphemes converted into kana character string, and the same connecting morpheme extraction means for extracting the same connecting morphemes from the connecting morpheme is a morpheme linked by pseudonym morpheme connecting means, by The erroneous conversion indication device according to (1), wherein a character string corresponding to the extracted connected morpheme is extracted as an inspection target word from the sentence after the Kanji conversion.

  According to the configuration of (2), the inspection target word extraction unit divides the sentence after the kanji conversion into morphemes, converts the divided morphemes into kana character strings, and connects the morphemes converted into kana character strings. Then, the same connected morpheme is extracted from the connected morpheme that is the connected morpheme, and the character string corresponding to the extracted same connected morpheme is extracted from the sentence after the kanji conversion as the inspection target word.

Here, in the invention of (1), the misconversion indication device extracts the longest common part as the inspection target word when the entire sentence after the kanji conversion is converted into the kana character, so the inspection target word is too long. Although there is a possibility that an erroneous conversion indication cannot be made accurately, the configuration (2) enables extraction of an inspection target word with an accurate length, so that an erroneous conversion indication can be made by an optimum method.

(3) the conversion break position storage means stores in association with the kana character string and the character string after kanji conversion and the conversion delimiter position, wherein the conversion delimited extracted by prior Symbol conversion break position extracting means If the positions are different between the inspection target word, before SL comprises an erroneous conversion flag activation means activates erroneously conversion flag to indicate that there is a possibility of inspection target word in erroneous conversion, the erroneous conversion determination means, if the erroneous conversion flag is activated, erroneous conversion pointed device according to, characterized in that you determine whether there is a conversion erroneous said object word the extracted (1) or (2).

According to the configuration of (3), wherein the conversion delimiting position storage means, in association with the kana character string and the character string after kanji conversion and the conversion delimiter position, erroneous conversion flag activation means, before Symbol conversion when the conversion delimiter position extracted by the break position extracting means is different between the test subject word starts erroneous conversion flag to indicate that there is a possibility of erroneous converted before Symbol inspection target word. Furthermore, the erroneous conversion determination unit, when the erroneous conversion flag is activated, it determines whether there is a conversion erroneous the extracted said object language.

Thus, the erroneous conversion pointed device, whether based on the conversion delimiter position extracted by converting delimiting position extracting means, only if there is a possibility of converting erroneous inspection target word, it is converted erroneous inspection target word Therefore, it is possible to narrow down the determination target for erroneous conversion. Therefore, when the erroneous conversion determination process is unnecessary, the erroneous conversion indication device does not need to perform the process, and thus the processing efficiency of the erroneous conversion determination process can be increased.

(4) computer is included within a sentence constituted by the character string after conversion, a false conversion point out how to point out erroneous conversion, storage for storing the converted delimiter position after kanji conversion kana string Step and test word extraction for extracting the kanji converted character string corresponding to the longest common part of the kana character string as a test target word when the whole sentence after kanji conversion is converted into a kana character string steps and, on the basis of the stored the converted delimiter position, and conversion break position extracting a conversion delimiter position of each of the inspection target words extracted by the inspection target word extracting means, converting delimited the extracted When the position is different between the inspection target words, it is determined that there is an erroneous conversion determination step for determining whether or not the extracted inspection target word has an erroneous conversion, and that there is an erroneous conversion. Case, pointed out steps and, Yusuke erroneous transformation point out methods wherein Rukoto to point out that there is a conversion erroneous the extracted said object language.

  According to the structure of (4), there exists an effect similar to the erroneous conversion indication apparatus of (1).

(5) In the step of extracting as the inspection target word, divide the sentence after the kanji conversion into morphemes, the divided the morphemes into a kana character string, connecting the morphemes converted into kana character string extracts identical coupling morphemes from the connecting morphemes are concatenated the morpheme extracting from the text after the kanji conversion, the extracted character string corresponding to the same connection morphemes as an inspection target word (4) The erroneous conversion indication method according to (4).

  According to the structure of (5), there exists an effect similar to the erroneous conversion indication apparatus of (2).

