JPH01229368A - Character processor - Google Patents

Abstract

PURPOSE:To prevent the variance of application for the use of the word pair of the same meaning, and to determine a first candidate by both the means of individual word example and meaning attribute example by giving a meaning attribute code classified in a meaning to an individual word. CONSTITUTION:Example information in a word dictionary DIC stores the word ID of the word of the companion of an example if the example type of it is an individual example, and is constituted of the example, etc. to store a meaning attribute to be the companion of the example if it is a meaning example. Then, the meaning attribute code classified in a meaning is given to the individual word, and as for the example applicable to all the words of the same meaning attribute, the example is made to be capable of being registered by the pair of the word and the meaning attribute or the pair of the meaning attribute and the meaning attribute. Thus, the variance of the application can be suppressed for the use of the word pair of the same meaning. Besides, for a peculiar word or an idiomatic phrase incapable of being classified in a meaning, the first candidate can be determined by both the means of the individual word example and the meaning attribute example by applying the individual word pair example.

Description

DETAILED DESCRIPTION OF THE INVENTION [Fields of Application in Graduation] The present invention relates to a homophone polysemy determination method in a character processing device that converts character strings.

(Prior Art) Conventionally, character processing devices for converting character strings have been widely used, such as kana-kanji conversion devices in Japanese word processors. As a result of Kana-Kanji conversion, many homophone candidates appear, and the ambiguity of the homophones has been resolved based on the grammatical characteristics of the parts of speech of individual words. Since no consideration was given to determining homophone pairs, erroneous conversions often occurred. Therefore, attempts have been made to use word-word pairs, that is, combinations in a word-pair dictionary that stores usage examples, to determine the first candidate.

(The problem that the invention is trying to solve) However,
It is not efficient or realistic to cover all combinations of individual words because the number of combinations is enormous.

Even if a proper conversion first candidate is obtained for a word pair that happens to be registered, it cannot be used for a word pair that is not registered, and the same effect cannot be obtained.

(Means for Solving the Problems (and Effects)) An object of the present invention is to eliminate the drawbacks of the above-mentioned conventional example.

In other words, a semantic attribute code that is semantically classified is given to each word, and for examples of word pair usage that apply to all words with the same semantic attribute, the word and semantic attribute code is assigned instead of the individual word pair. By making it possible to register usage examples using pairs of semantic attributes and semantic attributes, it is possible to suppress variations in application when using pairs of words with the same meaning, and also to prevent unique usage that cannot be categorized semantically. To provide a character processing device capable of determining a first candidate for word pairs and idiomatic phrases by applying individual word pair usage examples and by means of both individual word usage examples and semantic attribute usage examples. It is.

To this end, the present invention includes the following means.

In other words, an input means for inputting the pronunciation of the word to be converted;
A word dictionary that stores semantic attributes that encode the semantic classification of individual words in association with their pronunciations and spellings, as well as co-occurrence pairs between individual words and co-occurrence pairs between individual words and semantic attributes. A usage dictionary, a phrase recognition means for analyzing readings into phrases,
The object of the present invention is to include an example search means for searching for examples that hold true for the analyzed phrase pair, and a conversion means that can give priority to the examples that hold true and determine the first candidate in homophone polysemy.

As a result, for example, the semantic attribute example ``release an animal'' is registered, and all words to which the semantic attribute ``animal'' is given can be converted by applying the above semantic attribute example.

(Example) Hereinafter, the present invention will be described in detail with reference to the drawings.

FIG. 1 is an example of the overall configuration of the present invention.

In the illustrated configuration, the CPU is a microprocessor that performs calculations, logical judgments, etc. for character processing, and uses an address bus AB, a control bus CB, and a data bus D.
B to control each component connected to those buses.

Address bus AB transfers address signals indicating the components to be controlled by the microprocessor CPU. The control bus CB transfers and applies control signals for each component to be controlled by the microprocessor CPU. The data bus DB transfers data between each component device.

Next, ROM is a read-only fixed memory.

Further, the RAM is a writable random access memory having a configuration of 1 word and 16 bits, and is used for temporary storage of various data from each component.

TBLIF is a document buffer, which is a memory for storing document information input from the keyboard KB.

