JPH08272792A - Device and method for character processing - Google Patents

Abstract

PURPOSE: To suppress misconversion due to an inadequate sectioning position and improve the conversion rate by judging whether or not a sectioning position in a character string is proper and holding information on that, and learning the specified sectioning position with this held information. CONSTITUTION: A KANA(Japanese syllabary)-KANJI(Chinese character) conversion process part 201 converts an inputted KANA character string into a KANJI-KANA-mixed character string according to a sectioning position that is considered to be adequate as the result of an analysis by a sectioning position specification data holding part 202. When an undesired conversion result is obtained, an operator specifies a desired sectioning position in the input character string and it is judged whether or not the sectioning position is grammatically correct by looking up the character strings before and after the sectioning position in a word dictionary. When it is judged that the sectioning position is correct, the sectioning position specified by the operator is stored as new sectioning position specification data in the sectioning position specification data holding part 202 while made to correspond to the input character string, and the conversion result based on the specified sectioning position is outputted.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character processing device and a method thereof, and more particularly to a character processing device and a method thereof for performing kana-kanji conversion.

[0002]

2. Description of the Related Art Conventionally, in a character processing device for processing Japanese, a so-called kana-kanji conversion function for converting kana characters input by an operator into appropriate kanji characters is essential.

Kana-Kanji conversion processing in a conventional character processing apparatus will be described below with reference to FIG.

FIG. 10 is a flowchart showing a kana-kanji conversion process in a conventional character processing device. In FIG. 10, first, in step S901, the operator inputs a kana character string or the like. Then, in step S902, kana-kanji conversion is performed on the input kana character string to convert it into a kana-kana mixed character string. At this time, if delimiter position designation data corresponding to the input character string exists in the predetermined delimiter position designation data area, the delimiter position in the kana character string is determined by referring to the delimiter position designation data. And convert it.

Subsequently, in step S903, the kana-kana mixed character string or the like obtained as described above is output as a conversion result. Then, in step S904, the operator determines whether or not the obtained conversion result is the desired conversion result. If it is the desired conversion result, the process proceeds to step S910 to finalize the conversion result and ends, but if it is not the desired conversion result, step S90.
Go to 5.

In step S905, the operator designates a desired delimiter position for the input character string. Then, the process proceeds to step S906, and the input character string is converted into a kanji-kana conversion mixed sentence or the like so that the word is preferentially divided at the delimitation designated position designated by the operator. In step S907,
The kana-kana mixed character string etc. obtained as described above,
Output as the conversion result. Then, in step S908, the conversion result obtained in step S907 is determined, and then in step S909, step S90
The delimiter position designated by the operator in 5 is stored as the delimiter position designation data in association with the input character string.

Next, an example of kana-kanji conversion in a conventional character processing device will be described with reference to FIGS. Hereinafter, consider the case where "here kimono wa" is used as the input character string.

First, FIG. 11 shows an example of conversion when there is no delimiter position designation data corresponding to the input character string "here kimono wo". In this case, even if the delimiter position designation data area is referred to when performing kana-kanji conversion, there is no corresponding data, so it seems to be most appropriate based on the result of certain morphological analysis and likelihood information. "Here / Kimono" is output as the conversion result.

In the following, in the character string of the kana-kanji conversion result, "/" indicates a position recognized as a delimiter position,
It is not actually output as a conversion result character string.

In FIG. 12, the conversion result “here / Kimono” output as shown in FIG. 13 is not the desired conversion result. Here is an example: For example, "here"
It is assumed that the space between "and footwear" is newly specified as a delimiter position. In this case, kana-kanji conversion is performed with the designated delimiter position as the highest priority. Therefore, the desired conversion result "here / footwear" can be obtained. Then, the newly designated delimiter position is stored in the delimiter position designation data area in the format of "here / footwear", for example, so that the desired conversion result can be obtained at the next conversion.

FIG. 13 shows a conversion example when the same sentence is input again after the processing shown in FIG. in this case,
In the delimiter position designation data area, delimiter position designation data corresponding to the input character string “here kimono” is present, so that the delimiter designated position is prioritized in the conversion process. As a result, the desired conversion result “here / footwear” is obtained as the first candidate.

