JP3814000B2 - Character string conversion apparatus and character string conversion method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a character string conversion device, and more particularly to conversion of a Roman character mixed character string registered in a dictionary.
[0002]
[Related Art and Problems to be Solved by the Invention]
In Kana-Kanji conversion, an English character input Kana-Kanji conversion method for performing Kana-Kanji conversion by inputting English characters is known. In this alphabet input Kana-Kanji conversion method, when an English character mixed character string is input, there is a method of inputting the spelling of the English character mixed character string as it is and performing reverse conversion later. For example, in order to input an English character mixed character string “original operation”, “original nodousa” is input in the English character input mode, and the conversion key is pressed. Then, it is temporarily converted as “Origin l operation”. In this state, when “Origina 1” which is the part to be reversely converted is designated and the reverse conversion key is pressed, the designated part is reversely converted to “original” and an English character string. Thus, in the character string conversion device, when inputting an English character mixed character string in the English character input mode, a post-conversion function that once inputs an English character with a correct spelling and performs reverse conversion later is employed. ing.
[0003]
However, even if such a post-conversion function is used, it is necessary to perform an inverse conversion by designating a portion to be inversely converted once. Therefore, when an English character mixed character string is input, workability is lower than when a character string in which no English characters are mixed is input.
[0004]
In order to solve this problem, when the post-conversion is performed, it is automatically registered in the dictionary, and from the next time it is converted into English characters just by pressing the conversion key in the same way as normal kanji. Is also possible. However, even if such a dictionary function is used, there is a problem that it cannot be accurately converted in the following cases.
[0005]
For example, if the converted character string “original” is registered in the dictionary for the input character string “Origina 1”, enter “originallande” to convert it to the character string “in original”. Suppose that the conversion key is pressed. In this case, the input character string is understood as a kana character string “Originana de” based on a predetermined Roman reading rule (in this case, “la” = “a”). It will be converted to the string “Folding technique”.
[0006]
An object of the present invention is to solve the above problems and to provide a character string conversion device and a conversion method capable of accurately and easily converting an English character string input in a Roman character input mode.
[0007]
[Means for Solving the Problems]
A character string conversion device according to the present invention includes:A) A dictionary means for storing a post-conversion character string corresponding to this pre-conversion character string as a pre-conversion character string including English characters, and B) an English character string input in the Roman character input mode. Based on reading rulesRomaji kana conversion processConversion means for generating a reading character string and searching the dictionary means to output an alphabetic character string corresponding to the character string before conversion when the reading character string exists in the reading character string In the character string conversion device provided, C) the conversion means includes:A flag indicating the break of the Roman character kana conversion process is stored for the input English character string, and an English character mixed character string is obtained by performing a Roman character kana conversion process at a position that is not the break for the reading character string. When the character string mixed character string acquisition process is performed and the dictionary means is searched and the English character mixed character string exists in the pre-conversion character string, the corresponding English character string is set as a post-conversion character string. Output.
  Thereby, a reading character string including English characters can be obtained from the input character string. Therefore, the English character string input in the Roman character input mode can be converted accurately and easily.
[0008]
In the character string conversion device according to the present invention, the English character mixed character string acquisition process is repeatedly executed by sequentially shifting the position from the last or frontmost character string in the reading character string. Thereby, a reading character string including English characters can be obtained from the input character string.
[0009]
In the character string conversion device according to the present invention, the conversion means includes determination rule storage means for storing a determination rule as to whether or not a conversion candidate character string is preferable as a conversion candidate, and the conversion based on the determination rule When it is determined that the candidate character string is not preferable as a conversion candidate, a reading character string including an English character in the input English character string is obtained. Therefore, it is possible to obtain a kana character string including an English character in the input English character string simply by referring to the English character string storage means.
[0010]
In the character string conversion method according to the present invention, A) a dictionary means for storing a post-conversion character string corresponding to this pre-conversion character string as a pre-conversion character string including an English character, and B) a Roman character input mode The English character string input in is subjected to a romaji kana conversion process based on the roman character reading rules to generate a read character string, and the dictionary means is searched for and the pre-conversion character string exists in the read character string A character string conversion method using a character string conversion device provided with conversion means for outputting a corresponding English character string as a converted character string, and C) the conversion means is subjected to the following processing: To perform, c1) A flag indicating the break of the Roman character kana conversion process is stored for the input English character string, and a Roman character kana conversion process is performed on the read character string at a position that is not the break to obtain an English character mixed character string. , c2) When the dictionary means is searched and the English character mixed character string exists in the pre-conversion character string, the corresponding English character string is output as a post-conversion character string.
