JPH0414168A - Word processor - Google Patents

Abstract

PURPOSE:To attain a word processor which has the excellent operability and the reduced candidate correcting frequency by providing a means which converts automatically a word modified by a word changed by a KANA (Japanese syllabary)/KANJI (Chinese character) conversion means into a correct word according to the changed word. CONSTITUTION:Some words contained in a character string converted by a KANA/KANJI converter 3 are changed into other homonyms with retrieval of a KANA/KANJI conversion dictionary 4. Thus a meaning sorting dictionary 5 which stores the meaning sorts obtained by sorting the meaning categories of the substantives by the words modified by the changed words is retrieved together with a case frame dictionary 6 which stores the meaning sorts obtained by sorting the meaning categories of the substantives controlled by the declina ble words as well as the descriptions, the parts of speech, and the case informa tion of the declinable words. Then the words are automatically converted into the correct Japanese words in accordance with the changed words. As a result, the correcting operations of a homonym part can be decreased together with the excellent operability for a word processor which contains a KANA/KANJI conversion means.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a document creation device that creates Japanese documents using kana-kanji conversion, and in particular, converts input Japanese character strings into kanji and kanji characters. This relates to a device for converting mixed sentences.

[Prior Art] A method of inputting Japanese pronunciation and converting it into a sentence containing kanji and kana is as described in Japanese Patent Application Laid-open No. 189565/1982, the conversion result is held in a network state and the result is transferred from the network. There is a method of displaying a string of highly likely candidates on the display and inputting Japanese sentences while selecting the displayed candidates. According to this bulletin, it is possible to use a learning method to display the correct sentence as a conversion result in order to distinguish between the homophone parts of "reporter" and "train", such as "a newspaper reporter was on a train."
come.

[Problems to be solved by the invention] However, in such conventional methods, the following problem 7
There is directness.

First, the learning is performed only when the user performs correction operations while checking the converted kana-kanji content on the display device. operation is required.

For example, if the user's intended conversion result is "listen to a lecture" and the conversion result is "listen to a protest," first modify "protest" to "lecture," and then change "listen" to "listen." It is necessary to perform the correction operation twice.

In addition, because learning is performed using compound words or affixes without breaking them down into independent words, the determination of homophones for independent word parts is not resolved.

As described above, the conventional method has a problem in that, without learning information, homophones cannot be determined, making correction operations for the conversion results complicated.

The purpose of the present invention is to solve the problems of such conventional methods, and to reduce the number of candidate corrections by simply correcting the first term in the homonym part and automatically correcting the homonyms of the following terms. The object of the present invention is to provide a document creation device equipped with an easy-to-use Kana-Kanji conversion function.

[Means for Solving the Problems] The present invention provides an input device for inputting Japanese pronunciations, a kana-kanji conversion means for converting the readings input from the input device into a kanji-kana mixed sentence, and the kana-kanji conversion means. A conversion candidate changing means for changing the converted character string, a display device for visually displaying the converted Japanese sentence, a storage device for storing the converted Japanese sentence, etc., and a control device for controlling these. In the document creation device configured,
When some of the terms in the character string converted by the kana-kanji conversion means are changed to other terms by the conversion candidate changing means, the terms modified by the changed terms are converted into Japanese words corresponding to the changed terms. This is achieved by having a means of automatically converting it into the correct terminology.

[Operation: By using the technical means described above, when the homophone of the preceding term is corrected, the homonym of the subsequent term can also be automatically changed, resulting in improved operability that reduces the need to correct the homophone part. A document creation device equipped with Gana-Kanji conversion can be realized.

[Example 1] Hereinafter, a document creation device according to an embodiment of the present invention will be described.
This will be explained using FIGS. 1 to 10.

FIG. 2 is a block diagram showing the configuration of a document creation device in one embodiment.

1 is a control device such as a microprocessor that controls the overall operation, 2 is an input device for inputting Japanese readings and giving instructions for converting kanji, and 3 is converting the input Japanese reading character string into a mixture of kanji and kana. Kana-kanji conversion device that converts into sentences,
4 is a kana-kanji conversion dictionary that analyzes connections between morphemes, for example used in morphological analysis processing; 5 and 6 are dictionaries used in syntactic-semantic analysis processing, for example, to analyze modification relationships of terms from the results of the morphological analysis processing. , the semantic classification dictionary in 5 stores the semantic categories that classify the semantic categories of nominal words, and the case frame dictionary in 6 classifies the notation of predicates, parts of speech, case information, and the semantic categories of nominal words that are dominated by predicates. 67 is a storage device for storing created documents, etc. 8 is a random access memory (hereinafter referred to as RAM) for temporarily storing ephemeral kanji conversion results, data, etc.
, 9 is an area for temporarily storing a conversion candidate network created by morphological analysis, and 10 is a display device for displaying the input reading character string or kanji conversion result.