(6) In the storing step, the same kana character string, the character string after kanji conversion and the conversion delimiter position are stored in association with each other, and the extracted conversion delimiter position is different between the inspection target words the front SL has an activation step of activating erroneous conversion flag to indicate that there is a possibility of inspection target word in erroneous conversion, in the erroneous conversion determination step, when the erroneous conversion flag is activated, the extracted It has been erroneously converted pointed method according to the inspection target word in erroneous conversion whether the characterized determine Teisu Rukoto there (4) or (5).

  According to the structure of (6), there exists an effect similar to the error conversion indication apparatus of (3).

According to the present invention, the erroneous conversion indicating device can indicate a conversion error even when the conversion delimiter positions are different for the homonyms (test target words) in the sentence.

It is a functional block diagram showing the functional configuration of the erroneous conversion pointed equipment of the present embodiment. Is a diagram showing an erroneous conversion pointed equipment hardware configuration of the present embodiment. Is a flowchart at the time the character input processing of erroneous conversion pointed equipment of the present embodiment. Is a flowchart of the erroneous conversion report processing of erroneous conversion pointed equipment of the present embodiment. It is a figure which shows the example of the input character string at the time of input. It is a figure which shows the example of the text after conversion. It is a figure which shows the conversion unit character string table. It is a figure which shows a conversion position table. It is a figure which shows a test object word extraction result table. It is a figure which shows a conversion position extraction result table. It is a figure which shows an erroneous conversion flag starting result table. It is a figure which shows the example of an incorrect conversion indication. Server is a diagram showing further comprising a converting position storage means. It is a flowchart of a test object word extraction process. It is a figure which shows the process of a morphological analysis, kana conversion, and the connection of 2 adjacent morphemes. It is a figure which shows the same kana search table.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

[Functional configuration of erroneous conversion indication device 1]
FIG. 1 is a functional block diagram showing a functional configuration of an erroneous conversion indication device 1 according to an embodiment of the present invention.

  The erroneous conversion indication device 1 includes an identical kana character string extraction unit 110, a conversion position storage unit 120, a sentence reception unit 130, an inspection target word extraction unit 140, and an erroneous conversion indication unit 150.

  Further, the erroneous conversion indication unit 150 includes a conversion position extraction unit 151, an erroneous conversion flag activation unit 152, and an erroneous conversion determination unit 153. The conversion position extraction unit 151 refers to the conversion position storage unit 120 and extracts the conversion position.

  The same kana character string extraction unit 110 extracts the same kana character string from the kana character string constituting the sentence, and converts the extracted conversion position of the same kana character string after the kanji conversion into the conversion position storage unit 120 (in FIG. 8). It stores in a conversion position table (to be described later).

  The sentence accepting unit 130 accepts input of a sentence after conversion to kanji (converted sentence described later in FIG. 6), the inspection target word extracting unit 140 extracts the inspection target word, and the erroneous conversion indication unit 150 extracts If there is an incorrect conversion in the checked word, point out that there is an incorrect conversion.

  At that time, only when the erroneous conversion flag starting means 152 provided in the erroneous conversion indicating means 150 starts the erroneous conversion flag (when there is a possibility of erroneous conversion based on the conversion position), the erroneous conversion determining means 153 , Ngram + PLSA (Takuya Sanna, Sadamitsu September, Mikio Yamamoto “Detection and correction of Japanese homonyms using stochastic LSA”, Information Processing Society of Japan, September 2004, Vol. 45, No. .9, p.1-9), the presence or absence of erroneous conversion is determined.

[Hardware configuration of erroneous conversion indication device 1]
FIG. 2 is a diagram illustrating a hardware configuration of the erroneous conversion indication device 1 of the present embodiment. As shown in FIG. 2, a central processing unit (CPU) 210 (a plurality of CPUs such as CPU 220 may be added in a multiprocessor configuration), a bus line 100, a communication I / F (I / F) Interface) 230, main memory 240, BIOS (Basic Input Output System) 250, I / O controller 260, hard disk 270, optical disk drive 280, and semiconductor memory 290. The hard disk 270, the optical disk drive 280, and the semiconductor memory 290 are collectively referred to as a storage device 310.