[0011]
In the character string conversion method according to the present invention, the English character mixed character string acquisition process is repeatedly executed by sequentially shifting the position from the last or frontmost character string in the reading character string. Thereby, a reading character string including English characters can be obtained from the input character string.
[0013]
A storage medium according to claim 7 is a computer-readable storage medium storing a computer-executable program, and the program realizes the apparatus or method according to any one of claims 1 to 6. It is characterized by being.
[0014]
【The invention's effect】
In the character string conversion device or the character string conversion method according to any one of Claims 1 and 6, when the reading character string is generated, a part of the generated reading character string is not limited to the Roman character reading rule. For a character string, a reading character string including an English character in the input English character string is obtained, and whether or not the pre-conversion character string exists in the reading character string including the English character in the English character string. to decide. Thereby, a reading character string including English characters can be obtained from the input character string. Therefore, the English character string input in the Roman character input mode can be converted accurately and easily.
[0015]
3. The character string conversion apparatus according to claim 2, wherein the conversion means includes determination rule storage means for storing a determination rule as to whether or not a conversion candidate character string is preferable as a conversion candidate, and the conversion based on the determination rule. When it is determined that the candidate character string is not preferable as a conversion candidate, a reading character string including an English character in the input English character string is obtained. Therefore, only when the conversion candidate character string is not preferable as a conversion candidate, a reading character string including the English characters in the input English character string can be obtained. Thereby, conversion efficiency can be improved.
[0016]
4. The character string conversion apparatus according to claim 3, further comprising an English character string storage unit that stores the input English character string, wherein the conversion unit is configured to convert the English character string for a part of the generated character string. With reference to the storage means, a kana character string including an English character in the input English character string is obtained. Therefore, it is possible to obtain a kana character string including an English character in the input English character string simply by referring to the English character string storage means.
[0017]
5. The character string conversion apparatus according to claim 4, wherein the conversion means includes an English character string in the input English character string when a forced classification instruction command is added to the input English character string. Get a kana string containing. Therefore, when it matches the intention of the character string input person, it is possible to obtain a reading character string including the English characters in the input English character string. Thereby, conversion efficiency can be improved.
[0018]
6. The character string conversion apparatus according to claim 5, wherein the conversion means is adapted to read the roman character as a kana character string that is the same as the kana character string located before the English character in the pre-conversion character string stored in the dictionary means. It is determined whether or not it exists in the reading character string generated based on the rule, and if it exists, for the kana character string that continues in the generated reading character string, the corresponding English character is obtained, Judge whether both English character strings match. If they match, a reading character string including the English characters in the input English character string is obtained. Therefore, when a kana character string including the pre-conversion character string stored in the dictionary means exists in the reading character string, a reading character string including English characters can be obtained from the input character string. . Thereby, the English character string input in the Roman character input mode can be converted accurately and easily.
[0019]
BEST MODE FOR CARRYING OUT THE INVENTION Functional block diagramLight
BookAn embodiment of the invention will be described with reference to the drawings. The character string conversion device 1 shown in FIG. 1 includes an input unit 41, an input / output control unit 42, a display unit 43, an output unit 44, a character string storage unit 63, a conversion unit 50, and a dictionary unit 70.
[0020]
Various commands and alphanumeric character strings to be converted are input to the input means 41. The input / output control means 42 gives the input alphanumeric character string to the character string storage means 63. The character string storage unit 63 includes an English character string storage unit 64, a reading character string storage unit 65, and a written character string storage unit 66. The input alphanumeric character string is stored in the English character string storage means 64. The reading character string storage unit 65 stores the reading character string converted by the conversion unit 50. The character string stored in the reading character string storage means 65 may be a kana character string only, an alphanumeric character string only, or a combination of both. The written character string storage unit 66 stores the written character string converted by the conversion unit 50. The notation character string may be a kana character string, a kanji character string, or an alphanumeric character string.
[0021]
When the input / output control means 42 receives the output command, the input / output control means 42 outputs the written character string stored in the written character string storage means 66 to the display means 43. The display means 43 displays this notation character string. The output unit 44 outputs the written character string stored in the written character string storage unit 66.
[0022]
The dictionary means 70 includes grammatical information storage means 71, word information storage means 72, co-occurrence example information storage means 73, and learning information storage means 74.
[0023]
The grammatical information storage unit 71 stores information about correctness of grammatical connection between words, information about the strength of the connection, and the like. For example, information such as “noun (no particle) + suffix is strong, noun (no particle) + noun is strong” is stored.
