JP3566873B2 - Kana-kanji conversion apparatus and method, and storage medium storing kana-kanji conversion program - Google Patents

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
INDUSTRIAL APPLICABILITY The present invention is applied to a personal computer having a Japanese word processor, a Japanese document processing function, and the like, and converts a reading character string including a plurality of phrases into a kanji kana mixed string at a time, and a kana-kanji conversion program. Related to a medium that stores
[0002]
[Prior art]
Conventionally, in a kana-kanji conversion device that converts a kana character string used in a Japanese word processor or the like into a kana-kana kana sequence, in particular, a continuous phrase kana-kanji conversion process that converts a kana character string consisting of multiple phrases into a kanji-kana kana sequence at a time. In the case where the break position of a phrase is incorrectly converted, it is necessary to correct the break position of the same phrase every time for conversion.
[0003]
As a conventional technique for solving this problem, according to the description in Japanese Patent Application Laid-Open No. Hei 2-35564, when a character string that needs to be corrected for a segment break is to be converted to kana-kanji, once corrected, no further correction is required thereafter. There has been proposed a Japanese input device that learns a phrase break so that the conversion result is as corrected immediately before.
In this conventional Japanese input device, in the conversion of a reading string composed of a plurality of phrases, when the phrase break is incorrectly converted, the phrase break position is moved, and the phrase break position is corrected and converted. The kanji-kana character string is registered in the provisional dictionary to learn the phrase segmentation position.
[0004]
[Problems to be solved by the invention]
However, the Japanese input device according to the above-described conventional technology cannot perform segmentation delimitation when kana-kanji conversion is performed on a phrase basis. Although the phrase separation can be performed correctly, immediately after that, if a kana-kanji conversion is performed by inputting the reading character strings of a plurality of phrases at once, there is a problem that the phrase separation cannot be performed correctly. If you want to leave only the correct kanji-kana character string of the conversion candidate, delete the part that was incorrectly converted, and re-enter the read character string of the erased part to convert it, use phrase delimiter. There is a problem that if a character string of a plurality of phrases is input and converted at a time, the phrase segmentation cannot be correctly performed.
[0005]
The present invention has been made in view of the above circumstances, and if a reading character string consisting of a plurality of phrases is input in units of a phrase and correctly converted, a correct phrase separation position is learned and a plurality of characters to be input from the next time are read. Provided are a kana-kanji conversion apparatus and method for correctly converting kana-kanji conversion of a reading character string of a phrase, and a storage medium storing a kana-kanji conversion program.
[0006]
[Means for Solving the Problems]
The present invention A phrase-separation learning dictionary that stores phrase-separation information from continuous phrase conversion results, and kana-kanji conversion has already been performed. A pre-phrase reading character string storage unit for storing a reading character string, It is input following the reading character string stored in the previous phrase reading character string storage A post-sequence reading character string storage unit for storing the reading character string, The reading character string in the post-sequence reading character string storage unit is converted into a single phrase by referring to the phrase-separation learning dictionary for single-phrase conversion or phrase-separation to obtain a kana-kana-kanji string as a kana-kanji conversion result. A conversion unit for converting a combined read character string obtained by combining the read character strings stored in the read character string storage unit and the post-sequence read character string storage unit into a continuous phrase, and a conversion result of the combined read character string being two phrases; And the reading character strings of the respective phrases are the reading character strings stored in the pre-phrase reading character string storage unit and the post-phrase reading character string storage unit, respectively. A judgment unit for judging whether they are the same, If the determining unit determines that the two are not the same, the delimiter information in units of the read character strings stored in the pre-sequence reading character string storage unit and the post-sequence reading character string storage unit is used as the phrase separation information of the synthesized read character string. Phrase separation learning control means for storing in the phrase separation learning dictionary It is a kana-kanji conversion device characterized by the following.
[0007]
In the present invention, the input unit may be configured by, for example, an input device including a keyboard, a pen / tablet, and a mouse.
The pre-sentence reading character string storage unit, the post-sentence reading character string storage unit, the pre-sentence kanji kana mixed string storage unit, and the post-sentence kanji kana mixed string storage unit are configured by a storage device including, for example, a RAM, an EEPROM, a floppy disk, and the like. May be.
The conversion unit, the storage control unit, the determination unit, and the learning unit may be configured by, for example, a computer CPU and a control program.
[0008]
According to the present invention, if a reading character string consisting of a plurality of phrases is divided and input in units of a phrase and is correctly converted, a correct phrase separation position is learned, and a reading character string of a plurality of phrases input next time is correctly converted into kana-kanji characters. can do.
[0009]
BEST MODE FOR CARRYING OUT THE INVENTION
In the present embodiment, for example, the reading character string “jishi”, which is the pre-phrase of the two-character reading character string “jishihaigyo”, is input and converted to obtain a conversion candidate “independent”. A reading character string “Haigyo”, which is a later sentence, is input and converted to obtain a conversion candidate “Closed”. The reading character strings of two phrases are input and converted in units of phrases, and the reading character strings and conversion candidates corresponding to the delimiting positions are internally stored.