  The control unit 200 is a part that controls the misconversion indication device 1 in an integrated manner, and by appropriately reading and executing various programs stored in the hard disk 270, the control unit 200 cooperates with the hardware described above, and relates to the present invention. Various functions are realized.

  The communication I / F 230 is a network adapter when the erroneous conversion indication device 1 transmits / receives information to / from another device via a network.

  The BIOS 250 records a boot program executed by the CPU 210 when the erroneous conversion indication device 1 is started, a program depending on the hardware of the erroneous conversion indication device 1, and the like.

The I / O controller 260 can be connected hard disk 270, the optical disk drive 28 0及 beauty semiconductor memory 290 or the like of the storage device 310.

  The hard disk 270 stores various programs for causing the hardware to function as the erroneous conversion indication device 1, a program for executing the functions of the present invention, a data table to be described later, and the like. Note that the erroneous conversion indication device 1 can also use a hard disk (not shown) separately provided as an external storage device.

As the optical disk drive 280, for example, a DVD-ROM drive, a CD-ROM drive, a DVD-RAM drive, and a CD-RAM drive can be used. In this case, the optical disc 300 corresponding to each drive is used. A program or data can be read from the optical disk 300 by the optical disk drive 280 and provided to the main memory 240 or the hard disk 270 via the I / O controller 260.

  The computer in the present invention refers to an information processing apparatus including a storage device, a control unit, and the like, and the erroneous conversion indication device 1 includes an information processing device including the storage device 310, the control unit 200, and the like. .

  In the above example, the erroneous conversion indication device 1 has been mainly described. However, the functions described above can also be realized by installing a program in a computer and operating the computer as a server device. Therefore, the function realized by the erroneous conversion indication device 1 described as one embodiment in the present invention is executed by executing the above-described method by the computer or by introducing the above-described program into the computer. This is also possible.

[Character input process flowchart]
FIG. 3 is a flowchart of the character input process.

First, in step S1, the control unit 200 receives an input of a kana character string. Specifically, the control unit 200, an input device such as a keyboard provided in erroneous conversion pointed apparatus 1 de (not shown), receives an input of the kana character string.

In step S2, the control unit 200 performs kanji conversion. Specifically, the control unit 200 converts the kana character string received in step S1 into a kanji character string including kanji.

  Here, a specific example of the conversion process in step S1 and step S2 of FIG. 3 will be described with reference to FIGS.

  FIG. 5 is an example of an input character string at the time of input, and kanji conversion is performed each time a box portion (“Yatsuto”, “Renshu”, etc. in FIG. 5) is input. The kanji conversion is performed at a timing when a user inputting a character presses a key (not shown) corresponding to the kanji conversion of the input device.

  FIG. 6 is an example of the sentence after conversion, and kanji conversion is performed corresponding to the encircled portion of FIG. More specifically, when “Yatsuto” is input and a key corresponding to Kanji conversion of the input device (hereinafter referred to as “Kanji conversion key”) is pressed, “Yatsuto” is converted to “Renshu”. "Is input and the Kanji conversion key is pressed, it is converted to" practice ", and the same is repeated thereafter.

  Returning to FIG. 3, in step S3, the control unit 200 creates a conversion unit character string table (FIG. 7). This conversion unit character string table is created in a predetermined area of the main memory 240.

  Here, the conversion unit character string table will be described with reference to FIG. In this conversion unit character string table, the character string before conversion and the character string after conversion are associated with each Kanji conversion unit, and at the timing when the Kanji conversion key is pressed, the character string before conversion is displayed at the bottom. The character string and the converted character string are added and stored.

  For example, when the box “Seika” in FIG. 5 is input and the Kanji conversion key is pressed, “Seika” is added to the “Before conversion” field, and “Result” is added to the “After conversion” field. Stored.