[0024]
The word information storage means 72 stores the pre-conversion character string and the corresponding post-conversion character string. The pre-conversion character string includes the case of only Kana characters, only English characters, or Kana characters mixed with alphanumeric characters. The converted character string includes kanji, katakana, English characters, and a combination character string thereof. Specifically, a reading character string, a written character string, part-of-speech information, utilization information, and the like of each word are stored.
[0025]
The co-occurrence example information storage unit 73 stores binary relation information between words having strong semantic connections. This co-occurrence example information includes co-occurrence examples of attribute units such as “person” and “flower” and additional restriction information, in addition to two connections. Additional restriction information includes, for example, particle information such as “O” and “GA”, information on the direction of establishment, and the like. As examples of co-occurrence, “hot-summer”, “thick-book”, “hot-hot water”, and the like are stored as word ties. As an example of attribute-based co-occurrence, “Meeting a person (he, girlfriend, teacher, lover, etc.)” is stored. As an example of co-occurrence in attribute units, “flowers (tulips, chrysanthemums, etc.) are blooming” and the like are stored. As the particle restriction information, “speak-listen”, “medicine-effective”, “motivate-effective”, etc. are stored. “Home-education” and “education-process” are stored as direction restriction information.
[0026]
Such co-occurrence example information is semantic information, and the use of this information eliminates the ambiguity of homophones. For example, an example of a homonym word based on the co-occurrence example information 73 indicates that, when the conversion process is performed on the reading “hot / hon”, the “thick-book” information in the co-occurrence example information 73 indicates “ “Thick” is selected, and “hot, hot” or the like is not selected. Therefore, the conversion result of “thick / book” is obtained immediately. Thus, the co-occurrence example information is used to improve conversion efficiency.
[0027]
The learning information storage means 74 is usage information when a recently used notation is preferentially adopted among a plurality of notations.
[0028]
The conversion unit 50 includes a reading character string generation unit 51, a reading rule storage unit 52, a basic analysis unit 53, a phrase delimiter processing unit 54, an automatic dictionary registration unit 55, a notation selection unit 56, and a Roman character delimiter changing unit 58. .
[0029]
The reading rule storage means 52 stores a Roman character reading character string for an English character string. The reading character string generation means 51 generates a reading character string from the English character string stored in the English character string storage means 64 based on the Roman character reading rules stored in the reading rule storage means 52, and stores the reading character string storage. Store in means 65.
[0030]
The basic analysis means 53 estimates the segment break position with reference to the grammar information storage means 71 and the word information storage means 72 of the dictionary means 70.
[0031]
The phrase break processing means 54 checks the grammar / word information for the phrase break candidate output from the basic analysis means 52 and performs a homophone selection process with reference to the co-occurrence example information storage means 73. Further, the phrase break processing means 54 narrows down the phrase break candidates.
[0032]
The notation selection means 56 determines the notation for the phrase whose notation has not been determined by the processing of the phrase delimiter processing means 54. The notation selection means 56 has reverse Roman character conversion means 57 for converting the word portion of the extracted phrase from a kana character string to a Roman character string. The reverse Roman character conversion means 82 can also change the display character size.
[0033]
When the word determined by the notation selection means 56 is not registered in the word information storage means 72 of the dictionary means 70, the automatic dictionary registration means 55 displays information such as the reading of this word, the notation character string, and the part of speech. It is automatically registered in the word information storage means 72.
[0034]
The Roman character delimiter changing means 58 has a judgment rule storage means 59, and changes the Roman character delimiter as follows. The determination rule storage means 59 stores a determination rule as to whether or not a conversion candidate character string is preferable as a conversion candidate. Based on this determination rule, the Roman character delimiter changing means 58 refers to the English character string stored in the English character string storage means for some characters of the read character string stored in the read character string storage means 65. Thus, a kana character string including the English characters in the input English character string is obtained. For the kana character string including the obtained English character, the basic analysis means 53 and the phrase delimiter processing means 54 perform character string conversion again.
[0035]
With this configuration, the conversion means 50 generates a reading character string based on the Roman character reading rule for the English character string input in the Roman character input mode, and searches the dictionary means to search for the character string in the reading character string. If there is a pre-conversion character string, the corresponding English character string is output as the post-conversion character string. Further, when it is determined that the conversion candidate character string is not preferable as a conversion candidate based on the determination rule, the kana character string that is a part of the generated reading character string without being caught by the Roman character reading rule For the above, the dictionary means can be searched by obtaining a reading character string including an English character in the input English character string.