As described above, in the conventional character processing apparatus, when there are a plurality of delimiter positions that make sense as a sentence for the input character string, the operator designates one of the delimiter positions. The conversion result desired by the operator can be obtained by performing the conversion process based on the designated delimiter position.

As described above, in the conventional character processing device, an appropriate delimiter position for an input character string is determined by referring to delimiter position designation data for Kana-Kanji conversion, and Kana-Kanji conversion is performed.

Further, in the conventional kana-kanji conversion processing by the method described above, in the delimiter position designation data area, a delimiter position corresponding to the input character string, that is, a delimiter which is established as a sentence with respect to the input character string. This may happen if there are multiple locations. In this case, when the operator designates one of the plurality of delimiter positions, performs a desired conversion, and stores the designated delimiter position, the same character string is input again. Converted at the delimiter position specified last time. As a result, the conversion rate in kana-kanji conversion could be improved.

As described above, the method of performing kana-kanji conversion by designating the delimiter position has been quite effective for improving the conversion rate.

[0016]

However, in the kana-kanji conversion processing in the above-mentioned conventional character processing device,
There were the following problems. This will be described in detail with reference to FIGS.

FIG. 14 shows an example in which the character string "josha" is input to perform kana-kanji conversion. In this case, since there is no particular delimiter position data in the delimiter position data area, the kana-kanji conversion result of "ride" is obtained. However, it is assumed that the input character string is an erroneous input by the operator, and the original hope of the operator is to input “joushi” to obtain the conversion result of “boss”.

In this case, the operator does not re-input the input character string "joushi" as shown in FIG. 15, but performs kana-kanji conversion for the erroneously input character string "josho" by changing the delimiter position. Is common. For example, the operator specifies the delimiter position between “jo” and “ya”,
After that, kana-kanji conversion is performed to obtain the conversion result “boss / ya”. Then, by confirming this "boss / ya", new break position data "joushi / ya" is stored in the break position designation data area. Then, by deleting “ya” after the confirmation, the conversion result “boss” desired by the operator can be obtained.

The above-described example of delimiter position designation "Joshi / ya" shown in FIG. 16 is not grammatically correct even if a desired conversion result is obtained. As described above, most of the delimiter position designations at positions that are not grammatically correct occur due to erroneous input by the operator. Therefore, it is not necessary to consider the possibility that the same character string will be input next time, and it is not necessary to learn the delimiter position.

However, in the kana-kanji conversion processing in the conventional character processing device, learning is performed even at the delimiter designated position generated by the operator's erroneous input. As a result, erroneous conversion as shown in FIG. 16 has occurred.

FIG. 16 shows an example of inputting and converting the character string "joshaken" after the processing shown in FIG. 15 described above. In this case, the conversion result desired by the operator is, of course, the "ticket", but since "jou / ya" exists as the delimiter position designation data, the delimiter between "joshi" and "yaken" is present. It will be converted with the position as the priority.
Therefore, the conversion result of "boss / ya ticket" was obtained.

As described above, in the kana-kanji conversion process of the conventional character processing device, the delimiter position designation by the position that is grammatically incorrect, which is highly likely to have occurred due to an incorrect input by the operator, is learned. By storing
In some cases, the desired conversion results could not be obtained, or unexpected conversion errors could occur.

The present invention has been made to solve such a problem, and in the kana-kanji conversion process for learning delimiter position designation data, the delimiter position is grammatically related to the input character string to be converted. An object of the present invention is to provide a character processing device and a method thereof, which learn erroneous delimiter positions when appropriate and thereby prevent erroneous conversion due to improper delimiter positions and improve the conversion rate.

Another object of the present invention is to provide a character processing apparatus and method capable of reducing the required memory capacity because grammatically inappropriate delimiter positions are not stored.

[0025]

In order to achieve the above-mentioned object, the present invention comprises the following constitutions.

That is, input means for inputting the first character string, designating means for designating a delimiter position for the first character string, and whether or not the delimiter position designated by the designating means is appropriate. A determination unit that determines whether or not the delimiter position is appropriate, and a retaining unit that retains information regarding the delimiter position and the first character string that are designated by the designating unit. And a delimiter unit that delimits at the delimiter position regarding the information held in the holding unit.