YBUF is an input reading buffer memory that stores readings input from the keyboard KB.

DIC is a word dictionary for performing kana-kanji conversion.

DBP0OL is a homophone candidate storage memory that stores information obtained by analyzing and converting the pronunciation of YBUF into phrases.

YJPOOL is a memory that stores information on examples applied to homophone candidates in conjunction with DBP0OL.

LRNDAT is a learning data storage memory that stores the learning status of individual words and examples.

FZTBL is an adjunct word string conversion table for making adjunct 1ilj correspond to combination information stored in DIC.

KB is a keyboard, which includes alphabet keys, hiragana keys, katakana keys, and other character/symbol keys-1;
It is equipped with various function keys such as a conversion key for instructing conversion.

In the figure, YOM T is the key for inputting the reading;
CON is the conversion instruction key to convert the input reading,
NX and T are the next candidate conversion instruction keys to change the conversion candidate and convert it to the next candidate, and SEL is the selection key to confirm the current homophone display candidate and at the same time instruct to learn the candidate notation. That's the key.

The DISK stores documents created in a memory for storing standard documents, and the stored documents can be called up when necessary by instructions from the keyboard.

CR is a cursor register. The CPU can read and write the contents of the cursor register. A CRT controller CRTC, which will be described later, displays a cursor at a position on the display device CRT corresponding to the address stored here.

DBUF is a display buffer memory that stores patterns such as document information stored in TBUF.

CRTC is cursor register CR and buffer [)BL
It plays the role of displaying the contents read by the IF on the display device CRT.

Further, a CRT is a display device using a cathode ray tube or the like, and the display of a dot configuration pattern and a cursor on the display ff1cRT is controlled by a CRT controller.

Furthermore, CG is a character generator that stores patterns of characters and symbols to be displayed on the display device CRT.

The character processing device of the present invention, which is composed of each of these components, operates in response to various inputs from the keyboard KB, and when input from the keyboard KB is supplied, first, an interleaved signal or a microprocessor CP is output.
The microprocessor CPU reads various control signals stored in the ROM, and various controls are performed according to these control signals. It is a diagram showing the configuration.

Dictionaries are used to store word information used in orthographic conversion and the first information used in orthographic conversion.
It consists of example information for determining candidates. The word information consists of a word ID unique to the header word, a pronunciation that stores the pronunciation of the word, a notation that stores the notation of the word, a part of speech that stores the part of speech of the word, and a code that stores the semantic attribute of the word. The word ID is a consecutive number starting with 1 for the first word. The example information includes the number of example information that the heading word has, the example ID unique to the example, the example type that indicates whether the example is an individual example or a semantic example, and the example type if the example type is an individual example. If it is a semantic example, it stores the semantic attribute of the other word of the example, and a representative word that satisfies the conditions for the example to be true among the attached words attached to the other word of the example. It consists of combination information that stores adjuncts, and reversal information that stores information about whether the example can be used even if it is reversed. For words that do not have any examples, 0 is stored in the number of example information. The example ID is a consecutive number with the first example being 1, just like the previous word ID.

In the figure, the word ``buy'' is distinguished by consecutive numbers 300, and the reading is ``r-ka'' and the notation is ``r-buy''.
, has word information that the part of speech is r-verb J. It has two individual word examples: ``buy alcohol'' and ``buy a fight.'' Further, the word 'ua' is distinguished by a consecutive number of 500, has the pronunciation "sake J," the spelling "salmon," and the part of speech "r noun," and has the semantic attribute r animal j. Since 0 is stored in the number of examples, no example with "r salmon" as a headword is registered.

FIG. 3 is a diagram showing the configuration of the homophone pool DBP0OL and the example information pool YJPOOL. (a) shows the overall configuration of the homophone pool. The homophone pool is composed of each homophone information created by the kana-kanji conversion process. (b) is a diagram showing the structure of homophone information. The homophone information includes the homophone ID, which is a sequential number to distinguish the homophone from other homophones, the data length representing the memory size used for the homophone information, and the number of stored homophone candidates. The number of candidates for storing homophone candidates, candidate information for storing information about homophone candidates, and the priority number for storing the number of priority candidates (described later) among the candidate information stored in the homophones. There is. In such a homonym configuration, the candidate information at the beginning of the candidate information string should be output as the conversion result and become the first candidate, and the number of candidate information indicated by the priority number from the beginning of the string is as described below. This will be the preferred candidate. (C) is a diagram showing the structure of candidate information. Candidate information includes the notation of the candidate and its word ID.
, a clause-final conjugation that stores the conjugation at the end of a clause, and an example information pointer that stores a pointer to example information to be described later.