[0036]
In the present embodiment, the determination rule storage means 59 is provided, and the Roman character separation process is performed only when necessary. However, the Roman character separation process may be performed for all of them. .
[0037]
2. Hardware structureCompletion
FigureFIG. 2 shows an example of a hardware configuration in which the character string conversion device 1 shown in FIG. 1 is realized using a CPU.
[0038]
The character string conversion device 1 includes a CPU 23, a memory 27, a hard disk 26, a CRT 30, an FDD 25, a keyboard 28, a mouse 31, and a bus line 29. The CPU 23 controls each unit via the bus line 29 according to a control program stored in the hard disk 26.
[0039]
This control program is read from the flexible disk storing the program via the FDD 25 and installed in the hard disk 26. In addition to the flexible disk, it may be installed on a hard disk from a computer-readable storage medium in which a program such as a CD-ROM or an IC card is substantially integrated. Furthermore, it may be downloaded using a communication line.
[0040]
In the present embodiment, the program stored in the flexible disk is indirectly executed by the computer by installing the program from the flexible disk to the hard disk 26. However, the present invention is not limited to this, and the program stored in the flexible disk may be directly executed from the FDD 25. Note that programs that can be executed by a computer are not only those that can be directly executed by simply installing them, but also those that need to be converted to other forms once (for example, those that have been compressed) are decompressed. Etc.), and further executable in combination with other module parts.
[0041]
The hard disk 26 stores dictionary data. The memory 27 stores various calculation results and the like. Conversion candidates and the like are displayed on the CRT 30.
[0042]
The display on the CRT 30 will be described with reference to FIG. In the editing screen 100 of the CRT 30, an area 110 is provided in the lower part of the screen in addition to the area for displaying the text. In the area 110, various messages and states are displayed, and candidates for homonyms are displayed. An enlarged view of the area 110 is shown on the right side of the figure. Area 120 in area 110 is used as an area for displaying word candidates when an instruction for displaying the next candidate group is given.
[0043]
Here, [Insert] means that a character string input from the keyboard 4 is inserted immediately before the cursor 102, and [Confirm] means that the user performs Kana document conversion such as Kana-Kanji conversion and the like. It means that the selected word is not read back later. “Kana-Kana-Kan” indicates that the Kana-Kanji conversion mode is continuous phrase and Kana input.
[0044]
As for the above-mentioned [Insert], [Insert] and [Overwrite] can be appropriately changed by operating a predetermined control key. Similarly, [Kana Han] can be changed mutually with [R Han] for Roman character input.
[0045]
In FIG. 3, when a user inputs a character string from the keyboard 28 and presses a conversion key, a character string conversion process is executed as described later. As a result, a sentence (or phrase) consisting of a high-priority word converted by each reading is displayed in an undetermined state at the position of the cursor 102 on the screen. At this time, when the next candidate key is pressed, a candidate group of the first phrase among the displayed unconfirmed words is displayed in the area 120.
[0046]
In the present embodiment, the space key is associated with the kana document conversion key and the next candidate key. That is, when the space key is pressed immediately after inputting a reading, it is determined as a kana document conversion key. In this state, when the space key is continuously input, it is processed as the next candidate key. When a space key is simply input without inputting a reading, or when the control key, the shift key, and the space key are pressed simultaneously, the input is processed as an original blank character.
[0047]
The confirmation of the candidate is made by pressing the return key at that point or subsequently pressing the reading key if the state when the conversion key is pressed (first candidate) is acceptable. When the candidate group is displayed in the area 120 by pressing the next candidate key, one of the candidates is selected (the candidate group is numbered and the number is entered), and then the return key (confirmed) Confirm by pressing (key). Note that the return key input when the reading is confirmed literally indicates a line feed key.
[0048]
3. FLOWERG
OneNext, a program stored in the hard disk 26 will be described.
[0049]
3.1 Automatic registration and conversion processReason
FigureThe document processing operation will be described with reference to the flowchart of FIG. The user inputs an English character string from the keyboard 28. The CPU 23 determines whether or not an English character string is input (step ST1 in FIG. 4). When the English character string is input, the input English character string is stored in the Roman character buffer 27a (step 4 in FIG. 4). ST3). Next, it is determined whether or not the input alphabetic character string can be read based on the following Roman character reading rules (step ST5).
[0050]
Romaji reading rules
Rule 1: If the roman character string is a character string registered in the romaji kana conversion table, the corresponding romaji character string is converted to a kana character.
