JP3386147B2 - Roman character converter - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a Roman character conversion device for converting Roman character notation into Kana notation. In the case of inputting Japanese characters in Roman characters, not only one type of Roman characters but also a case where it is not possible to convert a commonly used Roman character into a desired Kana character. For example, looking at the case of a person's name, the kana notation "Ito" is "ITO", "ITO"
U "and" ITOH ". It is desired to convert to the correct kana notation including those used in the Roman notation. 2. Description of the Related Art Conventionally, when inputting kanji in Roman characters by word processing input or the like, usually, roman alphabet notation and kana notation are one.
It corresponds to one to one. However, notation such as a personal name is not always a single notation, and as shown in FIG. 9, there are several Roman notations, and these may correspond to one kana notation. In this way, Roman notation is
It does not necessarily correspond to kana notation one-to-one, and furthermore,
It is not necessarily the phonetic transcription itself. Therefore, for example, when "ITO" is input in Roman characters and converted to kana, "ito" is generated, and "ito" cannot be generated. [0004] As described above, when the Roman alphabet notation is converted to the kana notation using the conventional Roman alphabet kana conversion algorithm, there is a problem that some of the characters are not converted to the desired kana notation. there were. This is because Roman notation was converted to kana notation as a simple phonetic notation, and it is desired to resolve this and convert Roman notation to appropriate kana notation. SUMMARY OF THE INVENTION It is an object of the present invention to obtain a plurality of Roman alphabet notations by expanding the Roman alphabet notation, and to suppress unnecessary ones from the plurality of Roman alphabet notations to generate an appropriate appropriate Roman alphabet notation. . Means for solving the problem will be described with reference to FIG. In FIG. 1, an extended Roman character conversion device 2 converts a Roman alphabet notation 1 into a Roman alphabet expansion rule 3 and a Roman alphabet suppression rule 4.
Is applied. [0007] The Roman alphabet extension rule 3 is a rule for sequentially extracting characters from the beginning of the Roman alphabet notation and extending them to a plurality of Roman alphabet notations. The Roman character suppression rule 4 is a rule for determining whether or not to suppress characters by sequentially extracting characters from the head of a plurality of Roman character expressions expanded by the Roman character expansion rule 3. According to the present invention, as shown in FIG. 1, for an input Roman alphabet notation 1, an extended Roman alphabet conversion device 2 sequentially extracts characters from the beginning and applies a Roman alphabet extension rule 3 to form a plurality of Roman alphabet notations. To be expanded. In addition, the extended Roman alphabet conversion device 2 sequentially extracts characters from the beginning of the input Roman alphabet notation 1 and applies them to a plurality of Roman alphabet notations by applying the Roman alphabet extension rule 3, and further suppresses the expanded Roman alphabet notation. Only the Roman alphabet notation determined not to be suppressed by applying rule 4 is generated. Then, convert these extended or generated Roman alphabet notation to kana,
Based on the converted kana, the kana kanji dictionary is searched and converted to kanji. Therefore, by expanding the Roman alphabet notation to obtain a plurality of Roman alphabet notations, and by suppressing unnecessary ones among the plurality of Roman alphabet notations, an appropriate expanded Roman alphabet notation from the input Roman alphabet notation 1 can be obtained. , And these can be converted to kana notation. Next, the structure and operation of an embodiment of the present invention will be sequentially described in detail with reference to FIGS. FIG. 1 shows a configuration and explanatory diagram of one embodiment of the present invention. FIG. 1A shows a configuration diagram. In FIG. 1 (a), Roman alphabet notation 1 is a Roman alphabet notation that has been input, for example, a Roman alphabet notation input to convert Kana-Kanji to Roman from a keyboard using a word processor or the like, or a foreign language at the time of machine translation. Romanized notation translated into Japanese Romanized characters (for example, personal names, proper nouns, etc.). In particular, when a foreign language is translated into Japanese romaji, it is necessary to register kana (kanji) corresponding to all romaji notations in the kana-kanji conversion dictionary by expanding to a plurality of romaji notations by the present invention. Instead, it is only necessary to register kana (kanji) corresponding to typical Roman alphabet notation, and the number of entries in the dictionary can be reduced. The extended Roman alphabet conversion device 2 applies the Roman alphabet extension rule 3 and the Roman alphabet suppression rule 4 to the input Roman alphabet notation 1 to generate an appropriate extended Roman alphabet notation. The extended Roman character conversion device 2 is configured by applying a Roman alphabet expansion rule 3 and a Roman character control rule 4 to an input Roman alphabet notation 1 on a hardware inference / control engine. The Roman alphabet extension rule 3 is for generating a plurality of Roman alphabets by applying to the input Roman alphabet 1. For example, as shown in the right-hand side of Romaji notation 1 "TOKYO", "TOKYO, TOKYO, TOKYO, TOOK"
YO, TOOKYOO, TOOKYOU, TOUKY
O, TOKYO, TOKYOU ”(1) (to be described later with reference to FIGS. 4 and 5). The Roman alphabet suppression rule 4 applies the Roman alphabet extension rule 3 to the inputted Roman alphabet notation 1 and suppresses unnecessary Roman alphabet notation by expanding the intermediate candidates. Is generated. For example, in FIG.