Further, it is characterized by further comprising conversion means for converting the first character string delimited by the delimiter means into a second character string.

For example, the holding means holds the character strings before and after the delimiter position of the first character string.

For example, the holding means may hold a plurality of pieces of information regarding the delimiter position of the first character string in advance, and the information regarding the delimiter position of the first character string designated by the designating means may not be held. It is characterized by additionally holding.

For example, the judging means is judged to be appropriate if the character string before and after the delimiter position in the first character string is a word.

Further, it is appropriate that the judgment means has a word dictionary in which words are registered, and the judging means is such that character strings before and after the information regarding the delimiter position in the first character string exist in the word dictionary. It is characterized by determining that.

For example, the first character string is made up of a first character type, and the second character string is made up of a second character type including the first character type.

For example, the first character type is a kana character, and the second character type is a kanji character.

[0034]

With the above configuration, the delimiter position is designated for the input first character string, and when the delimiter position is determined to be grammatically appropriate, the delimiter position is held.
A unique effect of dividing the first character string at the dividing position can be obtained.

[0035]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment according to the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a block diagram showing the arrangement of an information processing system using the character processing apparatus according to this embodiment. In FIG. 1, 1 is a keyboard, a mouse, an OCR,
An input unit such as a microphone for inputting a character string to be converted into Kana-Kanji in this embodiment. Reference numeral 2 denotes a conversion processing unit that performs a Kana-Kanji conversion process in this embodiment, and converts the character string input by the input unit 1 into a kanji-Kana mixed sentence or the like. Reference numeral 3 is an output unit such as a CRT, and the processing result in the conversion processing unit 2 is visually output to the operator.
Reference numeral 4 denotes a storage unit such as a hard disk, which stores a history of conversion processing performed in the conversion processing unit 2 and is referred to as needed in the conversion processing of this embodiment.

FIG. 2 shows a detailed configuration of the conversion processing unit 2 described above. In FIG. 2, 201 is a kana-kanji conversion processing unit that includes a CPU and that actually performs kana-kanji conversion processing. Two
Reference numeral 02 denotes a delimiter position data holding unit that holds the delimiter position data shown in the above-mentioned conventional example, and is composed of a RAM or the like. 205 is a word dictionary. Reference numeral 203 denotes a ROM, which holds various dictionaries other than the word dictionary used in the document processing apparatus according to the present embodiment, and processing procedures shown in the flowchart of FIG. Also, reference numeral 204 denotes R used as a work area.
AM.

FIG. 3 shows a flowchart of the kana-kanji conversion processing in the kana-kanji conversion processing unit 201, which will be described below.

In FIG. 3, first, in step S1, the operator inputs a kana character string or the like through the input unit 1. Then, in step S2, the kana-kanji conversion processing unit 201 performs kana-kanji conversion on the inputted kana-character string to convert it into a kanji-kana mixed character string. At this time, the delimiter position designation data holding unit 202 searches the delimiter position designation data corresponding to the input character string, and if there is, delimiter position designation data in the kana character string is determined by the delimiter position designation data. Then, the conversion is performed. Also,
If the delimiter position designation data does not exist, as a result of analyzing the input character string, it is converted into a kanji / kana mixed sentence or the like depending on the delimiter position which seems to be optimal.

Subsequently, in step S3, the kanji-kana mixed character string and the like obtained as described above is output as a conversion result. Then, in step S4, the operator determines whether or not the obtained conversion result is the desired conversion result. If the desired conversion result is obtained, step S
The process proceeds to step 11 and the conversion result is confirmed and the process ends.

In step S5, the operator designates a desired delimiter position for the input character string. This is performed by designating the delimiter position by a method of inserting a predetermined symbol or the like into the input character string displayed on the output unit 3 such as a CRT with a keyboard or a mouse.

Then, the process proceeds to step S6 so that the words are preferentially divided at the delimitation designated position designated by the operator.
Converts the input character string into a kanji-kana conversion mixed sentence. In step S7, the kana-kana mixed character string or the like obtained as described above is output to the output unit 3 as the conversion result. Then, in step S8, the conversion result obtained in step S7 is confirmed.