Therefore, when a pre-stored two-segment reading character string “Jishuha Igyo” is input, a conversion candidate “independent is / great feat” based on an incorrect delimiter position “Jushiha / Igyo” is obtained. Instead, the correct conversion candidate “autonomous / closed” can be obtained by learning the correct break position “jishi / hagyo”. For convenience, “/” indicates a segment break position.
[0010]
According to another aspect of the present invention, there is provided a method of performing a kana-kanji conversion by performing a single-segment conversion or a continuous-segment conversion of a reading character string input for each single phrase or a plurality of phrases, wherein the kana-kanji conversion is performed for kana-kanji conversion. The sequence is stored in the pre-sentence reading character string storage unit, and the read character string input following the reading character string stored in the pre-sentence reading character string storage unit is stored in the post-sentence reading character string storage unit. The reading character string stored in the reading character string storage unit is converted into a single phrase or a continuous phrase conversion with reference to the phrase separation learning dictionary for phrase separation to obtain a kanji kana mixed string as a kana kanji conversion result, and the previous phrase reading character string The combined reading character string formed by combining the storage unit and the reading character string stored in the post-sequence reading character string storage unit is subjected to continuous phrase conversion, Whether the conversion result of the synthesized read character string is two phrases, and whether the read character string of each phrase is the same as the read character strings stored in the preceding phrase read character string storage unit and the subsequent phrase read character string storage unit, respectively. If it is determined that they are not the same, the delimiter information in units of the reading character strings stored in the pre-phrase reading character string storage unit and the post-phrase reading character string storage unit is used as the phrase separation of the composite reading character string. Store it in the phrase segmentation learning dictionary as information A kana-kanji conversion method is provided.
[0011]
In another embodiment, if the two-syllable reading character string “Jishu Haigyo” is converted at a time and the conversion candidate “independent is / feat” is obtained, the part of the kanji kana exchange string “ha / feat” is Delete and input the reading character string "Hello" again and convert it to obtain a conversion candidate "Closed". At that time, the character that remains without being erased by memorizing which character of the reading character string “jisyuhaigyo” corresponds to each character of the first conversion candidate “autonomous / feat” To determine what is.
[0012]
In this case, the character string to be read is “jishi”, and the kanji kana mixed string is “independent”. The remaining phrase is regarded as the first phrase, and the reading character string input subsequently is regarded as the second phrase. The reading character string is "Haikyo", and the kanji kana mix line is "Closed". A reading character string and a conversion candidate corresponding to a delimiter position where the reading character strings of the first and second phrases are connected are internally stored.
Therefore, when a pre-stored two-segment reading character string “Jishuha Igyo” is input, a conversion candidate “independent is / great feat” based on an incorrect delimiter position “Jushiha / Igyo” is obtained. Instead, the correct conversion candidate “autonomous / closed” can be obtained by learning the correct break position “jishi / hagyo”.
[0013]
An editing unit that partially deletes the kanji kana mixed string converted by the conversion unit is further provided, and the storage control unit stores the kanji kana mixed string remaining after the partial deletion by the editing unit in the previous sentence kanji kana mixed string storage unit. , The corresponding reading character string is stored in the preceding phrase reading character string storage unit, and the input reading character string following the deletion is stored in the subsequent phrase reading character string storage unit. The cross-sequence is stored in a post-sequence kanji kana cross-sequence storage unit, and the determination unit stores, in each of the storage units, a delimiter position when a read character string of the same plurality of phrases is input and converted to a kanji kana cross-sequence. It is determined whether or not it is the same as the delimiter position of the read character string, and upon receiving a different determination result from the determiner, the learning unit learns the delimiter position of the read character string. You may.
According to this configuration, even when a part of a character string that has been input and converted once is deleted, the correct kana-kanji conversion can be performed by learning the delimiter positions of the phrases of the character string remaining without being deleted.
[0014]
It further comprises a reading kanji kana mixing string correspondence table for storing the correspondence between each character of the reading character string and each character of the kanji kana mixing string. A kanji-kana-kanji column correspondence table that reads whether or not the kanji-kana-kane column converted from the reading character string connected to the column and the kanji-kana-kane column connected to the preceding sentence kanji-kana-kane sequence and the post-sequence kanji-kana-kana-sequence column are the same. When the learning unit receives different judgment results from the judgment unit, the learning unit connects the pre-sentence reading character string and the post-segment reading character string, and reads the pre-sentence kanji kana mixed sequence and the post-sentence kanji kana character. It may be configured to learn a correspondence relationship between the kanji string and the kanji kana string.
According to this configuration, it is possible to learn the delimiter position input in units of phrases based on the correspondence between the kanji kana character sequence that connects the pre-segment kanji kana character sequence and the kanji kana character sequence that connects the subsequent sentence kanji character.
[0015]
It further comprises a deciding section for moving a cursor in accordance with an instruction of the input section to fix the character string, and moving the cursor to a character string of another phrase by deciding a reading character string of the previous phrase by the deciding section. At this time, the storage control unit may be configured to delete all character strings stored in the pre-segment reading character string storage unit and the pre-segment kanji kana combination column storage unit, respectively.
According to this configuration, it is not necessary to perform the segment learning between discontinuous segments, which is caused by the movement of the cursor in the kana-kanji conversion by the user, and it is possible to prevent erroneous learning.