Returning to FIG. 3, in step S <b> 4, the control unit 200 determines whether or not the sentence creation is finished. When this process is determined to be YES, control unit 200 moves the process to step S5, and when it is determined to be NO, the process returns to step S1. A specific method for determining the end of sentence creation is, for example, that the control unit 200 detects that the file in which the sentence has been created is closed. When the file in which the sentence is created is closed, the file in which the converted sentence shown in FIG. 6 is stored in the hard disk 270. Further, a file composed only of the kana character string before conversion (corresponding to the example shown in FIG. 5 that omits the enclosing part and does not perform kanji conversion) is stored in the main memory 240.

  Here, a method of creating a file composed only of kana character strings before conversion is based on the kana character strings stored in the uppermost column of the “before conversion” column of the above-described conversion unit character string table (FIG. 7). For example, a kana character string is sequentially extracted from the bottom to the bottom and expanded into a file.

Returning to FIG. 3, in step S5, the control unit 200 extracts the same kana character string. Specifically, the control unit 200 searches for and extracts the same kana character string from the file composed only of the kana character string before conversion described above.

  Here, there are a plurality of character strings to be extracted, such as “Ryushu Seika”.

  In step S6, the control unit 200 creates a conversion position table (FIG. 8).

  Here, the conversion position table will be described with reference to FIG. This conversion position table is assigned to a predetermined area of the hard disk 270.

  In FIG. 8, for convenience of explanation, it is limited to “relenty”, but in reality, all the character strings extracted in step S5 of FIG. 3 are written in the conversion position table.

Specifically, the control unit 200 searches the “before conversion” field in the conversion unit character string table (FIG. 7), combines two consecutive kana character string data, and all the character strings extracted in step S5. Compare with As a result of the comparison, if there is a matched character string, the control unit 200 determines the conversion position of this character string (after what character the Kanji conversion key was pressed), and for this character string, the conversion position Data is stored in the “input word” column, “after conversion” column, and “conversion position” column of the table (FIG. 8).

  Referring to the “before conversion” column of the conversion unit character string table (FIG. 7), since “Ryushu” and “Seika” are two consecutive, “Ryushu Seika” is converted into the conversion position table (FIG. 8). ) Is stored in the data format in the input word field. Further, “practice results” obtained by combining “practice” and “results” corresponding to “renshu” and “seika” are stored in a data format in the converted column. Regarding the conversion position, “Renshu” is five characters from the left of “Renshuiseika”, so “after the fifth character” is stored in the conversion position column in the data format.

  Here, storing in the data format means storing data such as character codes. To give a specific example, the unit characters constituting the character strings “Ryushu Seika” and “Practice Outcome” Each character code is stored. For “after the fifth character”, numerical data such as “5” is stored.

  In the same way for “Ryushusei” and “Ka”, the “input word” field, “after conversion” field, and “conversion position” field in the conversion position table are respectively “Renesas Seika” and “Practice Student Allowed”. ”And“ after the seventh character ”are stored in the data format.

  When the process of step S6 in FIG. 3 is completed, the control unit 200 ends the character input process.

[Flow chart of erroneous conversion indication processing]
FIG. 4 is a flowchart of the erroneous conversion indication process.

In step S11, the control unit 200 receives a text input. Specifically, the control unit 200 reads from the hard disk 270 a file storing the converted text shown in FIG. 6 and develops the text data of the read file in the main memory 240.

Returning to FIG. 4, in step S <b> 12, the control unit 200 performs the inspection target word extraction. Specifically, the control unit 200 converts the entire sentence after the kanji conversion shown in FIG. 6 into a kana character, and uses the character string after the kanji conversion corresponding to the longest common part of the converted kana character as an inspection target word. Extract.