Next, in step S9, it is determined whether the delimiter position specified in step S5 is a grammatically correct (not an error) delimiter position. This determination method may be performed by determining whether or not the result of delimiting the input character string at the delimiter position is a significant word. For example, in the input character string, character strings located before and after the delimiter position are words. Judgment is made based on whether or not That is,
The word dictionary 205 is referred to for character strings before and after the delimiter position in the input character string, and if there is no corresponding character, the delimiter position is considered to be grammatically incorrect.

If it is determined that the delimiter position is not grammatically correct, the process is corrected as it is. If it is determined that the delimiter position is correct, the process proceeds to step S10 and the delimiter position designated by the operator in step S5 is set. , Is stored in the delimiter position designation data holding unit 202 as new delimiter position designation data in association with the input character string.

The history of the above-described processing shown in FIG. 3 is held in the storage unit 4 as a journal, for example. Therefore,
When previous information is needed, the information at a predetermined time can be restored by referring to the storage unit 4.

A specific example of the kana-kanji conversion processing in this embodiment will be shown and described below.

As an example of conversion in the case where the delimiter position designated by the operator is judged to be grammatically correct in step S9 of FIG. 3, a case where "here kimono" is used as the input character string is shown in FIG. ~ Shown in FIG.

First, FIG. 4 shows an example of conversion when there is no delimiter position designation data corresponding to the input character string "here kimono wo". In this case, when the kana-kanji conversion is performed, there is no corresponding data in the delimiter position designation data holding section 202, so that the kana-kanji conversion section 20 is selected in step S2.
In step 1, based on the result of the predetermined morpheme analysis and the likelihood information, "here / Kimono" which seems to be most appropriate is output as the conversion result.

In the following, "/" indicates the position recognized as the delimiter position in the character string of the kana-kanji conversion result,
It is not actually output as a conversion result character string.

In FIG. 5, the conversion result “here / Kimono” output as shown in FIG. 4, for example, was not the desired conversion result, so in step S5 the operator changed and designated the break position. An example of conversion in the case is shown. For example, in the output unit 3 such as a CRT, the operator newly designates as a delimiter position by inserting a predetermined symbol or the like between "here" and "footwear" with the input unit 1 such as a keyboard or a mouse. Suppose In this case, kana-kanji conversion is performed with the designated delimiter position as the highest priority (step S6). Therefore, the desired conversion result "here / footwear" can be obtained (step S7,
8).

Then, it is judged whether or not the newly designated delimiter position is a grammatically correct delimiter position (step S9). Various methods can be considered as a criterion for judging whether the delimiter position is grammatically appropriate. For example, as an example thereof, "in the input character string, the character strings before and after the specified delimiter position are significant words. Consider the case in which a judgment criterion "whether or not it can become" is established.
Then, in this case, when the delimiter position is specified after "here" in the input character string "here kimono", the character strings before and after the delimiter position are "here" and "footwear". Becomes These are detected as words as a result of searching the word dictionary 205. Therefore, the delimiter position "here / footwear" is determined to be valid.

Therefore, since the delimiter position is grammatically correct (no particular error), the delimiter position designation data is retained in the format of "here / footwear" so that the desired conversion result can be obtained at the next conversion. It is held in the unit 202 (step S10).

FIG. 6 shows an example of conversion when the same sentence is input again after the processing shown in FIG. In this case, first, in step S2, the delimiter position designation data holding unit 2
It is searched whether or not the delimiter position designation data corresponding to the input character string "here kimono wo" exists in 02. Then, since the delimiter position designation data "here / footwear" exists in the delimiter position designation data holding unit 202, the delimiter position designation by the delimiter position designation data is given priority in the conversion process. As a result, the desired conversion result “here / footwear” is obtained as the first candidate.

Similarly, in step S4 of FIG. 3, a conversion example in the case where the specified delimiter position is determined to be grammatically incorrect is shown in FIG. 7 to 9 show.

First, FIG. 7 shows an example of conversion in the case where the delimiter position designation data corresponding to the input character string "josha" does not exist. In this case, when the kana-kanji conversion is performed in step S2, since there is no corresponding data in the delimiter position designation data holding unit 202, "ride" is obtained as the conversion result.

Here, it is assumed that the input character string is an erroneous input by the operator, and the original hope of the operator is to input "joushi" to obtain the conversion result of "boss".