Applying the Roman alphabet suppression rule 4 and suppressing as described on the right side, as a result, "TOKYO, TOKYO, TOKYO, TOUKO
YOU ”is generated as a candidate for Roman alphabet notation. The Roman alphabet / kana conversion unit 5 converts the Roman alphabet notation of candidates after being suppressed by the Roman alphabet suppression rule 4 into kana. For example, as described on the right side, Romanized notation of four candidates is converted into kana of "Tokyo, Tokyo, Tokyo, Tokyo" as shown in the figure. FIG. 1B shows a specific example. In (b) of FIG. 1, the intermediate candidate is Romanized notation 1 “TOK”.
The roman alphabet notation extended by applying the Roman alphabet extension rule 3 to “YO” is shown. The candidate indicates a Roman alphabet notation in which the Roman alphabet suppression rule 4 is applied to the intermediate candidate Roman alphabet notation. Kana indicates that the Romanized notation of the candidate is converted to kana. FIG. 2 shows an example of a Roman character extension rule according to the present invention. The “rule number” on the left is a rule number. The "target character" in the center is the input Roman alphabet notation 1
Is a character to be extracted one character at a time from the beginning and to be expanded, and to generate an output character on the right side when applicable. The “output character” on the right side is a character to be output for the target character, for example, for outputting two of “I” and “II” when the target character is “I”. FIG. 3 shows an example of a Roman character suppression rule according to the present invention. "Target character" on the left side is Roman extended rule 3.
The Romanized notation of intermediate candidates expanded by applying
This is a character (character string) that is extracted from the beginning and to which the suppression rule is applied. The previous connection rule is a rule for determining whether a character to be suppressed is suppressed or not suppressed based on a relationship with a preceding character. The post-connection rule is a rule for determining whether a character to be suppressed is to be suppressed or not to be suppressed based on a relationship with a subsequent character. Next, in accordance with the order shown in the flow chart of FIG. 4, the Roman alphabet extension rule 3 of FIG. 2 is applied to the input Roman alphabet notation “TOKYO” and the configuration of FIG. This will be described in detail. In FIG. 4, a step S1 is for inputting a Roman character notation. This is done by inputting “TOKYO”, for example, as shown in Roman notation in FIG. In step S2, the Roman alphabet notation input in step S1 is expanded as a notation tree branch. This expands nodes as T, O, K, Y, and O with-as branches, such as "TOOKYO" at the top of the notation tree in FIG. In step S3, one branch is extracted from the notation tree. That is, one branch is extracted from “TOOKYO” at the top of the notation tree obtained by expanding the Roman alphabet notation input in S2. In S4, it is determined whether or not the processing has been completed (whether branch extraction has been completed). Y
In the case of ES, the processing ends. In the case of NO, the process proceeds to S5. In step S5, one character is extracted from the branch (including the succeeding one character for rule application). In step S6, it is determined whether or not the processing has been completed (ie, whether or not character extraction has been completed). If YES, the character extraction from the branch has been completed, and the process is repeated for the next branch in S3. On the other hand, in the case of NO, the process proceeds to S7. In step S7, it is determined whether or not the branch is the first character in the second and subsequent branches. This is to prevent application to characters generated by expansion. In the case of YES, the process returns to S5 because the character is generated by extension.