[0016]
If no character string is stored in the pre-sentence reading character string storage unit and the pre-sentence kanji kana combination line storage unit, the conversion unit converts the determined kanji kana combination sequence immediately before the cursor to a read character string. However, the storage control unit may be configured to store the preceding phrase reading character string storage unit and store the kanji kana mixed string in the previous phrase kanji kana mixed string storage unit.
According to this configuration, the burden on the user for the conversion processing operation can be reduced.
[0017]
Hereinafter, the present invention will be described in detail based on an embodiment shown in the drawings. The present invention is not limited by this.
[0018]
FIG. 1 is a block diagram showing the configuration of the kana-kanji conversion device of the present embodiment. In FIG. 1, reference numeral 1 denotes a control unit that controls each of the units 3 to 18 via a bus 2 based on a control program. The control unit 1 is, for example, a computer including a CPU, a ROM, a RAM, and an I / O port. Be composed. Further, the control unit 1 also functions as a storage control unit, a determination unit, a learning unit, an editing unit, and a determination unit.
Reference numeral 3 denotes an input unit for inputting an edit code such as a character code, conversion, or deletion instruction. The input unit 3 is configured by an input device including, for example, a keyboard, a pen / tablet, and a mouse.
Reference numeral 4 denotes a kana-kanji conversion unit for converting kana-kanji conversion of the reading kana input by the input unit 3, and is composed of, for example, a CPU of a computer, a ROM storing a kana-kanji conversion program, and an EEPROM.
[0019]
Reference numeral 5 denotes a kana-kanji conversion dictionary. The kana-kanji conversion dictionary 5 is, for example, a storage device including a ROM, an EEPROM, a floppy disk (FD), and a hard disk (HD) storing kana-kanji conversion data for kana-kanji conversion. It consists of.
Reference numeral 6 denotes a phrase segmentation learning dictionary. The phrase segmentation learning dictionary 6 includes, for example, a RAM and an EEPROM for storing learning data for learning segmentation positions of a plurality of phrases.
[0020]
Reference numeral 7 denotes a display buffer unit that stores an input character string, a converted character string, and the like as display data, and includes, for example, a RAM and an EEPROM.
Reference numeral 8 denotes a display unit that displays an input read character string, a converted character string, and the like. The display unit 8 is a display device including, for example, a CRT display, a liquid crystal display (LCD), and a plasma display (PD). Be composed.
Reference numeral 9 denotes an instruction unit for instructing a position such as insertion or deletion of a character by a cursor pointer. The instruction unit 9 is configured by, for example, an input device including a keyboard, a pen / tablet, and a mouse.
Reference numeral 10 denotes a printing unit that prints a character string displayed on the screen on the display unit 8, and the printing unit 10 includes, for example, a thermal printer, an inkjet printer, or the like.
[0021]
Reference numeral 11 denotes a pre-phrase reading character string storage unit that stores a reading character string of the previous phrase among the determined plural phrases, and is configured by, for example, a RAM and an EEPROM.
Reference numeral 12 denotes a post-segment reading character string storage unit that stores the currently input reading character string, and is composed of, for example, a RAM and an EEPROM.
Reference numeral 13 denotes a pre-sentence kanji kana interlacing sequence storage unit for storing the converted kanji kana interlacing sequence of the previous sentence, and is composed of, for example, a RAM and an EEPROM.
Numeral 14 denotes a post-sequence kanji kana mixed string storage unit for storing the converted kanji kana mixed string of the post-sentence, and is composed of, for example, a RAM and an EEPROM.
[0022]
Reference numeral 15 denotes a combined read character string storage unit that stores a read character string obtained by combining the read character string of the previous phrase input immediately before and the read character string that is currently input, and includes, for example, a RAM and an EEPROM.
Reference numeral 16 denotes a combined kanji kana string storage section for storing a kanji kana string string obtained by combining the converted kanji kana string string of the previous sentence and the currently converted kanji kana string string of the later sentence. For example, RAM, EEPROM, etc. It consists of.
Reference numeral 17 denotes a reading kanji kana-sequence row correspondence table for storing the input kanji character string and the conversion result of the kanji-kana kana-sequence corresponding to the positional information, and is composed of, for example, a RAM and an EEPROM.
[0023]
Reference numeral 18 denotes a storage medium storing the kana-kanji conversion program of the present invention, and the storage medium 18 is constituted by, for example, a computer-readable FD, HD, CD-ROM, or the like.
Reference numeral 19 denotes a storage medium reading unit that reads a kana-kanji conversion program from the storage medium 18 and installs the program in a RAM or an EEPROM.
[0024]
FIG. 2 is a block diagram illustrating the configuration of the storage medium of the present embodiment. In the area shown in FIG. 2A, a kana-kanji conversion program 201 for converting an input character string into a kanji-kana mixed character string with reference to a kana-kanji conversion dictionary, and deletes, corrects, and adds characters An editing program 202, a phrase break determination program 203 for determining whether or not the break positions of the phrases are the same, a phrase break learning program 204 for learning the break positions of the phrases, a character string comparison program for comparing character strings to determine whether they are the same 205, each code of the character string learning program 206 for learning the character string when the character string is not the same is stored.