In this case, as shown in FIG. 8, the operator performs kana-kanji conversion for the erroneously input character string "josha" by changing the delimiter position. That is, the operator is
Specify a delimiter position between the and "ya" (step S5),
After that, kana-kanji conversion is performed (step S6) to obtain the conversion result "boss / ya" (step S7). Then, this "boss / ya" is confirmed (step S8).

Then, in step S9, it is determined whether the delimiter position specified in step S5 is grammatically correct for the input character string. In this case, "Joshi"
At a delimiter position existing between and, the word beginning with "a" is not searched in the word dictionary 205, and thus the delimiter position is determined to be grammatically incorrect.
Therefore, this delimiter position designation data is not stored.

Next, FIG. 9 shows an example of a case where the character string "joshaken" is input and converted after the processing of FIG. In FIG. 9, the delimiter position designation data holding unit 20
Since the delimiter position designation data corresponding to the input character string does not exist in 2, the conversion result "ticket" desired by the operator can be obtained.

As described above, according to the present embodiment, the delimiter position is learned only when it is determined that the designated delimiter position is grammatically appropriate in the input character string, thereby eliminating waste. It is possible to perform more appropriate conversion without storing different delimiter positions and improve conversion efficiency.

In the above embodiment, the input character string is described as a kana character string, but the present invention is not limited to this, and the input character string may be, for example, katakana or alphabet ( Any format may be used as long as the character string can be converted by dividing the character string at a predetermined position, such as Roman characters.

Similarly, the Kana-Kanji conversion is described as an example of the conversion process, but the conversion is not limited to conversion into a character string containing Kana-Kana, but conversion into a character string including symbols, external characters, etc. may also be performed. good. It is also possible to convert a character string having a predetermined character type into a character string having the same character type but different meaning, such as encryption.

In the above embodiment, an example in which the delimiter position designated by the operator is stored in the delimiter position designation data holding unit 202 as new delimiter position designation data has been described. The designated delimiter position is the delimiter position designation data holding unit 20.
If it already exists in 2, there is no need to newly store it.

Further, the delimiter position designation data stored in the delimiter position data holding unit 202 is not necessarily one for one character string, and there are also character strings having a plurality of delimiter position data. In this case, it is possible to provide a priority for each of the plurality of delimiter position data and output the data in descending order of the priority. In this case, if the operator performs the same designation as the existing position designation data in the position designation data holding unit 202 as described above, the priority may be raised.

Although the method of referring to the word dictionary has been described as an example of the method of determining whether the delimiter position is grammatically correct, the determination method of the present invention is not limited to this example. For example, without referring to a word dictionary,
It is also possible that a character string that is clearly not the beginning of a word exists immediately after the delimiter position. Therefore, for example, the above-mentioned “ya”, “n”, and the like, which hold a word that can be clearly judged not to be the beginning of a word, are stored in advance as a pre-judgment dictionary and the like, and the pre-judgment dictionary is referred to before referring to the word dictionary. If so, the processing speed can be improved.
It should be noted that the data stored in the previous determination dictionary is not necessarily limited to words, and may hold a specific condition that is determined not to be the beginning of a word, for example.

The present invention may be applied to a system composed of a plurality of devices or an apparatus composed of one device. Further, it goes without saying that the present invention can be applied to the case where it is achieved by supplying a program to a system or an apparatus.

[0067]

As described above, according to the present invention, when a delimiter position that is not grammatically correct is designated when the delimiter position designation data is learned and the character string is delimited, in most cases, the operator's No learning is performed because it is due to an erroneous input. This can eliminate erroneous conversion due to an improper delimiter position in, for example, kana-kanji conversion processing, so that the conversion rate can be improved and the operability can be improved.

Further, since unnecessary delimiter position designation data is not registered, the data amount can be reduced, that is, the memory capacity can be reduced or effectively used.

[0069]

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a character processing device according to an embodiment of the present invention.

FIG. 2 is a block diagram showing a detailed configuration of a conversion processing unit in the present embodiment.

FIG. 3 is a flowchart showing a kana-kanji conversion process in this embodiment.

FIG. 4 is a diagram showing an example of kana-kanji conversion according to the present embodiment.