[0051]
Rule 2: When English consonant characters are continuously input, the first English character is converted to “”, and then Rule 1 is repeated.
[0052]
CPU23 performs kana conversion, when a Roman character can be read (step ST7). The converted kana character is stored in the reading character string buffer 27b (step ST9). For example, when an English character string as shown in FIG. 5A is input, since it is a vowel when “DA” is input, kana conversion processing is performed, and “DA” is stored in the read character string buffer 27b. Is memorized.
[0053]
By repeating such processing, a reading character string is generated from the input English character string. For example, when the above-described Roman Kana-Kana conversion rule is applied to the character string shown in FIG. 5A, a reading character string as shown in FIG. 5B is generated and stored in the reading character string buffer 27b.
[0054]
In this way, data is sequentially stored in the Roman character buffer 27a and the reading character string buffer 27b. In the case of English letters, numbers, and symbol keys that cannot be read in Roman letters, they are stored in the read character string buffer 27b as they are.
[0055]
The CPU 23 determines whether or not the conversion key is operated (step ST11), and steps ST1 to ST9 are repeated as long as the conversion key is not operated. In this state, when the user operates the conversion key, the CPU 23 determines that the conversion key has been operated, and performs the phrase break and dictionary conversion processing (step ST13).
[0056]
The process of step ST13 is the same as the conventional phrase segmentation and dictionary conversion process. Briefly, the phrase candidate of the reading character string is extracted from the data in the reading character string buffer 27b by the data (word information, grammatical information, etc.) stored in the dictionary unit 26b. Furthermore, using the other data stored in the dictionary unit 26b, the extracted phrase candidates are checked grammatically and semantically to narrow down the phrase candidates. Along with such narrowing down, a phrase break process is performed.
[0057]
Next, the CPU 23 tentatively determines conversion candidates by the minimum cost method, and stores the tentatively determined conversion candidates in the written character string buffer 27c (step ST15). In addition to the minimum cost method, a known method may be used.
[0058]
Next, the CPU 23 determines whether or not a Roman character break changing process is necessary (step ST19 in FIG. 8). The Romaji break change process will be described later. If it is determined that the Romaji break changing process is not necessary, the character string stored in the display character string buffer 27c is displayed on the CRT 30 (step ST21).
[0059]
The user determines whether the displayed conversion result is correct or not, and if it is correct, operates the confirmation key. If you determine that the paragraph break is incorrect,
Operate the phrase break correction key. If the phrase break is correct but the wrong homonym is displayed, the next candidate key is operated.
[0060]
The CPU 23 determines which key has been operated in step ST23. If the phrase break correction key is operated, the phrase break position is changed in accordance with the contents of the key (step ST25), and the processing from step ST13 onward in FIG. 4 is repeated.
[0061]
When the next candidate key is operated, the CPU 23 performs reverse romaji conversion processing (step ST27). The reverse romaji conversion process will be briefly described. Here, it is assumed that the reverse Roman character conversion process is performed on the “date” portion of FIG. 5C. The CPU 23 converts “date” into an English character string with reference to the Roman character buffer 27a. Then, a first candidate (“DATE”) expressed in full-width and a second candidate (“date”) expressed in half-width size are generated.
[0062]
In this way, the input English character string can be converted by performing the reverse romaji conversion process.
[0063]
Next, in addition to these two candidates, the CPU 23 combines the third to seventh candidates read from the dictionary unit 26b and presents them in the candidate display area 120 of the CRT 30 (see FIG. 5D). In this way, the character string that has been subjected to reverse romaji conversion is displayed together with the candidates registered in the dictionary unit 26b and presented as a user's selection target.
[0064]
The user gives a candidate specifying command for selecting “1.DATE” from the next candidates presented in the area 120. When such a candidate specifying command is given, the CPU 23 changes the display of the area 120 as shown in FIG. 5E and displays the change candidates as shown in FIG. 5F.
[0065]
Since the desired conversion result is obtained, the user operates the confirmation key. Thereby, the CPU 23 performs a confirmation process (step ST31). As a result, the character string shown in FIG. 5F is stored in the written character string buffer 27c.
[0066]
Next, the CPU 23 determines whether or not a character string that has been subjected to reverse Romaji conversion exists in the determined character string (step ST33). When it exists, CPU23 memorize | stores the said character string by which the reverse Roman character conversion was carried out in the dictionary part 26b. In this case, as shown in FIG. 9, the CPU 23 reads and adds the character string “DATE” as “date” and updates the learning information.