If the example is not applied to the candidate information and no example information exists, an invalid value -1 is stored in the example information pointer.

(d) shows the overall configuration of the example information pool YJPOOL. The example information pool is composed of example information that stores information on examples that are commonly used for each homophone candidate. (e) is a diagram showing the structure of example information. The example information stores the pair homophone ID that stores the ID of the homophone that is the partner of the example, the example ID that stores the ID of the applied example, and the type of applied example (individual example or semantic example). It consists of an example type, and an example information pointer that stores a pointer to another example information that is used for the same candidate information. In this way, the example information has a list structure. If the next example information does not exist, an invalid value -1 is stored in the example information pointer.

In the configuration of the homophone pool and example information pool as explained above, when an example is applied to a certain candidate for a certain homophone, the example information is stored in the candidate information, and the example information is stored in the candidate information for the homophone. It is assumed that example information containing information on the same example is also stored as candidate information for the candidate.

FIG. 4 is a diagram showing the structure of the adjunct word string conversion table FZTBL. The adjunct word string conversion table is a correspondence conversion table for making adjunct word strings constituting a clause correspond to combination information in the example information held by the word dictionary DIC. The adjunct word reading represents the pronunciation of the adjunct word string, and the corresponding combination information represents the combination information in the example information of the dictionary in which the adjunct word string can be converted.

For example, the adjunct ``koso'' uses the conjunction information ``ga1 or r to J
The combination can be determined by replacing the adjunct word string ``saemonoga-'' with the combination information rga.

FIG. 5 is a diagram showing the structure of learning data LRNDAT. (a) shows the overall structure of the learning data. The learning data consists of two parts: word learning data and example learning data. The pointer to word learning stores the index to the word learning data within the learning data, and the pointer to example learning stores the index to the beginning of the example learning data string within the learning data. (b) is a diagram showing the structure of a word learning data string. The word learning data string is composed of the usage size and each word learning data. The word learning data usage size stores the usage size of the word learning data string. The word learning data corresponds to the ID of each word, and 1 is stored for words that have been learned, and O is stored for 4 words that have not been learned. (C) is a diagram showing the structure of an example learning data string. The example learning data string is composed of the usage size and each example learning data. The example learning data usage size stores the usage size of the example learning data string. The example learning data corresponds to each example ID, and 1 is stored for suppressed prohibited usage examples, and 0 is stored for unsuppressed use or brain usage examples.

The operation of the above embodiment will be explained according to the flow.

FIG. 6 is a flowchart illustrating the operation of the character processing apparatus of the present invention.

At step 56-1, a key is pressed on the keyboard and a wait is made for an interrupt to occur. 36-2 when the key is input
56-3.5 depending on the type of key.
Branch to any of steps 6-4.56-5.56-6.56-7.

56-3 is a process when the reader key YOM I is pressed, and the code of the pressed reading is stored in the input reading buffer memory YBUF.

56-4 is the process when the conversion key CON is pressed, and is input in 56-3 and stored in YBUF.
Converts the character string subject to Kana-Kanji conversion into Kanji and outputs it to the output buffer. When converting to kanji, use the word dictionary D
The first candidate for the homophone is determined using the usage example information present in the IG.

At 56-5, the next candidate conversion key NXT is pressed.

56-4, another homophone candidate in the output buffer output by 56-4 is displayed.

56-6 is a process performed when the selection key SEL is pressed, in which the homophones in the output buffer displayed on the screen are determined, and the determined character strings are output into the document. moreover,
Performs processing to learn the selected words.

56-7 is a process performed when a key other than YOMI, CON, or NXT%SEL (for example, a key used for document editing such as a cursor movement key) is pressed, and is generally performed in character processing devices of the same type. It is a process that
Since it is publicly known, it will not be specifically described.