In the case of NO, since the character is not a character generated by extension, the process proceeds to the next S8. In S8, the rules are applied sequentially from the top of the Roman character extension rule 3. This is applied sequentially from the top of the Roman character extension rule 3 in FIG. In step S9, it is determined whether or not the rule is completed. If YES, proceed to the next branch in S5. In the case of NO, the process proceeds to S10. In S10, it is determined whether or not the rule can be applied. In the case of YES, a branch is applied to the application character of the notation tree, and characters subsequent to the application character are copied to the branch. This corresponds to, for example, the input Romanized notation “TOKY-Y-
The rule number 55 of the Roman alphabet extension rule 3 of FIG. 2 is applied to the second character “O” from the top of “O”,
The output characters “O, OU, OO” are extracted, and for example, for the second “OU”, as shown in FIG. 5, a branch is formed from O in the original Roman alphabet notation, followed by U, followed by After the original Roman alphabet notation, that is, "KYO"
Copy and extend Similarly, for "OO", FIG.
Expand as shown in. Then, S5 and subsequent steps are repeated. The above processing is performed by using the romaji notation “TOK” in FIG.
YO ”is repeatedly applied to create the notation tree of FIG. If this expanded notation tree is expressed in Roman alphabet notation, it becomes the extended Roman alphabet notation (Roman alphabet notation of intermediate candidates) described on the right side of FIG. FIG. 5 shows an application example of the Roman character extension rule of the present invention. In FIG. 5, Roman alphabet notation 1 is the input Roman alphabet notation, and is “TOKYO” here. The inference / control engine 21 applies the Roman alphabet extension rule 3 to the Roman alphabet notation 1. The Roman character extension rule 3 is a Roman character extension rule as shown in FIG. The extended Roman alphabet notation is a Roman alphabet notation obtained by applying the Roman alphabet extension rule 3 to the Roman alphabet notation 1 and expanding it. The notation tree is obtained by making each character of the Roman alphabet notation into nodes, and connecting these nodes with branches (arcs). Here, with respect to the original notation tree “TOKYO” of the Roman alphabet notation, it branches to O, OU, and OO at the time of the second O, and at the time of O in each of the branched notation trees. , OU, and OO, and expanded to a total of nine Roman characters. Next, in accordance with the order shown in the flowchart of FIG. 6, the configuration of FIG. 1A is added to the Roman alphabet notation group (extended Roman alphabet notation of FIG. 5) extended in the flowchart of FIG. A detailed description will be given with reference to FIG. In FIG. 6, in step S21, a Roman character is input. For this, for example, the nine Roman alphabets shown in the extended Roman alphabet notation of FIG. 5 are sequentially input one by one. S22 is 1 from the Roman notation entered in S21.
Characters are extracted (including the following one character for rule application). In step S23, it is determined whether or not the processing has been completed (ie, the character extraction has been completed). If YES, the process ends because the extraction of a series of characters has been completed. In the case of NO, the process proceeds to S24.
In S24, the rules are applied sequentially from the top of the Roman character suppression rules. A step S25 decides whether or not the rule has ended. YE
In the case of S, the Roman notation determined to be necessary in S27 is output. On the other hand, in the case of NO, it is determined in S26 whether or not the rule is applied, and in the case of YES, it is determined that the rule is applied and deleted. On the other hand, in the case of NO, S22 and subsequent steps are repeated. When the above processing is applied to the extended Roman alphabet notation shown in FIG. 7, the ones marked with X conform to the Roman alphabet suppression rule 3 in FIG. 3 and are determined to be unnecessary Roman alphabet notations, and are suppressed (deleted). Four Roman notations will be generated. FIG. 7 shows an application example of the Roman character expansion / suppression rule of the present invention. In FIG. 7, the Roman alphabet 1 is
The entered Roman alphabet notation, in this case "TOKY
O ". The inference / control engine 21 applies the Roman alphabet extension rule 3 and the Roman alphabet suppression rule 4 to Roman alphabet notation 1. Roman alphabet extension rule 3
Is a Roman character extension rule as shown in FIG. The Roman character suppression rule 4 is a Roman character suppression rule as shown in FIG.