FIG. 5 is a diagram showing an example of kana-kanji conversion according to the present embodiment.

FIG. 6 is a diagram showing an example of kana-kanji conversion according to the present embodiment.

FIG. 7 is a diagram showing an example of kana-kanji conversion according to the present embodiment.

FIG. 8 is a diagram showing an example of kana-kanji conversion according to the present embodiment.

FIG. 9 is a diagram showing an example of kana-kanji conversion according to the present embodiment.

FIG. 10 is a flowchart showing a kana-kanji conversion process in a conventional character processing device.

FIG. 11 is a diagram showing a conventional kana-kanji conversion example.

FIG. 12 is a diagram showing a conventional kana-kanji conversion example.

FIG. 13 is a diagram showing a conventional kana-kanji conversion example.

FIG. 14 is a diagram showing a conventional kana-kanji conversion example.

FIG. 15 is a diagram showing a conventional kana-kanji conversion example.

FIG. 16 is a diagram showing a conventional kana-kanji conversion example.

[Explanation of symbols]

 1 Input unit 2 Conversion processing unit 3 Output unit 4 Storage unit 201 Kana-Kanji conversion processing unit 202 Delimitation position data holding unit 203 ROM 204 RAM 205 Word dictionary

Claims (16)

[Claims] 1. An input unit for inputting a first character string, a designating unit for designating a delimiter position for the first character string, and whether or not the delimiter position designated by the designating unit is appropriate. A determination unit that determines whether or not the delimiter position is appropriate, and a retaining unit that retains information about the delimiter position; and the first character string that is designated by the designating unit. And a delimiter unit that delimits the delimiter position with respect to the information held in the holding unit. 2. The character processing device according to claim 1, further comprising conversion means for converting the first character string delimited by the delimiter means into a second character string. 3. The character processing device according to claim 1, wherein the holding unit holds a character string before and after a delimiter position of the first character string. 4. The holding means preliminarily holds a plurality of pieces of information on the delimiter position of the first character string, and the information on the delimiter position of the first character string designated by the designating means may not be held. The character processing device according to claim 1, wherein the character processing device is additionally held. 5. The character processing device according to claim 3, wherein the determination means determines that the character string before and after the delimiter position in the first character string is a word. . 6. A word dictionary in which words are registered is further provided, and the determination means is suitable if character strings before and after the information regarding the delimiter position in the first character string exist in the word dictionary. 6. The character processing device according to claim 5, wherein the character processing device determines that 7. The first character string is composed of a first character type, and the second character string is a second character string including the first character type.
The character processing device according to claim 2, wherein the character processing device comprises the following character types. 8. The first character type is a kana character, and the second character type is a kanji character.
Character processing device listed. 9. An input step of inputting a first character string, a designating step of designating a delimiter position for the first character string, and whether or not the delimiter position designated by the designating step is appropriate. A determination step of determining whether or not the delimiter position is appropriate by the determination step, a retaining step of retaining information regarding the delimiter position in retaining means, and the first character string And a delimiter step of delimiting at a delimiter position designated by the means and a delimiter position relating to the information held in the holding means. 10. The character processing method according to claim 9, further comprising a conversion step of converting the first character string delimited by the delimiting step into a second character string. 11. The character processing method according to claim 9, wherein the holding step holds a character string before and after a delimiter position of the first character string. 12. The holding means holds in advance a plurality of pieces of information about the delimiter position of the first character string, and the holding means stores information about the delimiter position of the first character string designated by the designating step. The character processing method according to claim 9, wherein if the character is not held, the character is additionally held. 13. The determination step, if the character string before and after the delimiter position in the first character string is a word,
The character processing method according to claim 11, wherein the character processing method is determined to be appropriate. 14. The method further comprises a word dictionary in which words are registered, and the determination means is appropriate if character strings before and after the information regarding the delimiter position in the first character string exist in the word dictionary. 14. The character processing method according to claim 13, wherein the character processing method determines that 15. The method according to claim 10, wherein the first character string is composed of a first character type and the second character string is composed of a second character type including the first character type. Character processing method. 16. The first character type is a kana character,
The character processing method according to claim 15, wherein the second character type is Kanji.