56-8 is a display process for displaying the changed portion as a result of the above process. This is a commonly used process in which a single character of data in a document is read, developed into a pattern, and output to a display buffer.

FIG. 7 is a detailed flowchart of the process 56-4.

57-1" is a process that analyzes a character string to be converted into kana-kanji that is input in segments and outputs candidates for output of kana-kanji conversion to the homophone pool. A process that sequentially extracts character strings in written units, searches a word dictionary for analysis, and outputs only candidates that are recognized as phrases to a homophone pool, and is generally performed in character processing devices of the same type. Since this is a well-known process, it will not be specifically described.

57-2 checks whether there is an example pattern stored in the word dictionary based on the analysis result output to the homophone pool in 57-1, and if the example pattern exists, it is Homophone candidates to be used as examples are picked up as priority candidates.

57-3 is the priority candidate picked up by 57-2,
From the candidates that have been learned, select the first one for kana-kanji conversion.
Decide on candidates.

57-4 is a process for displaying the output of the kana-kanji conversion stored in the output buffer, and is a process commonly performed in character processing devices of the same type and is well known, so it will not be described in particular.

FIG. 8 is a detailed flowchart of the IA process of 57-2.

Homophones to be extracted as priority candidates are sequentially extracted from the beginning, and the following process is performed on all homophones until there are no more target homophones.

At step 58-1, candidate information in the homophone is extracted.

At 58-2, it is determined whether all candidate information has been processed, and if all candidate information has been processed, priority candidate extraction is ended. If the next candidate information is retrieved, it is determined in 58-3 whether or not the example information exists in the candidate information. Whether or not example information exists can be determined by the example information pointer in the candidate information. That is, if the value of the example information pointer is "-1", no example information exists.

Otherwise, example information exists in the area indicated by the example information pointer. If the example information does not exist, the priority candidate extraction process ends.

If example information exists, backward example information is extracted at 58-4, and forward example information is extracted at 58-5.

When the processing in 58-5 is completed, the processing in 58-1 is performed to process the next candidate information.

FIG. 9 is a detailed flowchart of the process 58-4.

In 59-1, the word ID stored in the candidate information
information on the examples that exist for the word is retrieved one by one from the dictionary.

In 59-2, it is determined in 59-1 whether an example exists in the dictionary for the word of the candidate information and whether all examples in the dictionary have been taken. If the example exists, the process advances to 59-3; if the example does not exist or all the examples in the dictionary have been taken, the backward example extraction process is completed and the process returns.

At step 59-3, it is checked whether the retrieved example can be reversed and applied. That information is stored in a dictionary. If it can be reversed then it is 59-
Go to 4, and if you can't turn it around, go to 59-1.

In step 59-4, it is determined whether or not the input reading can be established when the retrieved example is reversed and applied. If it can be established, the process proceeds to 59-5, and if it cannot be established, the process proceeds to 59-1.

In step 59-5, information on valid examples is stored in example information. In other words, example information is stored as example information in the example information pool, and if no example information exists in the target homophone candidate information, a pointer to the newly created example information is set in the example information pointer. Store. If example information exists in the candidate information, the newly created example information is added to the end of the example information list. In addition, example information is stored not only for the target homophone, but also for the homophone that is the counterpart of the example.

FIG. 10 is a detailed flowchart of the process 59-4.

5IO-1, it is checked whether there is a clause following the target clause. If the following clause exists, proceed to S10-2, otherwise proceed to 5IO
-Go to 7.

In SIO-2, it is checked whether the conjugated form at the end of the target clause is an adnominal form. If it is an adjoint form, proceed to 5IO-3; if not, proceed to 5IO-7.

In S 10-3, from the homophone to the following clause? ii. Get the information one by one.

In 5IO-4, it is determined in S10-3 whether candidate information has been retrieved. If candidate information has been retrieved, the process advances to 510-5, and if all candidate information has been retrieved and no candidate information has been retrieved, the process advances to 5IO-7.