[0067]
In this way, a new word is automatically registered in the dictionary unit 26b without requiring any user dictionary registration operation. Therefore, for example, when the user next inputs the characters “To date, enter date” and press the conversion key, “Enter DATE to check the date. ”Can be converted correctly and at once. In other words, even if you do not prepare a Japanese-English dictionary or an English-Japanese dictionary, you can enter foreign words as they are and mix them in kana-kanji mixed sentences. There is no need for new registration by the operation.
[0068]
As described above, in the present embodiment, when a word that is not registered in the dictionary unit 26b is detected, these kana character strings are reversely converted into Roman character strings at the time of key input as conversion candidates. Are converted into uppercase full-width, lowercase half-width, uppercase half-width, lowercase full-width, etc., and presented to the user as one of the next candidates. Furthermore, when a character string displayed as the original kana character string or a character string obtained by converting hiragana into katakana is presented as one of the next candidates, the user's selection range can be expanded.
[0069]
If a character string that has been subjected to reverse romaji conversion is preferentially presented as the next candidate presentation order, the user can handle Japanese and foreign languages without further distinction. In addition, whether the reverse Roman alphabet conversion to English capital letters first, the conversion to English characters first, the full-width notation first, the half-width notation character string to be displayed as one of the next candidates, etc. If it is possible to customize by selection, it becomes easier for the user to use reverse Romaji conversion.
[0070]
When a candidate character string that has been converted into a reverse romaji is adopted as a confirmed character string, it can be automatically and sequentially registered in the dictionary unit 26b, and this function creates a Japanese sentence in which Japanese and foreign languages are mixed. In this case, a new foreign language can be registered in the dictionary without requiring the user to perform any mode conversion key or dictionary registration operation.
[0071]
As described above, in this embodiment, when the next candidate key is pressed, the character string returned to the input English character is displayed as the next candidate target. Thereby, the user can obtain a desired English character string only by operating the conversion key, without operating the post-conversion key (the key for converting the kana-converted character string into the English character string again). . Further, when such an English character string is confirmed, it is automatically registered. This eliminates the need for new registration in the dictionary.
[0072]
In the present embodiment, the automatic registration process is performed when registering in the dictionary. However, the present invention is not limited to this, and the user operates the post-conversion key to convert it into an English character string, and the conversion is confirmed. At that time, the dictionary may be registered.
[0073]
3.2 Romaji separatorReason
OneNext, a description will be given of the process of changing romaji separators. Romaji delimiter change processing means that when reading an input character string in Roman letters, the input character string is converted to a character string registered in the dictionary because the character is read in a part different from the user's intention. Romaji reading process without being bound by the Romaji reading rules.
[0074]
For example, if “ATOK” is registered in the dictionary and “atokyourei” is input to convert it to “example for ATOK”, the Roman reading of the above-mentioned Roman character will result in “Romantic reading” and the user ’s It will deviate from the input intention. By performing the Romaji break changing process, it is possible to read the Romaji as intended by the user and to obtain a correct conversion result.
[0075]
In step ST19 in FIG. 8, when there is a high possibility that the Romaji break changing process is necessary, the CPU 23 performs the processes after step ST41 in FIG. In the present embodiment, when there is a one-character phrase, the CPU 23 determines that there is a high possibility that the Romaji break changing process is necessary.
[0076]
For example, when “atkyourei” is input to convert “ATOK example”, the text “Atokyorei” is read as romaji and converted to “after, strong cold” with reference to the dictionary shown in FIG. In this case, since there is a one-character phrase “after” in the character string that has been converted to Kanji, it is determined in step ST19 that Roman character delimiter change processing is necessary.
[0077]
The CPU 23 re-divides the kana characters of interest in the corresponding phrase with English characters whose flag rmpos is “0”. The flag rmpos is a flag for discriminating whether an input English character string is a vowel or a consonant when it is read in Roman characters. In this step, a portion that can be converted into a Roman character at the stage of reading the input character string is set as a flag rmpos [1], and the other portion is set as a flag rmpos [0]. In addition, the kana character of interest is initially a kana character that is the first after the clause and includes the flag rmpos [0].
[0078]
More specifically, as shown in FIG. 10A, when “atkyourei” is input, the flag rmpos is set to “1,0,1,0,0,1,1,0, 1, 1 ". In this case, the kana character of interest is “Re” of “A, To, Ki, Cho, U, Re, I”. Therefore, in this case, “re” is converted into an English character string “r”, “i” and converted into an English character mixed reading character string “Ato-r” as shown in FIG. 10D.