Explaining in detail, when “practice results finally came out… I think this is also a trainee…” converted to kana characters, I'm sure it'll be ... " Furthermore, by extracting the longest common part character string from “It was a great deal, I guess it ’s a great deal”, you can get “Renshuiseika”. Furthermore, since the kanji converted character strings corresponding to “Renshuiseika” are “practice results” and “practice trainees”, the test target words are “practice results” and “practices trainees”. Become. When the test target word is extracted , the control unit 200 associates the extracted test target word with the kana character string and creates a test target word extraction result table (FIG. 9). This inspection target word extraction result table is created in a predetermined area of the main memory 240.

  The inspection target word extraction result table will be described with reference to FIG. This inspection target word extraction result table is a table representing the correspondence between the same kana character string and the inspection target word.

  According to this table, the test target words corresponding to “Renshu Seika” are “practice result” and “practicable”.

Returning to FIG. 4, in step S13, the control unit 200 performs conversion position extraction. Specifically, the control unit 200 extracts the conversion position for the inspection target word with reference to the conversion position table shown in FIG.

Here, the conversion position extraction method will be described. The control unit 200 uses the data stored in the “test target word” column of the test target word extraction result table shown in FIG. 9 as a key, and the conversion shown in FIG. The “post-conversion” column of the position table is searched to extract the conversion position.

  For example, when “practice result” stored in the “test target word” column of the test target word extraction result table shown in FIG. 9 is used as a key, “after the fifth character” is extracted as the conversion position. Similarly, when “Practice Student Allowed” is used as a key, “after the seventh character” is extracted as the conversion position.

  The extracted data is stored in the “conversion position” column of the conversion position extraction result table (FIG. 10).

  The conversion position extraction result table will be described with reference to FIG. This conversion position extraction result table is a table in which the “conversion position” column is added to the inspection target word extraction result table shown in FIG. 9, and “conversion position” is added to the inspection target word extraction result table stored in the main memory 240. Created by merging fields.

Returning to FIG. 4, in step S <b> 14, the control unit 200 activates an erroneous conversion flag. Specifically, the control unit 200 compares the conversion positions corresponding to the plurality of inspection target words stored in the “inspection target word” column in the conversion position extraction result table (FIG. 10), and when the conversion positions are different from each other. Activates (turns on) an erroneous conversion flag (a flag indicating that there is a possibility of erroneous conversion due to a difference in conversion position), and in the same case, does not activate the erroneous conversion flag (leaves it off).

In the embodiment, since the conversion positions in the conversion position extraction result table (FIG. 10) differ between “practice results” and “practice trainees allowed”, the control unit 200 sets the erroneous conversion flag to “ON”.

  Whether or not the erroneous conversion flag is activated is stored in the erroneous conversion flag activation result table shown in FIG.

  The erroneous conversion flag activation result table will be described with reference to FIG. This erroneous conversion flag activation result table is a table in which an “error conversion flag” column is added to the conversion position extraction result table shown in FIG. 10, and “error conversion flag” is added to the conversion position extraction result table stored in the main memory 240. "" Is created by combining the fields.

  Here, when the erroneous conversion flag is activated, “ON” is stored in the “erroneous conversion flag” field, and when not activated, “OFF” is stored. “On” may be “1” of numerical data, and “off” may be “0”.

  By referring to the erroneous conversion flag activation result table in FIG. 11, it can be determined whether or not there is a possibility of erroneous conversion due to a difference in conversion position for the inspection target word. In the example of FIG. 11, since the erroneous conversion flag is on, it can be seen that there is a possibility of erroneous conversion in “practice result” or “trainee acceptable”.

  Returning to FIG. 4, in step S <b> 15, the control unit 200 performs erroneous conversion determination. This erroneous conversion determination processing (and erroneous conversion indication processing in step S16) is executed only when the erroneous conversion flag is activated in step S14.

In the erroneous conversion determination process, the control unit 200 determines which of “practice result” or “trainee acceptable” is an erroneous conversion.

  First, a method for determining whether or not there is an erroneous conversion in “Practice trainee” will be described.