At 5IO-5, it is determined whether the word of the candidate information obtained at 310-3 can be used as an example. That is, if the type of example is an individual example, the Jn word ID stored in the example item of the example information in the dictionary and 310-3
If the AA 3i 1 D of the candidate information obtained in step 2 matches, it will be used as the partner of the example, and if it does not match, it will not be the partner of the example. If the type of example is a semantic example, the semantic attribute stored in the example item of the example information in the dictionary and the semantic attribute obtained from the word ID of the candidate information obtained in S10-3 match. If it does, it will be the target of the example, and if it does not, it will not be the target of the example. If it is determined that it will be the partner of the example, it will proceed to S10-6, and if it is determined that it will not be the partner of the example, it will proceed to 510-3.
Proceed to.

In SIO-6, a return code indicating that the example is true is set.

In SIO-7, set a return code that the example does not hold.

FIG. 11 is a detailed flowchart of the process 58-5.

In 5ll-1, the word I stored in the candidate information
Using D, information on the examples of the word is retrieved from the dictionary one by one.

In step 5ll-2, it is determined in st t-t whether an example exists in the dictionary for word A1 in the candidate information and whether all examples in the dictionary have been taken. If the example exists, the process advances to 5ll-3, and if it is determined whether the example exists or not, and if all the examples in the dictionary have been taken, the forward example extraction process is completed and the process returns.

In step 511-3, it is determined whether or not the input reading can be established if the extracted example is passed in the forward direction. If it can be established, proceed to 5ll-4,
If it cannot be established, proceed to Sl 1-1.

In 5ll-4, information on valid examples is stored in example information. In other words, example information is stored as example information in the example information pool, and if no example information exists in the target homophone candidate information, a pointer to the newly created example information is set in the example information pointer. Store. If example information exists in the candidate information, the newly created example information is added to the end of the example information list. In addition, example information is stored not only for the target homophone, but also for the target homophone.

FIG. 12 is a detailed flowchart of the processing of 5ll-3.

At Sl2-1, it is checked whether there is a clause before the target clause. If the previous clause exists, the process advances to 512-2; otherwise, the process advances to S12-8.

In S+2-2, candidate information is obtained one by one from the homophones for the previous phrase.

In SI 2-3, it is determined in 512-2 whether candidate information has been retrieved. Once the candidate information has been retrieved, proceed to 512-4, and after all candidate information has been retrieved,
If candidate information is not retrieved, the process advances to S12-8.

At 512-4, it is determined whether the word of the candidate information obtained at 512-2 can be used as an example. In other words, if the type of example is an individual example, if the word ID stored in the example item of the example information in the dictionary matches the word ID of the candidate information obtained in step 312-2, then If they do not match, they will not be used in the example. If the example type is a semantic example, the semantic attribute stored in the example item of the example information in the dictionary,
If the semantic attributes obtained from the word ID of the candidate information obtained in S12-2 match, the word will be used as an example, and if they do not match, it will not be used as an example. If it is determined that the user will be the partner of the example, the process advances to 512-5, and if it is determined that the user is not the partner of the example, the process advances to 512-2.

In 512-5, the remaining pronunciations (adjunct word strings) other than independent parts of the candidate information obtained in 512-2 are converted into combined information using the adjunct word string conversion table.

In 512-6, determine whether the combination condition of the homophone candidate of the previous clause or the example is (1:4).In other words, in S I 2-5, the adjunct word string If it cannot be converted into combined information using the conversion table, it is assumed that the combined condition is not satisfied.Also, if there is matching information between the combined information obtained in 512-5 and the combined information of the example information in the dictionary, It is assumed that it satisfies the binding condition, and if there is no matching information, it is assumed that the binding condition is not satisfied.If the binding condition is satisfied, then S I 2-7
If the connection condition is satisfied, the process proceeds to 512-2.

In SIO-7, a return code is set that the example is true.

At 512-8, a return code is set that the example does not hold.

FIG. 13 is a detailed flowchart of the process 57-3.

The homophones for which the first candidate determination must be made are sequentially extracted from the beginning, and the following process is performed on the remaining homophones until there are no target homophones left.