[0025]
In the area shown in FIG. 2B, a pre-segment reading character string storage control program 207, a post-segment reading character string storage control program 208, a pre-segment kanji kana mixed string storage control program 209, a post-sentence kanji kana mixed string storage A control program 210, a synthetic read character string storage control program 211, a synthetic kanji kana mixed string storage control program 212, a phrase separation learning data storage control program 213, and a character string learning data storage control program 214 are stored respectively.
[0026]
According to another aspect of the present invention, the computer 1 sequentially converts the reading character strings of a plurality of phrases input by the input unit 3 in units of phrases into kanji kana mixed sequences, and reads the reading character strings of the preceding phrase out of the two phrases. Is stored in the pre-bunsetsu reading character string storage unit 11, the post-bunsetsu reading character string following the pre-bunsetsu is stored in the post-bunsetsu reading character string storage unit 12, and the converted kanji kana mixed string of the pre-bunsetsu is stored in the pre-bunsetsu kanji. The kana mixed string storage unit 13 stores the converted kanji kana mixed string of the post-sequence in the post-sentence kanji kana mixed string storage unit 14. To determine whether or not the delimiter position when converted to the same as the delimiter position of the read character string stored in each of the storage units 11 to 13. Learning break positions It is possible to provide a recording medium 18 that stores the kana-kanji conversion program for causing.
[0027]
Therefore, by reading the kana-kanji conversion programs 201 to 214 of the present invention from the recording medium 18 and installing them in the RAM and the EEPROM, it is possible to input a read character string composed of a plurality of phrases in units of phrases, and to convert them correctly. It is possible to realize a kana-kanji conversion device that learns a phrase separation position and correctly converts kana-kanji conversion of a plurality of phrase reading character strings to be input next time.
[0028]
FIG. 3 is a flowchart showing the procedure (1) of the kana-kanji conversion process of the present embodiment. FIG. 4 is a flowchart showing the procedure (2) of the kana-kanji conversion process of the present embodiment. FIG. 5 is a diagram illustrating a storage state of each storage unit according to the kana-kanji conversion process of the present embodiment.
In the present embodiment, for example, when a two-segment reading character string “jishihaigyo” is converted to a kanji kana mixed character string “independent is feat”, the reading character strings “jishiha” and “haigyo” are read. Is divided into two phrases and converted to obtain the Kanji kana mixed character strings "independent" and "closed business". From the next time, the correct Kanji kana mixed string for the input of the reading character string "jisyuhaigyo" Obtain “voluntary closure”.
[0029]
In the flowchart of FIG.
Step 301: A key code is input from the input unit 3.
Step 302: Check whether the input key code is a character code. If the key code is a character code, proceed to step 303; otherwise, proceed to step 306.
Step 303: Store in the later-sentence reading character string storage unit 12.
FIG. 5A shows that “jishi” (5-1) is input to the post-sequence reading character string storage unit 12. In the present embodiment, the phrase to be converted by the user is treated as a later phrase for convenience.
[0030]
Step 304: Store in the display buffer 7.
Step 305: Display the conversion result on the display unit 8.
Step 306: Check whether the key code is a conversion key. If the key code is not a conversion key, proceed to step 307; otherwise, proceed to a kana-kanji conversion process of step 314.
Step 307: Check whether the key code is an instruction for confirmation. If it is not a confirmation instruction, the process returns to step 301. In the case of the confirmation instruction, the process proceeds to step 308.
[0031]
In FIG.
Step 315: Kana-Kanji conversion of the post-sentence reading character string (5-1). The kana-kanji conversion unit 4 performs conversion with reference to the kana-kanji conversion dictionary 5. FIG. 5B shows that the reading character string “jishi” is converted to kana-kanji and the result is stored as “independent” (5-2) in the post-sequence kanji kana-mixing column storage unit 14. I have.
Step 316: Check whether the conversion result is a plurality of clauses. If YES, proceed to Step 317; otherwise, proceed to Step 319.
Step 317: Check whether the conversion result is registered in the phrase segmentation learning dictionary 6, and if YES, proceed to Step 318, otherwise proceed to Step 319.
[0032]
Step 318: Store the learning candidates in the display buffer 7.
Step 319: Store the conversion result in the display buffer 7. Next, the process proceeds to step 305, where the conversion result “independent” is displayed on the display unit 8.
When the user gives the “independent” confirmation instruction, the process returns to step 301 and proceeds to step 308.
[0033]
Step 308: Check whether there is a preceding phrase reading character string in the preceding phrase reading character string storage unit 11. In this case, since it is not stored, the process proceeds to step 313.
Step 313: As shown in FIG. 5c, the post-sequence reading character string is copied to the pre-sequence reading character string storage unit 11 as the pre-sequence reading character string (5-3), and It is copied to the previous sentence kanji kana interlacing column storage unit 13 as a kanji kana interlacing column (5-4), and the process returns to step 304 and step 305.
[0034]
Subsequently, as shown in FIG. 5e, the input of the reading character string "Haigyo" and the kana-kanji conversion are performed, and the kanji kana mixed string "closed out" is obtained and determined.