[0079]
The CPU 23 determines whether or not this “Ato-r” is registered in the dictionary unit 26b (step ST43). If it is not registered, it is determined whether or not the next separation is possible (step ST45). If it is possible, the process of step ST41 is repeated.
[0080]
In this case, since the next delimiter is possible, kana conversion is performed with the kana character including the next flag rmpos [0]. As a result, “after ky” is obtained as shown in FIG. 10E. The CPU 23 determines whether or not “after ky” is registered in the dictionary unit 26b (step ST43). If it is not registered, it is determined whether or not the next separation is possible (step ST45). If it is possible, the process of step ST41 is repeated.
[0081]
In this case, since the next delimiter is possible, kana conversion is performed with the kana character including the next flag rmpos [0]. As a result, “k” is obtained as shown in FIG. 10F. The CPU 23 determines whether or not “after k” is registered in the dictionary unit 26b (step ST43). In this case, since “k” is registered as shown in FIG. 7, a Roman character kana conversion process is performed at the position (step ST 49), and the processes after step ST 13 in FIG. 4 are performed.
[0082]
By performing the process of step ST13, the input character string “atokyourei” is converted to “example for ATOK”.
[0083]
If the next delimiter is not possible in step ST45, it is determined that the Roman delimiter changing process is unnecessary, and the conversion candidate provisionally determined in step ST15 in FIG. 4 is set as a conversion candidate (step ST47). Then, the processing from step ST21 onward in FIG. 8 is performed.
[0084]
In this way, a character string including English characters can be accurately converted. In this embodiment, the search is performed in the long reading order until the search is successful. However, the search may be performed in the short reading order. Furthermore, if the search is successful, the processing is stopped, but if possible, the search may be performed for all, and the highest evaluation may be selected.
[0085]
In the present embodiment, when there is a one-character phrase, it is determined that there is a high possibility that the Romaji break changing process is necessary, and the Romaji break changing process is performed. A case where there is a high possibility that the delimiter changing process is necessary may be set, and the same process may be performed. For example, there is a clause of an unregistered word.
[0086]
As an unregistered word clause, for example, “original” is registered in the dictionary, and “original appli” is input to convert it to “original app”. In this case, Kana conversion is performed as “Originapuri”, and Kanji conversion is performed, for example, “Folding technique Napuri”. In this case, “Naapuri” is a phrase of an unregistered word. In such a case as well, it can be determined that there is a high possibility that the Romaji break changing process is necessary as in the case where the one-character phrase is present.
[0087]
4). Other implementationsstate
UpIn the embodiment, it is determined whether or not the Roman character break changing process is necessary after the Kanji conversion, but it is determined whether or not the Roman character break changing process is necessary in a state where the character is input. Also good. For example, when a character is input in a state where the shift key is pressed, it is determined that there is a high possibility that the Roman character delimiter changing process is necessary, and the Roman character reading rule is converted from the first English character string. It is possible to decide to perform separation processing that is not restricted by
[0088]
For example, if “ATOK” is registered in the dictionary, and “KonoAtokyoiwomotiteite” and “A” are input using the shift key to convert “Using this example for ATOK”, Romaji kana conversion.
[0089]
The CPU 23 performs Roman character kana conversion from the character “A” input using the shift key to a portion that does not belong to the alphabetic break. That is, of “konoAtokyoiwomotite”, “Atokyouei” is delimited by the character string “atkyoour” up to a part that does not belong to the delimiter of alphabetic characters, and is converted into “Atokyo”. Then, it is determined whether or not this “Ato-r” is in the dictionary.
[0090]
If not, re-separation processing is performed in the same manner as described above. In this manner, a character string including English characters can be reliably converted without a user command.
[0091]
In this case as well, it can be performed in the order of short reading.
[0092]
In this way, when an English character string is input to the input means, if a compulsory classification instruction command is added, a kana character string including an English character string in the input English character string is obtained. It may be.
[0093]
Moreover, you may make it obtain the reading character string containing the English character in the said input English character string as follows. In the character string existing in the dictionary, it is determined whether or not there is a portion including a kana up to the corresponding English character in the character string including the English character at the stage of converting the Roman character. Then, if it exists, it is determined whether or not the following letter in the input character matches the following letter in the character string stored in the dictionary. Do.