With regard to the Roman alphabet notation extended by the above, unnecessary Roman alphabet notation is suppressed (deleted), and necessary Roman alphabet notation is generated. FIG. 8 shows another embodiment of the present invention. This is because you are trying to enter the kanji "Todo" in Roman notation,
An example is shown in which Roman key notation "TODO" is key-inputted from a keyboard. When the Roman alphabet extension rule 3 of FIG. 2 is applied in correspondence with the Roman alphabet notation “TODO”, a total of nine extended Roman alphabet notations described as “illustrated extension” are obtained. When the Roman alphabet suppression rule 3 of FIG. 3 is applied to the expanded nine Roman alphabet notations, the Roman alphabet notation shown as “suppression x” is suppressed (deleted), and as a result, the output Roman alphabet notation “TODO, TODOU, TOU
DO, TOUDOU "is obtained. Then, these Roman characters are converted to kana to obtain "Todo, Todo, Todo, Todo". If one kanji character “Todo” is registered in association with one of these kana characters, for example, “Todo”, the kana kanji character can be converted. According to this embodiment, kana-kanji conversion can be performed without registering other kana “todo”, “todo”, and “todo” in the kana-kanji conversion dictionary. This is particularly effective for proper nouns such as a person's name and land, and there is no need to register kanji in a dictionary in association with various kana, and it is sufficient to register it in correspondence with one kana. As described above, according to the present invention,
By expanding the Roman alphabet notation to obtain multiple Roman alphabet notations, further suppressing unnecessary Roman alphabet notations from among these multiple Roman alphabet notations to generate appropriate Roman alphabet notation, and converting these Roman alphabet notations to Kana , It is possible to generate an appropriate extended Roman alphabet notation from the input Roman alphabet notation 1. The Roman notation can be further converted to Kanji based on the Kana notation obtained by performing the Kana conversion. Thereby, (1) By inputting one Roman alphabet notation, it is possible to automatically generate a plurality of Roman alphabet notation candidates deemed necessary at the time of Kana-Kanji conversion. By converting the generated Roman alphabet notation into kana, converting the kana-kanji conversion dictionary to kanji, and registering the same kanji with fewer kana readings, the kana-kanji conversion dictionary is simplified. . (2) In particular, since the number of words to be registered for proper nouns such as personal names and geography is enormous, reading by kana to be registered in a dictionary is minimized. The required Roman characters are converted to kana, then converted to kana, and a dictionary is searched based on these kana to find the corresponding kanji. This is useful when converting a Roman language notation such as a person's name into a plurality of Roman notations by the processing of the present invention when machine-translating a foreign language into Japanese, then converting it to Kana and looking up a dictionary to find Kanji. .

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a configuration and explanatory diagram of one embodiment of the present invention. FIG. 2 is an example of a Roman character extension rule of the present invention. FIG. 3 is an example of a Roman character suppression rule of the present invention. FIG. 4 is a flowchart of a Roman character extension rule of the present invention. FIG. 5 is an application example of a Roman character extension rule of the present invention. FIG. 6 is a flowchart of a Roman character suppression rule according to the present invention. FIG. 7 is an application example of a Roman character extension / suppression rule of the present invention. FIG. 8 is another specific example of the present invention. FIG. 9 is an example of kana notation / Roman notation. [Description of Signs] 1: Roman alphabet notation 2: Extended Roman alphabet conversion device 21: Inference / control engine 3: Roman alphabet expansion rule 4: Roman alphabet suppression rule 5: Roman alphabet kana conversion unit

Continuation of the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) G06F 17/21-17/26

Claims (1)

(57) [Claims] (1) In a Roman alphabet conversion device for converting a Roman alphabet notation into a plurality of Roman alphabet notations, one character and a succeeding character are sequentially extracted from the head of the Roman alphabet notation to form a plurality of Roman alphabet notations. A Roman character expansion rule to be expanded, and a Roman character suppression rule that determines whether or not to suppress one character and at least one of a pre-connection and a post-connection in order from the top of a plurality of Roman character notations expanded by the Roman character expansion rule Means for extracting one character and the succeeding character in order from the beginning of the input Roman alphabet notation and applying the Roman alphabet extension rule to extend to a plurality of Roman alphabet notations; in order from the beginning of these expanded Roman alphabet notations One character and at least one of a pre-connection and a post-connection are taken out and the Roman character suppression rule is extracted. Romaji conversion apparatus characterized by comprising a means for generating only Roman alphabet that is determined not to inhibit by applying.