Step 307: Check whether it is a confirmation instruction.
Step 308: Check whether there is a preceding phrase reading character string in the preceding phrase reading character string storage unit 11. In this case, since it is stored, the process proceeds to step 309.
Step 309: The pre-sentence reading character string and the post-sentence reading character are combined, and the combined reading character string “jishihaigyo” is stored in the combined reading character string storage unit 15, as shown in FIG.
Step 310: The kana-kanji conversion of the synthesized reading character string is performed, and in a state where learning is not performed, the conversion result “independent / great achievement” is stored in the kana-kanji mixed string storage unit 16 as shown in g of FIG.
[0035]
Step 311: It is checked whether the conversion result is two phrases and the delimiter position is between the preceding phrase reading character string and the subsequent phrase reading character string. In this case, the delimiter position of “jishiha / igyo” is between the pre-sequence reading character string and the postsequence read character string, and is not the same as the delimiter position of “jishi / haigyo”. It is necessary to proceed and learn the break position. If they are the same, the process proceeds to step 313 because the segment break learning is not performed.
Step 312: The learning registration process of the phrase segmentation is performed, the learning information is stored in the phrase segmentation learning dictionary 6, and the process proceeds to step 313. H in FIG. 5 shows a state when the process proceeds to step 313.
[0036]
FIG. 6 is a diagram showing the stored contents of the phrase segmentation learning dictionary of the present embodiment. As shown in FIG. 6A, the phrase segmentation learning registration process may be configured such that a kanji kana exchange sequence “autonomous out of business” is stored for the reading character string “jishihaigyo”. . In this case, it means that the kanji kana exchange string “independent closing” has been learned with respect to the reading character string “jishuhaigyo”.
In addition, as shown in FIG. 6B, a configuration may be adopted in which only the read character strings “jishi” and “haigyo” are registered, and the kanji kana mixed string is not registered.
[0037]
In the flowcharts of FIG. 3 and FIG. 4, after learning the phrase separation position of the kanji kana combination string “independent closure” in the reading character string “jishi haigyo”, the user at one time reads “jishihaigyo”. An example of input conversion will be described.
Step 315: As the conversion result candidates for the reading character string “Jishu Igyo”, “independent / great business”, “independent / closed business”, etc. are obtained.
Step 316: Check whether the conversion result is a plurality of clauses. If YES, proceed to Step 317; otherwise, proceed to Step 319.
[0038]
Step 317: Check whether the conversion result is registered in the phrase segmentation learning dictionary. If YES, proceed to Step 318; if NO, proceed to Step 319.
In this case, in step 312, as shown in FIG. 6A, the phrase segmentation learning is being performed, and “voluntary closure” is a learning candidate.
Step 318: The learning candidate “voluntary closure” is stored in the display buffer 7. Next, the process proceeds to step 305, where the conversion result "voluntary closure" is displayed on the display unit 8.
[0039]
FIG. 7 is a diagram showing the storage contents (1) of the reading kanji kana mixed character string correspondence table of the present embodiment. FIG. 8 is a diagram showing the storage contents (2) of the reading kanji kana mixed character string correspondence table of the present embodiment. FIG. 9 is a diagram showing the storage contents (3) of the reading kanji kana mixed character string correspondence table of the present embodiment. In particular, FIGS. 7 and 8 are explanatory diagrams in which information for examining the reading of one kanji is stored.
[0040]
FIG. 7A shows a state in which the reading character string “Joshiha Haigyo” has been input, and the numbers in the figure indicate the number of the character from the beginning.
FIG. 7B shows a kana-kana kana mixing column “independent is great” as a result of the kana-kanji conversion.
FIG. 7C shows the position of each character in the reading character string corresponding to each character in the kanji kana mixed string, and is stored in the reading kanji kana mixed string correspondence table 17.
[0041]
FIG. 10 is a diagram showing the storage contents of the kana-kanji conversion dictionary of the present embodiment. In the kana-kanji conversion dictionary shown in FIG. 10, the kanji code is stored using two bytes of the JISX0208 code, and the number of characters read for one kanji is two bytes that are not used in the JISX0208 code. It is stored using one bit of the MSB of the first byte and the second byte of the represented kanji code.
[0042]
As shown in FIG. 10A, when the number of characters read for one kanji is one, the MSB of the first and second bytes of the kanji code is set to 0. When the number of characters read for one kanji is two, the MSB of the first byte of the kanji code is 0 and the MSB of the second byte is 1. Hereinafter, when the number of characters to be read for one kanji is 5 or more, the MSB of the first byte and the second byte of the kanji code becomes 1, and the next byte contains a number of 5 or more, and the number is the number of characters to be read. And
[0043]
FIG. 10 (b) is an explanatory diagram showing the pronunciation of reading characters and kanji for words registered in the kana-kanji conversion dictionary, and further showing the number of reading characters for each display of each kanji of the word. .
[0044]
FIG. 10 illustrates an example of correspondence between the read character string “Igyo” and the kanji character “Kyogyo”.