[0094]
For example, when “after k” is registered, when “atokyourei” is input, “Akiyorei” is converted into a Roman character. It is determined whether or not there is a character string of the “after” part of “after k” registered in the dictionary (hereinafter referred to as “kana common part”) in such input characters. In this case, since there is a kana common part, the alphabetic characters that follow the kana common part are extracted for “Akiyorei” in the input character string. As for the alphabetic characters following the kana common part, all combinations up to the part before the next vowel part are extracted. Therefore, in this case, two types “k” and “ky” are extracted. As for the characters registered in the dictionary, the alphabetical character following the kana common part is “k”. In this case, since “k” is considered to be the same, “atokyourei” is divided into “atok” and “yourei” in this case.
[0095]
In this case, since it is inefficient to perform such processing for all locations, such processing may be performed only when there is the one-character phrase or when there is an unregistered word.
[0096]
In the above embodiment, the Roman character delimiter changing process is executed by referring to the Roman character buffer and returning to the English character. However, the Roman character reading rule in FIG. You may make it perform.
[0097]
In the present embodiment, the CPU 23 is used to realize the function shown in FIG. 1, and this is realized by software. However, some or all of them may be realized by hardware such as a logic circuit.
[Brief description of the drawings]
FIG. 1 is a functional block diagram of a character string converter 1 according to the present invention.
FIG. 2 is a diagram illustrating an example of a hardware configuration of the character string conversion device 1 illustrated in FIG. 1;
FIG. 3 is a diagram showing a display on a CRT 30;
FIG. 4 shows a flowchart of conversion processing.
FIG. 5 is a diagram showing data contents and display examples of a Roman character buffer, a Kana buffer, and a notation character buffer.
FIG. 6 is a diagram showing Roman character reading rules.
FIG. 7 is a diagram showing the contents of a dictionary unit 26b.
FIG. 8 shows a flowchart of a conversion process.
FIG. 9 is a diagram showing the contents of word data added to the dictionary unit 26b.
FIG. 10 is a diagram showing data contents of a Roman character buffer and a Kana buffer.
FIG. 11 is a flowchart of Roman character segment change processing.
[Explanation of symbols]
63 ... Character string storage means
50 ... Conversion means
70 ... dictionary means
23 ... CPU
27 ... Memory

Claims (5)

Dictionary means for storing a post-conversion character string corresponding to the pre-conversion character string as a pre-conversion character string including an English character;
A Roman character kana conversion process is performed on an English character string input in the Roman character input mode based on a Roman character reading rule to generate a read character string, and the dictionary means is searched for the pre-conversion character in the read character string. If the column exists, conversion means for outputting the corresponding character string as a converted character string,
In a character string conversion device comprising:
The conversion means stores a flag indicating a break in the Roman character kana conversion process for the input English character string, and performs a Roman character kana conversion process at a position that is not the break for the reading character string. In addition to executing an English character mixed character string acquisition process for obtaining a character string, if the English character mixed character string exists in the pre-conversion character string by searching the dictionary means, the corresponding English character string is Output as a string after conversion,
String conversion apparatus according to claim. The character string conversion device according to claim 1,
The English character mixed character string acquisition processing is repeatedly executed by sequentially shifting the position from the last or frontmost character string in the reading character string,
Character string converter characterized by the above. In the character string conversion device according to claim 1 or 2,
The conversion unit includes a determination rule storage unit that stores a determination rule as to whether or not a conversion candidate character string is preferable as a conversion candidate, and determines that the conversion candidate character string is not preferable as a conversion candidate based on the determination rule. And obtaining a reading character string including an English character in the input English character string,
Character string converter characterized by the above.   Dictionary means for storing a post-conversion character string corresponding to the pre-conversion character string as a pre-conversion character string including an English character;
  An English character string input in the Roman character input mode is subjected to a Roman character kana conversion process based on a Roman character reading rule to generate a read character string, and the dictionary means is searched for the pre-conversion character in the read character string. If the column exists, conversion means for outputting the corresponding character string as a converted character string,
  A character string conversion method using a character string conversion device comprising:
  Causing the conversion means to execute the following processing;
  A flag indicating the break of the Roman character kana conversion process is stored for the input English character string, and a Roman character kana conversion process is performed on the read character string at a position that is not the break to obtain an English character mixed character string. ,
  When the dictionary means is searched and the English character mixed character string is present in the pre-conversion character string, the corresponding English character string is output as a post-conversion character string.
  A character string conversion method using a character string conversion device characterized by the above. In the character string conversion method according to claim 4,
The English character mixed character string acquisition processing is repeatedly executed by sequentially shifting the position from the last or frontmost character string in the reading character string,
Character string conversion method characterized by

Images