Specifically, the control unit 200 uses “practice results” and “trainees allowed” as a homonym list, and uses the ngram + PLSA method (Takuya Sanna, Sadamitsu Seketsu, Mikio Yamamoto “Japanese homonyms using probabilistic LSA”). "Detection / correction of word errors", Journal of Information Processing Society of Japan, September 2004, Vol. 45, No. 9, p. 1-9), "Practice results" in the sentence read in step S11 Likelihood defined based on global appearance probability modeled by PLSA and local appearance probability modeled by ngram (hereinafter referred to as “likelihood”), and likelihood of “trainee acceptable” And calculate. Next, as an error determination calculation, the control unit 200 calculates the logarithm of the calculated likelihood ratio, and when the calculated logarithm value exceeds a certain threshold, an erroneous conversion is made to “trainee acceptable”. Judge that there is.

As a calculation method, d = log { (likelihood of “practice trainee”) / (likelihood of “practice result”) } is calculated. It can be determined that there is a conversion. However, when the judgment condition is d <0, the likelihood of “practice results” is almost the same as the likelihood of “trainees allowed” and is slightly higher (for example, the likelihood of “practice results” is 50, Even if the likelihood of “possible” is 49), it is determined to be erroneous conversion. Therefore, the value of the threshold value “0” in the determination condition d <0 is appropriately adjusted on condition that it is a negative number. (E.g., d <-0.5).

On the other hand, a method for determining whether or not there is an erroneous conversion in the “practice result” is the same method as described above, and d = log { (likelihood of “practice result”) / (likelihood of “practice student”). ) } Is calculated.

  In step S16, the control unit 200 performs misconversion indication. This erroneous conversion indication process is executed only when it is determined in step S15 that there is an erroneous conversion.

Specifically, when it is determined that there is an erroneous conversion in one of the inspection target words, the control unit 200 points out that the other inspection target word is correct. As a specific example, as shown in FIG. 12, when it is determined that there is an erroneous conversion in “practice trainee”, “practice result” is pointed out with respect to “practice trainee”. Furthermore, by displaying the content shown in FIG. 13 on a display device (not shown) included in the erroneous conversion indication device 1, the user can recognize that there is an erroneous conversion.

  The user who recognizes that there is an erroneous conversion determines whether or not to change “Practice Student Allowed” to “Practice Result”, and in the case of changing, a change confirmation button (not shown) provided in the erroneous conversion indication device 1 is shown. )) To confirm the change.

  When the process of step S16 in FIG. 4 ends, the control unit 200 ends the erroneous conversion indication process.

  By performing the above processing, a conversion error can be pointed out even if the conversion position is different for the homonym (inspection word) in the sentence.

  Furthermore, since it is determined whether or not there is an erroneous conversion in the inspection target word only when the erroneous conversion flag is activated, it is possible to narrow down the erroneous conversion determination target and increase the processing efficiency of the erroneous conversion determination process. Can do.

  As described above, in the embodiment, the conversion position table (FIG. 8) is provided in the erroneous conversion indication device 1, but is not limited thereto. For example, as shown in FIG. 13, the server 400 is provided with a conversion position table, and the conversion position table in the server 400 is created while communicating with the erroneous conversion indication device 1 (the process in step S6 in FIG. 3 is performed). Further, it may be referred to (the processing in step S12 and step S13 in FIG. 4 is performed).

Further, in the inspection target word extraction process in step S12 of FIG. 4 in the embodiment, the control unit 200 converts the entire sentence after the kanji conversion shown in FIG. 6 into a kana character, and sets the longest common part of the converted kana character. The corresponding character string after Kanji conversion is extracted as the inspection target word, but is not limited to this. The processing from step S31 to step S36 shown in FIG. 14 described later may be performed in place of the processing of step S12 in FIG.

  Hereinafter, the inspection target word extraction process will be described with reference to FIG.

In step S31, the control unit 200 performs morphological analysis. Specifically, the control unit 200 performs morphological analysis on the text data of the file read in step S11 of FIG. In step S32, the control unit 200 performs kana conversion, and in step S33, the control unit 200 connects adjacent two morphemes.