When examining the kanji code of the kanji "wei", since the first and second bytes of the MSB are 0, it is understood that the reading character of the "wei" is one character and the reading character is "i". I understand. Subsequently, when the kanji code of the kanji “work” is examined, the MSB of the first byte is 1 and the MSB of the second byte is 0, so that the number of read characters is three. Therefore, it can be understood that the reading character of the kanji character “Yoku” is “Gyo”.
[0045]
By using the Kana-Kanji conversion dictionary as described above, if the reading character string “Jishu Haigyo” is converted to Kana-Kanji and the result is “Independent is feat,” search the Kana-Kanji conversion dictionary 5 by Kana-Kanji conversion. Then, for each kanji, it is possible to know what the reading characters are.
In this example, “self” corresponds to “ji”, “lord” corresponds to “shu”, “ha” corresponds to “ha”, “great” corresponds to “i”, and “work” corresponds to “gyo”. (C) is stored in the reading kanji kana combination column correspondence table.
[0046]
The reading kanji kana mixed string correspondence table shown in (c) of FIG. 7 is a table indicating which character of the read character string points to each character of the kanji kana mixed string.
Therefore, the "self" of the kanji kana mixed string has a kanji kana mixed string position of 0 and the reading character string starts with 0, and the "main" has a kanji kana mixed string position of 1 and the reading character string starts with 1 and ""Ha" has a kanji kana mixed string position of 2 and the reading character string starts with 3, "Wai" has a kanji kana mixed string position of 3 and the reading character string starts with 4, and in the same manner, "work" has a kanji character The kana combination row position starts at 4, and the read character string starts at 5.
Since the end of the reading kanji kana mixed string correspondence table indicates the position +1 where the last character of the kanji kana mixed string ends, the kanji kana mixed string is 5 and the read character string is 8.
[0047]
FIG. 11 is a flowchart showing the procedure (3) of the kana-kanji conversion process of the present embodiment. In FIG. 11, the processing from step 1101 to step 1114 is the same as the processing from step 301 to step 314 shown in FIG.
FIG. 12 is a flowchart showing the procedure (4) of the kana-kanji conversion process of the present embodiment. Note that the processing from step 1130 to step 1137 shows a flowchart for determining the break position between the pre-phrase reading character string and the post-phrase reading character string by the conversion result partial deletion processing.
[0048]
Step 1130: Check whether the input key code is a delete key. If YES, the process proceeds to the character deletion process of step 1131. If NO, the process proceeds to step 1106.
Step 1132: It is checked whether or not the target of the deletion instruction is the part of the kanji kana mixed string of the previous phrase. If YES, the process proceeds to step 1133, and if NO, the process proceeds to step 1136.
[0049]
Step 1133: To delete the previous sentence, delete the character string to be deleted from the previous sentence kanji kana mixed string. Before the deletion, as shown in FIG. 7B, “independent is great” is input, and as shown in FIG. 8B, one character is deleted from “industry”.
Step 1134: As shown in FIG. 7 (c), characters starting at position 4 and ending at position 5 in the kanji kana mixed string are deleted, and the reading character string corresponding thereto is also started at position 5;
It is found that the read character string ends at position 8, and the read character string “Gyo” is deleted, and the state shown in FIG. 9A is obtained.
[0050]
Here, since the position of the ending character actually indicates the start position of the next character, it is a number one larger than the actual ending character position.
Step 1135: Update the reading kanji kana combination column correspondence table to a state as shown in FIG. 9C, proceed to step 1137, and end the character deletion processing.
Step 1136: Delete the character string other than the previous phrase and proceed to the end processing of step 1137.
[0051]
When the user performs the above character deletion processing and desires to obtain the kanji kana character string "autonomous out of business" by converting the reading character string "jishuhaigyo", the kana kanji conversion result Becomes “independent is a feat”, and when the user deletes “ha feat” in the control unit 1, the three-letter sentence of the kanji kana mixed string is deleted. The three characters of the later sentence shown in (c) are deleted, and the result is as shown in (c) of FIG.
[0052]
Therefore, as shown in FIG. 9B, it can be seen that the read character string of the kanji kana mixed string “independent” which remains without being deleted is “Jishi” shown in FIG. 9A.
This "jishi" is used as a pre-sentence reading character string, "independent" is used as a pre-sentence kanji kana mixed string, and the next input reading character string is converted as a post-sentence reading character string.
[0053]
Therefore, when the input character strings of a plurality of clauses are input and converted at once, only the correct kanji kana mixed string of the conversion result is left, the incorrect conversion part is deleted, and the input character string corresponding to the part is input again. Even after conversion, the segment break position can be learned.
According to this embodiment, even when a part of a character string that has been input and converted once is deleted, the correct kana-kanji conversion can be performed by learning the delimiter positions of the phrases of the character string remaining without being deleted.
[0054]
FIG. 13 is a flowchart showing the procedure (5) of the kana-kanji conversion process of the present embodiment. In FIG. 13, the processing from step 1301 to step 1314 is the same as the processing from step 301 to step 314 shown in FIG.
Step 1320: Check whether the key code input by the user is a cursor movement instruction. If YES, the process proceeds to step 1321, and if NO, the process proceeds to step 1306.