In step 513-1, priority candidates are picked up from candidate information within the homophone. To pick up the priority candidates, among the candidate information of the homophone, the number of example information pointers or those that are not "-1" is set as the priority number, and the candidate information counted in the priority number is sequentially rearranged to the top. When the priority candidates have been picked up for all candidate information, the process proceeds to 513-2, where a process is performed to exclude the suppressed usage examples from the priority candidates. By performing the following process on the priority number of candidate information starting from the beginning of the candidate information, suppression examples are preferentially excluded. First, the example information pointer in the candidate information is followed, the example ID in the example information is extracted, and it is checked whether the example learning data indicated by the extracted example ID is inhibited or deaf. Until the example information pointer reaches the end, that is, the pointer value "-t
As a result of checking whether the 1-column ID for the 1-column ID until `` is shown is suppressed, only if the example information pointer of the candidate information or the example + R report of the 1-column ID of the candidate information is suppressed, the candidate information is excluded from the priority candidates.In other words, the number of priority items is decreased by 1, and the candidate information is rearranged immediately after the end of the priority candidates.

In S13-3, as a result of 313-2, it is determined whether or not there are any priority candidates remaining, based on whether or not the number of priority candidates is O. If there is no priority candidate, the process goes to 513-5.
If more than one remain, the process branches to S13-4.

It is checked whether or not there is a candidate having an example of individual word usage among the candidate information left as priority candidates in Sl, 3-4. The existence check for individual word examples is performed by performing the following process on the priority number of candidate information starting from the beginning of the candidate information. First, follow the example information pointer in the candidate information, and set the type of example for the example information indicated by the pointer 1 to ti'i Pa. If there is at least one individual word example, it is It is determined that there is an example.Until the example information pointer reaches the end, that is, the pointer value is "-1".

If the type of the example is examined and all of the examples are layered or taste attribute examples, it is determined that it is a single individual example.

As a result of the determination, if there is an example of an individual word, the process branches to 513-6, and if there is no example of an individual word, the process branches to 313-7.

At 313-6, as a result of the individual word example search in 313-5, candidate information that is all semantic attribute examples is excluded from the priority candidates. That is, the number of priority items is decreased by 1, and the candidate information is rearranged immediately after the end of the priority candidates.

As a result of the above processing, it is determined in step 313-7 whether or not the priority candidates have been narrowed down to one. That is, it is determined whether the priority number is 1 or not. If the priority number is 1, then 313-
Proceed to step 9 and determine the priority candidate as the first candidate. (If the number of bean tips is not 1, the candidates have not yet been narrowed down to one, so the candidates are narrowed down by word learning in step 513-8.

S l 3-8 extracts the word ID of the priority candidate, checks the word learning data indicated by the word ID, and if there is a word-learned candidate, the word-learned candidate is selected as the first candidate in 513-11. Determine as. ! If there are multiple candidates under the Xi-language learning language, the previously stored candidate information is set as the first candidate. If none of the priority candidates have learned words, 313
-10, the candidate stored at the beginning of the priority candidates is determined as the first candidate.

If it is determined in 513-3 that there is no priority candidate, the word ID of the candidate information is extracted and the word ID is determined in 313-5 to determine the first candidate by learning the vulgar word.
The word learning data indicated by is examined, and if a word-learning candidate exists, the word-learning candidate is determined as the first candidate in 313-11. If there are a plurality of word-learning candidates, the previously stored candidate information is set as the first candidate. All cases? R-HQ for iTi information
If there is no learned information, the first candidate information is determined as the first candidate in 5I3-12.

FIG. 14 is a detailed flowchart of the process 56-6.

In step 514-1, the example information pointer is referred to for all candidate information in the homophone to determine whether an example exists in the homophone in which the candidate whose selection/confirmation is instructed by the selection key SEL is stored. Judgment is made by That is, if the value of the example information pointer is "-1", it is determined that there is no example. If there is no example for all candidate information, the process branches to 514-3 and \, and if there is even one example, the process branches to 514-2.

514-2 was selected. It is determined as follows whether an example established by a 4L word exists in the example information.