  Here, morphological analysis is one of the basic techniques of natural language processing using a computer such as a computer, and it includes knowledge of the target language grammar (gathering of grammar rules) and dictionary (word list with information such as parts of speech). ) As an information source, and a sentence written in a natural language is divided into columns of morphemes (roughly speaking, the smallest unit having meaning in a language).

  A specific example of morphological analysis will be described with reference to FIG. 15 including kana conversion and connection of adjacent two morphemes. FIG. 15 is a diagram illustrating a process of morphological analysis and a connection between adjacent two morphemes. For example, when morphological analysis is performed on “practice results”, it is divided into “practices” and “results” (circle numeral 1 in FIG. 15).

  Next, “practice” and “result” which are the divided morphemes are converted to kana and converted to “rensyu” and “seika”, respectively (circle numeral 2 in FIG. 15).

  Next, two adjacent morphemes (two morphemes adjacent to each other among the divided morphemes) are connected. Since “Ryushu” and “Seika” are adjacent to each other, they are connected to each other, "Seikai" (circled number 3 in FIG. 15).

  For “Practice Student Allowed”, morphological analysis is performed (circled number 1 in FIG. 15), divided into “Practice Student”, “OK”, and “DA”, and kana conversion is performed (circled number in FIG. 15). 2) Convert them into “Renshuisei”, “ka” and “da”, respectively, and connect two adjacent morphemes to make “Renshuiseika” and “Kada”.

  The processes of morphological analysis, kana conversion, and connection of adjacent two morphemes are temporarily stored in the main memory 240 as data by the control unit 200.

  Returning to FIG. 14, in step S <b> 34, the control unit 200 creates the same kana search table (see FIG. 16). Specifically, the correspondence relationship between each of the morphemes including adjacent two morphemes connected in FIG. 15 (hereinafter referred to as “connected morphemes”) with the character string in the text of the file read in step S11. And the correspondence relationship with the character string before being divided into morphemes are stored in the same kana search table assigned to a predetermined area of the hard disk 270.

  The same kana search table will be described with reference to FIG. As described above, the same kana search table includes a morpheme including a connected morpheme only of kana (morpheme including a connected morpheme (only kana)), a character string in a sentence (a morpheme including a connected morpheme (in a sentence)), and This is a table for storing a correspondence relationship between character strings before being divided into morphemes (before being divided into morphemes).

  This same kana search table is created based on data of morpheme analysis, kana conversion, and connection of adjacent two morphemes stored as data in the main memory 240.

  For example, the correspondence relationship in the same kana search table of FIG. 16 will be described for “rensyu seika” as the connected morpheme in FIG. 15, corresponding to the character strings “practice results” and “trainees allowed” in the sentence. Before dividing into morphemes, it corresponds to “practice results” and “practicable”.

Returning to FIG. 14, in step S <b> 35, the control unit 200 performs the same kana search. Specifically, the control unit 200 refers to the same kana search table (FIG. 16), refers to the “morpheme including linked morphemes (only kana)” column, and searches for the same kana (two or more same characters). If there is a column, the same pseudonym exists).

  For example, in the same kana search table of FIG. 16, “renyu seika” is extracted as the same kana as the search result.

Returning to FIG. 14, in step S <b> 36, the control unit 200 performs the inspection target word extraction. Specifically, the control unit 200 extracts a morpheme (in a sentence) including a connected morpheme corresponding to a morpheme including a connected morpheme searched as the same kana (only kana) from the same kana search table of FIG.

  For example, since the retrieved same kana is “Renshuiseika”, the row data corresponding to “Renshuiseika” (“Practice result” and “Practice trainee”) is extracted and the above-mentioned inspection is performed. A target word extraction result table (FIG. 9) is created.

  When the process of step S36 ends, the control unit 200 performs the process of step S13 in FIG.

  As mentioned above, although embodiment of this invention was described, this invention is not restricted to embodiment mentioned above. The effects described in the embodiments of the present invention are only the most preferable effects resulting from the present invention, and the effects of the present invention are limited to those described in the embodiments of the present invention. is not.