Step 1321: When the previous phrase reading character string and the previous phrase kanji kana string are stored in the previous phrase reading character string storage section and the previous phrase kanji mixed string storage section, these are cleared.
[0055]
According to this embodiment, when the cursor is moved and a character string is input, the character string input before that is rejected, so that there is no error in learning between unrelated phrases, Correct phrase segmentation learning can be performed, and the convenience of the user's conversion processing operation can be improved. In other words, it is not necessary to perform the segment learning between discontinuous phrases caused by the cursor movement, and it is possible to prevent erroneous learning.
[0056]
FIG. 14 is a flowchart showing the procedure (6) of the kana-kanji conversion process of the present embodiment. In FIG. 14, the processing from step 1401 to step 1414 is the same as the processing from step 301 to step 314 shown in FIG.
Step 1420: Check whether the key code input by the user is a cursor movement instruction. If YES, proceed to Step 1421; if NO, proceed to Step 1406.
Step 1421: When the pre-segment reading character string and the pre-segment kanji kana string are stored in the pre-segment reading character string storage unit and the pre-segment kanji mixed string storage unit, these are cleared.
[0057]
Step 1422: When the reading character string is not stored in the preceding paragraph reading character string storage unit, it is confirmed whether or not there is a determined character string in front of the current cursor. If YES, the process proceeds to step 1423, and if NO, the process proceeds to step 1413.
Step 1423: Return the determined kanji kana mixed string in front of the cursor to the read character string. The read character string is stored in the previous phrase reading character string storage unit, the kanji kana mixed string returned to the kana is stored in the previous phrase kanji mixed string storage unit, and the process returns to step 1408. In addition, the method of returning to the reading character string may be a method of storing the reading character string when the user has input, or a method of converting the character string into a kana character string by a known method.
[0058]
According to this embodiment, when the cursor is moved and a character string is input, the phrase before the cursor to which the cursor is moved can be read as a character string by reading the kanji kana mixed string. The convenience of the conversion operation can be improved.
Further, even if the cursor is moved, the relationship between the front and the back can be correctly determined, and the phrase segmentation learning can be correctly performed.
[0059]
【The invention's effect】
According to the present invention, if a reading character string consisting of a plurality of phrases is divided and input in units of a phrase and is correctly converted, a correct phrase separation position is learned, and a reading character string of a plurality of phrases input next time is correctly converted into kana-kanji characters. can do.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a configuration of a kana-kanji conversion device of the present embodiment.
FIG. 2 is a block diagram illustrating a configuration of a storage medium according to the present embodiment.
FIG. 3 is a flowchart illustrating a procedure (1) of a kana-kanji conversion process according to the embodiment;
FIG. 4 is a flowchart illustrating a procedure (2) of a kana-kanji conversion process according to the present embodiment.
FIG. 5 is a diagram illustrating a storage state of each storage unit according to a kana-kanji conversion process of the embodiment.
FIG. 6 is a diagram showing stored contents of a phrase segmentation learning dictionary of the present embodiment.
FIG. 7 is a diagram illustrating storage contents (1) of a reading kanji kana mixed character string correspondence table according to the present embodiment.
FIG. 8 is a diagram showing storage contents (2) of a reading kanji kana mixed character string correspondence table of the embodiment.
FIG. 9 is a diagram showing storage contents (3) of a reading kanji kana mixed character string correspondence table of the embodiment.
FIG. 10 is a diagram showing storage contents of a kana-kanji conversion dictionary of the present embodiment.
FIG. 11 is a flowchart illustrating a procedure (3) of a kana-kanji conversion process according to the embodiment;
FIG. 12 is a flowchart illustrating a procedure (4) of a kana-kanji conversion process according to the embodiment;
FIG. 13 is a flowchart illustrating a procedure (5) of a kana-kanji conversion process according to the present embodiment.
FIG. 14 is a flowchart illustrating a procedure (6) of a kana-kanji conversion process according to the present embodiment.