FIG. 7 shows a keyboard as an example of the input device 2, in which 21 is a 50-on key for inputting the Japanese reading, 22 is a conversion key for giving an instruction to convert the input reading into a sentence containing kanji and kana, and 23 is a keyboard. is the left cursor key that gives an instruction to move the cursor displayed on the display device of 1o to the left, 24 is the left cursor key that instructs to move the cursor displayed on the display device of 1o
The right cursor key 25 gives an instruction to move the cursor displayed on the display device of o to the right, and 25 is a key that is displayed on the display device of 1o, such as selecting a candidate from the candidate selection matrix or inputting a number. It is a numeric keypad for

Next, the present invention will be explained in detail using FIGS. 1 to 10.

First, from the keyboard 2 of the input device shown in FIG.
Enter. The input reading is temporarily stored in the reading storage area 81 on the RAM 5 shown in FIG. When inputting the reading is completed, the conversion key 22 is operated to start the kana-kanji conversion device 3. The operation of the kana-kanji conversion device will be explained below with reference to FIG.

In the kana-kanji conversion process, morphological analysis process 31 is first performed. In this morphological analysis process, for example, the conversion candidate network 9 shown in FIG.
1 is created and stored in the conversion candidate network-time storage area 9. Syntactic and semantic analysis processing 3 for all combinations stored in this conversion candidate network-time storage area 9
Do step 2. Note that candidate sequences determined to be optimal through morphological analysis are held in the RAM 8.

This method of morphological analysis processing is disclosed in Japanese Patent Application Laid-Open No. 60-189565.
Since the method described in the above publication is known, detailed explanation will be omitted here.

The operation of the syntactic and semantic analysis process 32 will be explained in detail using FIG.

First, the part of speech of the candidate string to be analyzed in the parsing process 321 is set in the part of speech storage table (hereinafter referred to as HINSHI-TBL) in FIG. Divide into units.

Next, in a simple sentence extraction process 322, the automorphism of the clauses is analyzed from the information divided into clause units, and the sentences are divided into sentences (hereinafter referred to as simple sentences) that include at least one pragmatic clause and one nominal clause. In the subsequent case frame analysis feasibility determination processing 323, the clause information of the delimited simple sentences is checked, and combinations in which no noun and predicate do not exist in the simple sentence, such as "protest wo chrysanthemum", are excluded from the case frame analysis processing 324 and are determined as ""HearProtest" is first sent to case frame analysis process 324.

It is determined whether there is a match with some of the patterns that cannot be processed by case frame analysis. If it does not match any of the patterns, it is determined that the case frame can be analyzed, and the process moves on to case frame analysis processing 324. If it matches any of the patterns, it is determined that the case frame cannot be analyzed, and the process moves on to case frame analysis processing 324.
Move on to 6.

If it is determined in the determination process 323 that the case frame cannot be analyzed, in the process 326 the conversion candidate network -
It is checked whether there is a candidate column to be extracted next in the time storage area 9. If there is a candidate string for the combination to be extracted, the candidate string is extracted in the next candidate search 327, and the process returns to the syntax analysis process 321.

In addition, if there is no combination of candidate sequences to be extracted, at the end of the morphological analysis process 31 shown in FIG. 1, the candidate sequences held in the RAM 8 are transferred to the process 33 shown in FIG. 1, and the syntax-semantic analysis process ends. do.

If the case frame can be analyzed in the case frame analysis possibility determination process 323, then in the case frame analysis process 324.

The case frame analysis possibility flag (hereinafter F
LAGI) is set to ``0''.

Based on the phrase “listen” in the simple sentence “listen to the protest” separated by the simple sentence extraction process 322, the case frame dictionary 6 and the nominal phrase “listen to the protest” are used.
The meaning classification dictionary 5 is searched based on "protest" and the relationship between "listen" and "protest" is analyzed. Examples of dictionaries used during analysis are shown in FIGS. 9 and 10. 9 shows the case frame dictionary 6, and FIG. 10 shows the semantic classification dictionary 5. The interrogation in the case frame analysis process 324 is an example of relational analysis, and checks whether the predicate to be converted includes a nominal semantic classification. In the case of ``listen to a protest,'' the predicate ``listen'' is registered in the case frame dictionary shown in Figure 9, and ``listen'' in the case frame dictionary is r voice and music.