DESCRIPTION OF SYMBOLS 1 Error conversion indication apparatus 110 Same kana character string extraction means 120 Conversion position storage means 130 Sentence reception means 140 Inspection object word extraction means 150 Error conversion indication means

Claims (6)

An erroneous conversion indication device for indicating an erroneous conversion included in a sentence composed of converted character strings,
A conversion break position storage means for storing a conversion delimiter position after kanji conversion kana character string,
When the entire sentence after the kanji conversion is converted into a kana character string, a test target word extracting unit that extracts the character string after the kanji conversion corresponding to the longest common part of the kana character string as a test target word, and
A conversion delimiter position extracting unit that extracts a conversion delimiter position of each of the test target words extracted by the test target word extracting unit by referring to the conversion delimiter position storage unit ;
When the conversion delimiter positions extracted by the conversion delimiting position extracting means is different between the inspection target word, to determine whether there is a conversion false before Symbol inspection target word extracted by said object word extraction means An erroneous conversion determination means ,
When the erroneous conversion determining means determines that there is an erroneous conversion, it points out that there is an erroneous conversion in the test target word extracted by the test target word extracting means. Conversion indication device. The inspection target word extraction means includes
Morpheme dividing means for dividing the sentence after the kanji conversion into morphemes;
A kana conversion means for converting the morpheme divided by the morpheme dividing means into a kana character string;
Kana morpheme linking means for linking morphemes converted into kana character strings by the kana conversion means;
The same connected morpheme extracting means for extracting the same connected morpheme from the connected morphemes that are connected by the kana morpheme connecting means,
2. The erroneous conversion indication device according to claim 1, wherein a character string corresponding to the extracted connected morpheme is extracted as an inspection target word from the sentence after the Kanji conversion. The conversion break position storage means stores in association with the conversion delimiting position and kana character string and the character string after kanji conversion,
False front Symbol converter break position extracting means and said converting delimiter position extracted by the when different between the inspection target word, activates erroneously conversion flag to indicate that there is a possibility of erroneous converted before Symbol inspected word Conversion flag activation means ,
The erroneous conversion determination unit, when the erroneous conversion flag is activated, according to claim 1 or 2, characterized in the Turkey to determine whether there is a conversion erroneous said object word the extracted Incorrect conversion indication device. An erroneous conversion indication method in which a computer indicates an erroneous conversion included in a sentence composed of converted character strings,
A storage step for storing the conversion delimiter position after the kanji conversion of the kana character string;
A test target word extraction step of extracting the kanji converted character string corresponding to the longest common part of the kana character string as a test target word when the entire sentence after the kanji conversion is converted into a kana character string;
On the basis of the stored the converted delimiter position, and conversion break position extracting a conversion delimiter position of each of the inspection target word which is the extraction,
An erroneous conversion determination step for determining whether or not there is an erroneous conversion in the extracted inspection target word when the extracted conversion break position differs between the inspection target words;
When it is determined that there is the erroneous conversion, an indication step for pointing out that there is an erroneous conversion in the extracted word to be examined;
Erroneous transformation point out methods wherein Rukoto to have a. In the inspection target word extraction step,
Divide the kanji-converted sentence into morphemes,
Converting the divided morpheme into a kana character string;
Concatenate the morphemes converted to kana strings,
Extracting the same connected morpheme from the connected morphemes that are the connected morphemes,
5. The erroneous conversion indication method according to claim 4, wherein a character string corresponding to the extracted connected morpheme is extracted as an inspection target word from the sentence after the Kanji conversion. In the storing step, the same kana character string, the character string after the kanji conversion and the conversion delimiter position are stored in association with each other,
Having an activation step of the extracted said conversion delimiter position said to vary between the inspection subject word starts erroneous conversion flag to indicate that there is a possibility of erroneous converted before Symbol inspection target word,
Wherein the erroneous conversion determination step, when the erroneous conversion flag is activated, according to whether there is a conversion erroneous the extracted the inspection target word to claim 4 or 5, wherein the stamp Teisu Rukoto Of misconversion indication.

Images