First, what are the 41 selected words? Follow the example information pointer in the ii information and find the bare homophone ID and example 10.
Take out. The usage example information pointer of the head candidate information in the homophone indicated by the extracted bare homophone ID is followed, and a search is made to see if there is any homophone ID whose bare homophone ID is the same as the homophone ID in which the selected word exists. Unless such example information is retrieved, it is determined that the example does not hold true. If the bare homophone H61D or example information matching the homophone ID of the selected word is retrieved, the example ID is taken out and checked to see if it matches the example ID in the example information of the selected word. If the example IDs match, it is determined that a valid example exists. If they do not match, it is determined that the example does not hold. If it is determined that the example is valid, the process advances to S14-6, and if it is determined that the example is not valid, a similar determination is made for the next example information of the bare homophone. If all examples are not established for the example information of the bare homophone, the same determination process is performed for the next example information of the selected word. If there is no example information for the selected word, it is determined that there is no valid example, and S
Branch to I 4-5.

514-4 is a process for canceling usage suppression. In step 514-2, the example learning data indicated by the example ID of the example determined to be a valid example is made available for use.

In 514-3, when there is no example of the word selected in 5I4-1, the existence of examples is determined in order to suppress examples of other homophones that were not selected.

To do this, it is determined whether or not example information exists in all example information other than the selected candidate information in the homophones in which the selected word exists, and if examples exist,
In step 514-5, the example learning data indicated by the example ID is disabled.

514-6 performs word learning processing. A learning letter of the base language learning data indicated by the word ID of the selected notation! μ, and performs a process of canceling the learning state of the word learning data indicated by the word ID of the unselected candidate information among the homophones.

At step 514-7, the selected notation is confirmed in the document. Confirmation processing transfers the confirmed character code to the document buffer TBUF.
This is a process that is normally performed in character processing devices of the same type and is well known, so it will not be described in detail.

[Other Examples]

The dictionary used in the above explanation has a single dictionary structure in which word information and usage example information are described under the same headword, but it can also be used similarly as each dictionary table with word information and usage example information separated. can be processed. Furthermore, dictionaries may be configured for each type of example, each dictionary having its own ID, and processing may be performed by giving priority to the examples according to the dictionary ID in which the example exists.

Furthermore, in the above description, examples are applied only between adjacent bunsetsu, but examples can also be easily applied between non-adjacent bunsetsu. In this case, by performing a syntactic analysis process between clauses (an analysis process for dependencies), incorrect application of the example can be avoided.

In the above explanation, we have described the case where usage examples are applied between clauses input in separate writing, but processing for automatically dividing into clauses can also be added to reading sequences input in solid writing. By doing so, you can apply the example in the same way. It is also possible to divide sentences into clauses without considering the existence of examples.

(Explanation of Effects) As is clear from the above explanation, according to the present invention, examples with semantic sameness can be registered all at once, so there is no need to exhaustively register examples of individual words. , it is possible to reduce the cost of the example dictionary, and it is easy to minimize the variation in applied word pairs. Also,
When used in conjunction with individual word examples, it is possible to uniquely determine the first homophone candidate for unique word pairs that cannot be fully described using semantic attribute examples, so there is no need to convert to the next candidate. This allows for efficient input work.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of the overall configuration of the present invention, FIG. 2 is a diagram showing an example of the dictionary configuration of the present invention, FIG. 3 is a diagram showing the configuration of the homophone pool and the example information pool of the present invention, and FIG. FIG. 5 is a diagram showing the structure of the adjunct word conversion table of the present invention, FIG. 5 is a diagram showing the structure of learning data of the present invention, and FIGS. 6 to 14 are flowcharts showing the operation of the character processing system of the present invention. be. Patent applicant: Canon Co., Ltd. Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure E A Figure 12

Claims (1)

[Scope of Claims] An input means for inputting the pronunciation of a word, a word dictionary that stores the semantic attributes of the word while making correspondence between the pronunciation and the notation of the word, and a pair of words having a co-occurrence relationship. an example dictionary that stores a plurality of types of examples of individual words and semantic attribute examples that are pairs of words and semantic attributes; a phrase recognition means for recognizing a phrase as a phrase by the phrase recognition means; a usage example search means for searching the example dictionary for a pair of phrases recognized by the phrase recognition means and extracting valid individual word examples or semantic attribute examples; A character processing device comprising: a conversion means for determining a first candidate notation by giving priority to a word pair constituting an individual word example or a semantic attribute example searched by a search means.