The meaning classification of "negotiation, order" has been registered.

Furthermore, the single word "protest" is also registered in the meaning classification dictionary shown in FIG. 10, and "protest" in the meaning classification dictionary is registered with the meaning classification "negotiation". From these two dictionary information,
Since ``listen'' includes the meaning category ``negotiation'' of ``protest,'' ``listen to a protest'' is determined to be correct as Japanese, and RA
Store the simple sentence “Listen to the protest” in the simple sentence storage area (hereinafter referred to as 5I-LIST) that exists in M8, and write “Listen to the protest” in FLAGI.
1”.

Next, a determination process 325 is performed to determine whether or not the case frame analysis process 324 was successful.

The determination process 325 is a determination process 32 if FLAG1=O.
6, and if FLAGl=1, the process moves to determination processing 328.

In the determination process 328, it is determined whether the processing has been completed for all the simple sentences separated by the simple sentence extraction process 322. If the process has not been completed, the process returns to the simple sentence extraction process 322 again, and if it has been completed, in the output notation creation 329, 5I
- Deliver all simple sentences stored in LIST to the process 33 shown in FIG. 1, and end the syntactic and semantic analysis process.

After completing the above syntax and semantic analysis process 32, the delivered candidate string 501 is set and displayed as a conversion result in the character string display buffer 502 of FIG. 5 in process 33.

In the subsequent process 34, other candidate groups registered in the conversion candidate network are set in the display selection matrix 503 and displayed. The conversion result 501 at that time is kept in a state where it can be converted again.

Next, in the key acceptance process 35, if the conversion result is the user's intention, the user inputs a key to perform the next process, and if the conversion result candidate is to be modified, the user inputs the numeric keypad 25 shown in FIG.

In the input key determination process 36, it is determined whether the input key 25 is the keypad 25 for instructing modification of conversion candidates.

If the key is other than the numeric keypad 25, the kana-kanji conversion process is finished and the conversion result is determined.

In the key reception processing 35, when the numeric keypad 25 corresponding to the instruction key for modifying, for example, "protest J" to "lecture J" is input, the notation "lecture" is fixed, and the subsequent conversion candidate network performs syntactic and semantic analysis again in step 32. Perform processing.

In the syntax-semantic analysis process 32, the receptionist first analyzes the relationship between the predicate "listen" and the fixed notation "lecture."

As a result, since "listen" does not include the semantic classification "discourse" of "lecture", we will analyze the first-order terms "listen" and "lecture J."'', the meaning classification ``discourse'' is included, so ``listen to a lecture'' is displayed as the optimal conversion result on the display device 10 as shown in FIG.

With this method, for the converted character string that has been converted into kana-kanji,
When the homophone of the term initially displayed on the display device is corrected to another term, the subsequent term can also be automatically converted into a term that is correct as Japanese.

[Effects of the invention] As described above, according to the present invention, when the first homophone is corrected by converting it into kana-kanji using syntax-semantic analysis, the homophone of the following term is also automatically converted. , it is possible to configure a document creation device equipped with Kana-Kanji conversion that is easy to operate and reduces the number of correction operations for conversion candidates.

[Brief explanation of drawings]

Fig. 1 is a flowchart showing kana-kanji conversion processing in an embodiment of the present invention, Fig. 2 is a system configuration diagram showing an embodiment of the invention, Fig. 3 is a conversion candidate network showing ephemeral-kanji conversion candidates, and Fig. 4 5 is a flowchart showing the syntax-semantic analysis process, 5. 6 is an example of displaying the result of ephemeral kanji conversion, 7 is an input device such as a keyboard, 8 is a part-of-speech storage table, and 9 is an example of a case frame dictionary. , 10th
The figure shows an example of a meaning classification dictionary. DESCRIPTION OF SYMBOLS 1... Control device, 2... Input device, 3... Kana-Kanji conversion device, 7... Storage device, 10... Display device,
25... Cause Figure Figure Figure Cause Figure 4 Figure Figure 6 Figure 7 Figure 8

Claims (1)

[Claims] 1. An input device for inputting a Japanese reading, a kana-kanji conversion means for converting the reading input from the input device into a sentence containing kanji and kana, and a conversion candidate change for changing the character string converted by the kana-kanji conversion means. In a document creation device comprising a means for visually displaying the converted Japanese sentences, a storage device for storing the converted Japanese sentences, etc., and a control device for controlling these, the kana-kanji conversion is performed. When some terms in the character string converted by the means are changed to other homophone terms by the conversion candidate changing means, the words modified by the changed term are changed to 1. A document creation device comprising means for automatically converting words into correct terms.