[Explanation of symbols]
1 control unit
2 bus
3 Input section
4 Kana-Kanji conversion part
5 Kana-Kanji conversion dictionary
6. Clause-separated dictionary
7 Display buffer
8 Display
9 Indicator
10 Printing Department
11 Pre-phrase reading character string storage
12 Post-segment reading character string storage
13 Pre-Sentence Kanji Kana Mixing Column Storage
14. Post-sentence kanji kana mixed column storage unit
15 Composite reading character string storage
16 Synthetic kanji kana mixing column storage
17 Reading Kanji Kana Mixing Column Correspondence Table
18 Storage media
19 Storage medium reading unit

Claims (4)

A phrase segmentation learning dictionary that stores phrase segmentation information from the continuous phrase conversion result,
A pre-syllable reading character string storage unit that stores a reading character string already converted to kana-kanji ,
A post-sequence reading character string storage unit that stores a reading character string that is input following the reading character string stored in the pre-phrase reading character string storage unit ;
The reading character string in the post-sequence reading character string storage unit is converted into a single phrase by referring to the phrase-separation learning dictionary for single-phrase conversion or phrase-separation to obtain a kana-kana-kanji string as a kana-kanji conversion result. A conversion unit that converts a combined read character string obtained by combining the read character strings stored in the read character string storage unit and the post-sequence read character string storage unit into a continuous phrase;
Whether the conversion result of the synthesized read character string is two phrases, and whether the read character string of each phrase is the same as the read character strings stored in the preceding phrase read character string storage unit and the subsequent phrase read character string storage unit, respectively. A determination unit for determining whether
If the determining unit determines that the two are not the same, the delimiter information in units of the read character strings stored in the pre-sequence reading character string storage unit and the post-sequence reading character string storage unit is used as the phrase separation information of the synthesized read character string. A kana-kanji conversion device comprising: a phrase separation learning control means for storing the phrase separation learning dictionary . A post-sequence kanji kana-sequence column storage part for storing a kanji-kana-sequence column containing at least one of a kanji or a kana obtained by converting the reading character string of the post-sequence reading character string storage part to a kana-kanji character,
A pre-segmented kanji kana mixing column storage unit that stores a kanji kana mixing column already input and converted to kana kanji,
Each character of the reading character string of the pre-sentence reading character string storage unit and the characters of the reading character string of the post-sentence reading character string storage unit and each character of the kanji kana mixing string of the pre-segment kanji kana mixing column storage unit and A reading kanji kana mixing column correspondence table for storing correspondence,
Further comprising a kanji kana Pepper column converted by the conversion unit and editing unit for deleting portions,
When the character string to be partially deleted by the editing unit is a part corresponding to the contents of the previous phrase kanji kana mixing column storage unit and the content of the previous phrase reading character string storage unit, the target of the previous phrase kanji kana mixing column storage unit In addition to deleting and updating the character string, the corresponding character string in the previous sentence-reading character string storage is deleted and updated based on the contents of the reading kanji-kana kana-sequence column correspondence table. The column is stored in the post-sentence reading character string storage unit, and the kanji kana mixed string converted from the read character string is stored in the post-sentence kanji kana mixed string storage unit,
The determining unit is configured to convert the combined read character string obtained by combining the read character strings stored in the preceding phrase read character string storage unit and the subsequent phrase read character string storage unit into a continuous phrase conversion, and the conversion result obtained is two phrases. And it is determined whether the reading character string of each phrase is the same as the reading character strings stored in the pre-phrase reading character string storage unit and the post-phrase reading character string storage unit, respectively ,
When the determining unit determines that the two are not the same, the phrase segmentation learning control unit synthesizes the segmentation information in units of the read character strings stored in the preceding phrase read character string storage unit and the subsequent phrase read character string storage unit. 2. The kana-kanji conversion device according to claim 1, wherein the kana-kanji conversion unit stores the kana-kanji conversion information as the kana-kanji information of the reading character string . A method of performing a kana-kanji conversion by performing a single phrase conversion or a continuous phrase conversion on a reading character string input for each single phrase or a plurality of phrases,
The reading character string already converted to Kana-Kanji is stored in the previous phrase reading character string storage unit,
A reading character string input following the reading character string stored in the preceding phrase reading character string storage unit is stored in the subsequent phrase reading character string storage unit,
The reading character string of the post-syllable reading character string storage unit is referred to the phrase separation learning dictionary for single phrase conversion or phrase separation to perform continuous phrase conversion to obtain a kanji kana mixed sequence as a kana kanji conversion result,
A continuous reading conversion of a combined reading character string obtained by combining the reading character string stored in the preceding phrase reading character string storage unit and the reading character string stored in the subsequent phrase reading character string storage unit ,
Whether the conversion result of the synthesized read character string is two phrases, and whether the read character string of each phrase is the same as the read character strings stored in the preceding phrase read character string storage unit and the subsequent phrase read character string storage unit, respectively. Judge
If it is determined that they are not the same, the delimiter information in units of the reading character strings stored in the pre-phrase reading character string storage unit and the post-phrase reading character string storage unit is used as the phrase delimiter information of the composite reading character string. Kana-Kanji conversion method to store in learning dictionary . On the computer,
A kana-kanji conversion program for converting a reading character string inputted for each single phrase or plural phrases into a single phrase or a continuous phrase,
The reading character string already converted into kana-kanji is stored in the preceding phrase reading character string storage unit,
A reading character string input following the reading character string stored in the preceding phrase reading character string storage unit is stored in the subsequent phrase reading character string storage unit,
The reading character string of the post-sequence reading character string storage unit is referred to the phrase separation learning dictionary for single phrase conversion or phrase separation to perform continuous phrase conversion to obtain a kanji kana mixed string as a kana kanji conversion result,
The combined reading character string formed by combining the reading character string stored in the preceding phrase reading character string storage unit and the reading character string stored in the subsequent phrase reading character string storage unit is subjected to continuous phrase conversion,
Whether the conversion result of the synthesized read character string is two phrases, and whether the read character string of each phrase is the same as the read character strings stored in the preceding phrase read character string storage unit and the subsequent phrase read character string storage unit, respectively. To determine
If it is determined that they are not the same, the delimiter information in units of the reading character strings stored in the pre-phrase reading character string storage unit and the post-phrase reading character string storage unit is used as the phrase delimiter information of the composite reading character string. A storage medium storing a kana-kanji conversion program to be stored in a learning dictionary .

Images